evaluation of ideal wine and cheese pairs using a deviation-from-ideal scale with food and wine...

12
Journal of Food Quality 28 (2005) 245–256. All Rights Reserved. © Copyright 2005, Blackwell Publishing 245 EVALUATION OF IDEAL WINE AND CHEESE PAIRS USING A DEVIATION-FROM-IDEAL SCALE WITH FOOD AND WINE EXPERTS MARJORIE KING 1 and MARGARET CLIFF Agriculture and Agri-Food Canada Pacific Agri-Food Research Center Summerland, BC V0H 1Z0 Canada Accepted for Publication August 4, 2004 ABSTRACT Most information regarding the suitability of wine and cheese pairs is anecdotal information. The objective of this research was to provide recom- mendations based on scientific research for the most desirable “wine & cheese pairs” using nine award-winning Canadian cheeses and 18 BC wines (six white, six red and six specialty wines). Twenty-seven wine and food profes- sionals rated the wine and cheese pairs using a bipolar structured line scale (12 cm). The “ideal pair,” scored at the midpoint of the scale, was defined as a wine and cheese combination where neither the wine nor the cheese domi- nated. For each cheese, mean deviation-from-ideal scores were determined and evaluated by analysis of variance. Scores closest to six were considered “ideal,” while higher or lower scores represented pairs where the “wine” or the “cheese” dominated, respectively. In general, white wines had mean scores closer to six (“ideal”) than either the red or specialty wines. The late harvest, ice and port-type wines were more difficult to pair. Judges varied considerably in their individual assessments reflecting a high degree of per- sonal expectation and preference. INTRODUCTION Wine and cheese have had a long history of food pairing. Archaeologists have found evidence that cheese making dates back to at least 6000 BC. It is thought that early nomadic tribes used animal-skin bags to transport cow and goat’s milk. The fermentation of the milk sugars caused the milk to separate Blackwell Science, LtdOxford, UKJFQJournal of Food Quality0146-9428Copyright 2005 by Food & Nutrition Press, Inc., Trumbull, Connecticut.2005283245256Original Article WINE AND CHEESE PAIRSM. KING and M. CLIFF 1 Corresponding author. TEL: 250-494-7711; FAX: 250-494-0755; EMAIL: [email protected]

Upload: marjorie-king

Post on 21-Jul-2016

216 views

Category:

Documents


4 download

TRANSCRIPT

Page 1: EVALUATION OF IDEAL WINE AND CHEESE PAIRS USING A DEVIATION-FROM-IDEAL SCALE WITH FOOD AND WINE EXPERTS

Journal of Food Quality

28

(2005) 245ndash256

All Rights Reserved

copy

Copyright 2005 Blackwell Publishing

245

EVALUATION OF IDEAL WINE AND CHEESE PAIRS USING A DEVIATION-FROM-IDEAL SCALE WITH FOOD

AND WINE EXPERTS

MARJORIE KING

1

and MARGARET CLIFF

Agriculture and Agri-Food CanadaPacific Agri-Food Research Center

Summerland BC V0H 1Z0Canada

Accepted for Publication August 4 2004

ABSTRACT

Most information regarding the suitability of wine and cheese pairs isanecdotal information The objective of this research was to provide recom-mendations based on scientific research for the most desirable ldquowine amp cheesepairsrdquo using nine award-winning Canadian cheeses and 18 BC wines (sixwhite six red and six specialty wines) Twenty-seven wine and food profes-sionals rated the wine and cheese pairs using a bipolar structured line scale(12 cm) The ldquoideal pairrdquo scored at the midpoint of the scale was defined asa wine and cheese combination where neither the wine nor the cheese domi-nated For each cheese mean deviation-from-ideal scores were determinedand evaluated by analysis of variance Scores closest to six were consideredldquoidealrdquo while higher or lower scores represented pairs where the ldquowinerdquo orthe ldquocheeserdquo dominated respectively In general white wines had meanscores closer to six (ldquoidealrdquo) than either the red or specialty wines The lateharvest ice and port-type wines were more difficult to pair Judges variedconsiderably in their individual assessments reflecting a high degree of per-sonal expectation and preference

INTRODUCTION

Wine and cheese have had a long history of food pairing Archaeologistshave found evidence that cheese making dates back to at least 6000 BC It isthought that early nomadic tribes used animal-skin bags to transport cow andgoatrsquos milk The fermentation of the milk sugars caused the milk to separate

Blackwell Science LtdOxford UKJFQJournal of Food Quality0146-9428Copyright 2005 by Food amp Nutrition Press Inc Trumbull Connecticut2005283245256Original Article

WINE AND CHEESE PAIRSM KING and M CLIFF

1

Corresponding author TEL 250-494-7711 FAX 250-494-0755 EMAIL KingMagrgcca

246 M KING and M CLIFF

into whey which was drunk and curds which were removed drained andsalted to create a high-protein food (Smith 1995) Wine also dates backthousands of years As a fermented beverage it was safer to drink than waterand provided some level of nutrition (Phillips 2000) Because fermentationwas one of the earliest forms of food preservation cheese and wine were anatural combination providing a safe source of a complete protein along witha thirst-quenching liquid that was safe to drink

The early production of both wine and cheese was very rudimentaryresulting in crude products that did not resemble the quality products we seetoday Wine was produced in dirty open vats fermented to a lower alcoholthan most modern wines and was generally unbalanced It was consumed asa daily commodity and was not expected to have a shelf life of more than ayear Cheese was consumed with wine that was produced in the same regionIt has been suggested that both cheese and wine styles are a reflection of their

terrior

so that wines and cheeses fermented in the same region have a naturalpairing (Moessner 2003)

With time wine and cheese moved away from being a daily staple to anincreased emphasis on quality and style People began to try products fromdifferent regions and experiment with different combinations of wine andcheese General guidelines for positive wine and cheese pairings can be foundon the Internet (EampJ Gallo Winery 2002) For example EampJ Gallo Winery(2002) suggests that Softer cheeses need wines with acidity mild cheeses thattaste slightly sweet make dry wines seem acidic and strong cheeses requirestrong wines

It is generally acknowledged that there are a multitude of cheese andwine combinations and the perfect match for any individual has an elementof personal preference There are several factors that contribute to personalchoice Within the human population there is a wide range in perceivedintensities of basic tastes based on taste-bud density so to some extent thereis a physiological basis for preference differences (Bartoshuk 2000) Addi-tionally both cheese and wines are highly aromatic with a wide range oftextures The overall impression of flavor is an integration of aroma and tastewith other mouth-feel sensations such as texture and pungency Psychologicalfactors also play a role in this integration with perceived quality being asso-ciated with past experiences with a particular odor or taste (Noble 1996)

Despite the many factors that contribute to the variability in good wineand cheese matches it is generally agreed that some wine and cheese pairingsare better than others Conventional wisdom dictates that the intensity of thewine must match with that of the cheese (Jackson 2002) Suitable pairings areoften determined by group discussion or by one or two individual ldquoexpertsrdquoHowever there has been very little research conducted under controlled con-ditions to evaluate under blind conditions if a group of food and wine

WINE AND CHEESE PAIRS 247

professionals can establish ldquoidealrdquo pairings for still wine However someinteresting work has been done by Korbel Champagne Cellars in the 1980s todetermine how well ldquochampagnerdquo goes with nongourmet foods resulting ina Champagne Compatibility Index (Baldy 1995) but no such index is avail-able for still wine and cheese This study was designed to determine if sucha group of experts can determine under controlled experimental conditionsthe suitability of a larger number of wine and cheese pairs

MATERIALS AND METHODS

Judges

Twenty-seven judges 8 restaurateurs and 19 wine-industry personnel(winemakers and Vintners Quality Alliance [VQA] judges) were recruited forthe evaluation All had extensive wine and food-matching experience Judgeswere given a brief training session on the technical aspects of cheese tastingand the criteria to use for evaluating the ldquoidealrdquo cheese and wine pair Theldquoideal pairrdquo was defined as a combination of wine and cheese where neitherthe wine nor cheese dominated and the pair together was better than eachindividually

Cheeses

Nine award-winning Canadian cheeses were provided by the Dairy Farm-ers of Canada They were selected to cover a broad spectrum of artisan cheesesthat are available in the Canadian marketplace Hard cheeses were cut into30-g (1 oz) cubes and soft cheeses were cut into 30-g wedges Rind-ripenedcheeses were cut so that a section of rind was left on the sample Cheeseswere randomly assigned an alphabetic code (AndashI) and placed in covered 60-mL plastic cups and served at room temperature A list of cheeses and theircharacteristics are presented in Table 1

Wines

Eighteen commercial British Columbia (BC) VQA wines were selectedfrom the Okanagan Wine Festival wine library located at the Pacific Agri-FoodResearch Center (PARC) Summerland BC Wines were selected by VQApersonnel who had extensive knowledge of the range of character and qualityassociated with BC wines six white six red and six specialty wines werechosen to be representative of their type All wines were labeled with three-digitrandom numbers and served in random order Table 2 lists the varietals winesand their known characteristics as summarized by a panel of VQA judges

248 M KING and M CLIFF

TAB

LE

1

DE

SCR

IPT

ION

OF

NIN

E C

HE

ESE

S U

SED

IN

ST

UD

Y

Nam

eC

ateg

ory

Tast

eA

rom

aTe

xtur

e

Blu

e B

ened

ictin

Sem

isof

t(i

nter

ior

ripe

ned)

Piqu

ant

to s

tron

g an

d sa

ltySl

ight

ly p

unge

ntSm

ooth

cre

amy

use

sldquop

enic

illiu

m r

oque

fort

irdquocu

lture

Boe

renk

aas

Firm

Shar

pPu

ngen

tFi

rm a

nd s

light

ly g

ranu

lar

Cam

embe

rtSo

ftC

ream

y b

utte

ry f

unga

l nu

tty t

angy

as

it ri

pens

But

tery

nut

tySm

ooth

cre

amy

whi

tebl

oom

y ri

ndC

hedd

ar ndash

med

ium

Firm

But

tery

lac

ticM

ilky

mild

Smoo

th s

omew

hat

spri

ngy

Che

ddar

whi

te 3

yea

rsFi

rmA

cidu

lous

sal

ty l

actic

cul

ture

Milk

y s

light

ly f

ruity

Smoo

th t

o gr

ainy

firm

ho

mog

eneo

us n

o op

enin

gsL

e M

igne

ron

de

Cha

rlev

oix

Sem

isof

t(s

urfa

ce r

ipen

ed)

Fully

rip

e nu

tty fl

avor

but

ter

with

a t

ouch

of

acid

ityL

ight

not

es o

f cr

eam

an

d a

hint

of

yogu

rtW

ashe

d ri

nd s

oft

supp

le

slig

htly

ora

nge

surf

ace

rind

an

d so

met

imes

slig

htly

stic

kyO

ka ndash

Cla

ssiq

ueSe

mis

oft

(sur

face

rip

ened

)C

ream

y b

utte

ry n

utty

slig

ht

flora

l sh

arpe

r ta

ste

due

to r

aw m

ilkQ

uite

pro

noun

ced

ndash ag

edSu

pple

ten

der

and

soft

was

hed

rind

Prov

olon

e se

tte f

ette

Fi

rmL

ight

tas

te o

f bu

tter

slig

htly

tar

t an

d ta

ngy

with

ful

l bo

dy w

ith a

geSl

ight

ly m

ilky

fru

ityPa

sta

filat

a fi

brou

s in

teri

or d

ue

to a

ging

gou

rd s

hape

d r

iche

r ye

llow

col

orSw

iss

Firm

Slig

htly

alm

ond

som

e sw

eetn

ess

Plea

sant

nut

ty m

ilky

Uni

form

ly d

istr

ibut

ed r

ound

ldquoe

yes

rdquo so

mew

hat

gritt

y

shin

y s

uppl

e

WINE AND CHEESE PAIRS 249

Experimental Design

Wine and cheese pairings were assigned using a partially balancedincomplete block design The treatment combinations were arranged in aLatin-square formation (Cochran and Cox 1957) In this design each cheeseand wine pair occurred in each tasting position (first second third) the same

TABLE 2DESCRIPTION OF 18 WINES USED IN STUDY

Wine Wine description

Sauvignon Blanc Slight herbaceous hints mixed with light pear and gooseberry fruit and a hint of oak Medium to light body with a spicy vanilla aftertaste

Chardonnay (unoaked) Apple pear and citrus aromas and flavors No oak or malolactic character Full body with firm acid and long fresh fruity aftertaste

Pinot Gris Fruity apple pear aromas and flavors Touch of honey Clean wine with good body and balance and a long smooth finish

Chardonnay Citrus apple fruit on aroma and palate Rich oak character with caramel and butternotes Firm acid with full body and a long aftertaste

Gewurztraminer Floral perfume spicy and lychee aromas Some pear pineapple and tropical characters Clean with good balance Long smooth finish

Riesling Fruity and fresh with apricot peach and apple aromas and flavor Crisp dry well balanced wine with medium body and length of finish

Pinot noir (light) Light clean berry cherry fruit Delicate character with some toasty vanilla oak Soft tannins with a relatively short finish

Pinot noir (oaked) Oaky with a toasty caramel spicy character Jammy cherry and strawberry flavors Soft round finish with a hint of residual sweetness

Merlot Lots of ripe jammy fruit raspberry and cherry on aroma and palate Spicy character A complex wine and toasty caramel vanilla notes

Meritage Toasted oak black pepper and ripe fruit (cherry plum and prune) aroma smokey character Full tannins and good body Long aftertaste

Foch Smokey spicy oaky aromas Some earthy cedar vegetative character on palate Clean with good body and balance Medium long finish

Sparkling Clean fresh fruit Ripe fruit with citrus apricot and pear characters on both aroma and palate Good body and balance with a long smooth finish Nice mousse on palate

Gamay blush Fruity light clean wine Light strawberry and raspberry character with a hint of black pepper Good body balance and long smooth finish

Gewuztraminer (late harvest)

Lychee character with citrus apricot peach and pear fruit character Some perfume floral and honey notes Good late harvest concentration of flavors Sweet with lower acid

Riesling Icewine Concentrated apple and citrus fruit flavors with a touch of honey High residual sugar with acid to balance with viscous mouth feel

Pinot noir Icewine Smokey baked-fruit character on nose Some light raspberry and strawberry on palate Concentrated fruit sweetness and acid with a full body and long finish

Port type Dried fruit plum and oak on aroma and palate Slightly hot with some residual sugar and sweet lingering aftertaste

250 M KING and M CLIFF

number of times allowing experimental error from ldquocarryoverrdquo or ldquoresidualrdquoto be balanced across all samples

The wines were divided into three blocks of six white red and spe-cialty wines respectively The white and red wines were tasted in a randomorder while the specialty wines had an assigned order of sparkling winefirst and Icewines last This was thought to be necessary because Icewinesare very sweet and intense with a potential for a strong carryover effect thatcould interfere with a judgersquos ability to assess the other cheese and winepairs

Using the incomplete block design each judge tasted three cheeses withsix of the wines The cheeses were systematically paired with the wines sothat after the three blocks of wines a judge had tasted all nine cheeses andeach cheese had been evaluated nine times with each of the wines Judgeswere asked to evaluate the suitability of the pairing for each cheese with allof the wines in the block before moving on to the next cheese

Tasting Process

Judges were instructed to take a small bite of cheese sip the wine andevaluate the appropriateness of the pair The judges evaluated the match usinga variation of the ldquojust right scalerdquo (Shepherd

et al

1989) Judges scored thedeviation from an ldquoideal matchrdquo on a 12-cm line scale where the ideal matchwas the midpoint of the line (6 cm) If the judges thought the cheese domi-nated the wine they marked the line scale to the left of the midpoint whereasif they thought the wine dominated the cheese the judges marked the line scaleto the right of the midpoint The scale was labeled with ldquocheese dominatesexcessivelyrdquo ldquocheese dominates moderatelyrdquo and ldquocheese dominates slightlyrdquoat 0 2 and 4 cm The other half of the scale was labeled with ldquowine dominatesslightlyrdquo ldquowine dominates moderatelyrdquo and ldquowine dominates excessivelyrdquo at8 10 12 cm respectively

Judges were given sparkling water bread and melba toast in order toclear their palates between evaluations They were also instructed to take ashort break between cheeses After the tasting the scorecards were collectedand the judgesrsquo scores for each wine and cheese pairing were measured(0ndash12) for data analysis and interpretation

Statistical Analysis

A single factor analysis of variance (ANOVA) (Microsoft Excel Seat-tle WA) was used to determine if there were significant differences in thesuitability of the match between each cheese with each wine A Fisherrsquosleast significant difference (LSD) was calculated to determine if the ldquoidealrdquoscores were significantly different from each other The average standard

WINE AND CHEESE PAIRS 251

deviation from the mean was calculated for each cheese and wine combina-tion This was then averaged across all wines to give an indication of judgevariability

RESULTS AND DISCUSSION

Results of the ANOVA are shown in Table 3 As might be expected mostcheeses showed significant differences (

P

lt

0001) in the suitability of theirpairing with the 18 wines More surprisingly two cheeses (Swiss and mediumCheddar) showed no differences in their suitability with any of the winesTable 3 tabulates the mean score for each of the wine and cheese pairs Anideal pairing would have a mean score of six a mean score greater than sixwould indicate that the wine dominated the cheese a score less than six wouldindicate that the cheese dominated

Overall the relatively small deviations-from-ideal confirmed that wineand cheese are compatible food pairings In the three categories of wine whitewines tended to be a better match with these cheeses This was indicated byan average deviation-from-ideal of 085 compared to a deviation of 12 and17 for the red and specialty wines respectively The most versatile white winewas the Riesling followed by the Sauvignon blanc and the Pinot Gris asreflected by the smallest deviations The Pinot noir wine was the versatile redwine in that it paired best with the most number of cheeses The specialtysparkling and blush wines also paired well with many of the cheeses whilethe late harvest ice and port wines were the most difficult to pair with manyof the cheeses The Blue Benedictine Oka and Provolone cheeses were themost suitable match with these wines thus supporting the general rule thatstronger flavored cheeses tend to be better matched with stronger flavoredwines In general in this study the specialty wines (late harvest icewine porttype) dominated the cheeses However the addition of other foods or a dif-ference in the tasting method might change the perception of the pairing andlikely make these wines more compatible

Four cheeses (aged white Cheddar Migerone de Charlevoix Swiss andmedium Cheddar) had mean scores greater than six indicating that the wineswere more strongly flavored than the cheese Although the mean scores forthe 3-year old Cheddar were greater than six the means for the white winesthe Pinot noir and Merlot wines were not significantly different than six Thisindicated that there would be a range of wines that would be a suitable matchwith this cheese In general the stronger flavored cheeses were the mostvariable in their matches with the wines This indicated that more care wouldbe required in selecting wines to pair with these cheeses Figure 1 shows thedeviation-from-ideal for the match of the Benedictine Blue with each of the

252 M KING and M CLIFF

TAB

LE

3

ME

AN

DE

VIA

TIO

N-F

RO

M-I

DE

AL

SC

OR

ES

FOR

9 C

HE

ESE

S PA

IRE

D W

ITH

18

WIN

ES

(

n

=

9)

Win

ety

peW

ine

no

Win

eL

a pr

ovid

ence

d

cent

Oka

Blu

e B

ened

ictin

Prov

olon

eA

ged

3 ye

arC

hedd

ar

Mig

nero

n de

C

harl

evoi

x

Boe

renk

ass

Cam

embe

rtSw

iss

Med

ium

Che

ddar

Mea

n(

n

=

81)

acro

ss a

llch

eese

s

Whi

te1

Pino

t G

ris

50

61

62

68

76

65

55

75

77

654

Whi

te2

Rie

slin

g5

05

15

56

97

35

84

66

57

96

07W

hite

3Sa

uvig

non

blan

c4

04

45

16

86

15

76

46

77

25

82W

hite

4C

hard

onna

y (o

aked

)5

87

08

57

56

36

53

77

77

26

69W

hite

5G

ewur

ztra

min

er5

86

05

67

47

45

45

89

47

96

74W

hite

6C

hard

onna

y (u

noak

ed)

60

76

82

79

76

70

57

72

82

727

Red

7Pi

not

noir

(oa

ked)

52

59

58

65

77

71

68

65

67

647

Red

8R

ed M

erita

ge5

66

85

48

48

77

06

98

97

57

24R

ed9

Foch

(oa

ked)

48

65

80

83

90

76

70

87

86

761

Red

10Sy

rah

51

65

64

73

84

63

71

86

71

698

Red

11Pi

not

noir

(lig

ht)

40

40

57

52

67

66

57

62

68

566

Red

12M

erlo

t5

25

86

77

69

28

66

38

58

27

34Sp

ecia

lty13

Spar

klin

g6

35

45

86

57

06

15

37

67

36

37Sp

ecia

lty14

Blu

sh4

74

74

06

06

26

46

36

96

85

78Sp

ecia

lty15

LH

Gew

urzt

ram

iner

80

70

96

78

86

75

70

75

83

792

Spec

ialty

16R

iesl

ing

Icew

ine

79

73

87

83

97

82

83

90

100

860

Spec

ialty

17Pi

not

noir

Ice

win

e7

78

08

38

77

27

88

19

29

78

30Sp

ecia

lty18

Port

typ

e8

37

88

010

29

410

110

49

18

69

10St

anda

rd D

evia

tion

26

24

31

24

26

25

31

27

22

naL

SD1

891

752

241

751

961

852

19ns

nsna

P

pound

000

040

004

9E-0

40

003

002

000

60

0006

nsns

naR

ange

40ndash

79

40ndash

78

40ndash

87

52ndash

102

61ndash

97

54ndash

101

46ndash

104

65ndash

94

67ndash

10

WINE AND CHEESE PAIRS 253

wines In each of the wine categories there were wines where the cheesedominated and others where the wine dominated

The average standard deviation for each cheese across all wines rangedfrom 31 to 22 (Table 3) This relatively high deviation indicated the judgeswere not in agreement on their evaluation of the match For a trained descrip-tive panel using a 12-cm scale the standard deviation is generally 1ndash15Figure 2 showed the judge variation in assessing the match of Provolonecheese for three of the wines Although the mean score was not significantlydifferent from an ideal match of six the wide range of judgesrsquo assessmentswas evident For each of the three wines some of the judges thought thecheese dominated the wine while others thought the wine dominated thecheese Figure 3 showed that although there was some variation in the judgeassessment of how the medium Cheddar paired with the various wines therewas better agreement that the wine tended to dominate the cheese The dif-ference in variability is most likely due to the cheese with Provolone beinga stronger cheese and more difficult to match

It was very important to recognize the extent of the judge variation incheese and wine assessments Although all of the judges were individually

FIG 1 MEAN DEVIATION-FROM-IDEAL FOR BLUE CHEESE PAIRED WITH 18 WINES (

n

=

9)

ndash2

ndash15

ndash1

ndash05

0

50

1

51

2

52

Idea

l mat

ch

1

2

3

5

4 0196

21

11

87

41

7161 8151

31

winesetihWGristoniP1gnilseiR2

cnalbnongivuaS3)dekao(yannodrahC4

renimartzruweG5)dekaonu(yannodrahC6

Cheese dominates

Wine dominates

winesdeR)dekao(riontoniP7

egatireM8hcoF9haryS01

)thgil(riontoniP11tolreM21

winesytlaicepSeniwgnilkrapS31

hsulB-esoR41HLrenimartzruweG51

eniwecIgnilseiR61eniwecIriontoniP71

epyttroP81

254 M KING and M CLIFF

FIG 2 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING PROVOLONE CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

renimartzruweG

)thgil(riontoniPeniwgnilkrapS

STD = 315

STD = 361

STD = 417

Cheese dominates

Wine dominates

FIG 3 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING MEDIUM CHEDDAR CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

)thgil( rion toniPrenimartzruweG eniw gnilkrapS

STD = 266 STD = 247STD = 269

Wine dominates

Cheese dominates

WINE AND CHEESE PAIRS 255

experienced in food and wine pairing as a group their results under experi-mental conditions were not always consistent ie they did not always agreeTherefore the results obtained in this study are only broad recommendationsfor possible cheese and wine pairings In the light of the range of assessmentsobtained by knowledgeable judges it would be important to empower con-sumers to have the confidence to make their own assessment of how muchthey enjoy the cheese and wine pair

In conclusion white wines were easier to pair with a broader range ofcheeses than red or specialty wines The stronger more flavorful cheeses weremore difficult to pair with the wines but were more likely to be a good matchfor the late harvest and ice wines than other milder cheeses The results ofthis study were determined with a single tasting protocol and the wine andcheese were tasted without other foods The addition of other foods or agreater time period between the assessments would change the perception ofthe pairing Finally the variation in judge assessment indicates a high degreeof personal expectation or preference associated with the suitability of wineand cheese pairing Round-table discussion and consensus developmentmight produce ldquotrainedrdquo judges with less variation in the assessments althoughthis might be difficult given the ldquorespectedrdquo opinions of acknowledgedexperts Individuals should be encouraged to experiment to determine theirown preferences

ACKNOWLEDGMENTS

The authors would like to thank Reg Hendrickson of the Dairy Farmersof Canada for donating the cheese Christa-Lee Bond of the Okanagan WineFestival Society for allowing this research to be conducted as part of theIcewine Festival events Kimberly Dever for her technical assistance and theBC Winemakers and restaurateurs who participated in the study

REFERENCES

BALDY MW 1995

The University Wine Course

p 426 Wine AppreciationGuild San Francisco CA

BARTOSHUK LM 2000 Comparing sensory experiences across individu-als Recent psychophysical advances illuminate genetic variation in tasteperception Chem Senses

25

447ndash460COCHRAN WG amp COX GM 1957

Experimental Designs

p 611 JohnWiley amp Sons New York

256 M KING and M CLIFF

EampJ GALLO WINERY 2002 Gallo of sonoma wine and cheese pairingguide ABCrsquos of wine and cheese httpwwwwinewebcentralcomawineandcheeseabchtm (accessed October 24 2003)

JACKSON RS 2002

Wine Tasting A Professional Handbook

pp 261ndash262Academic Press San Diego CA

MOESSNER M Inniskillinrsquos guide to wine and cheese pairings httpwwwwineandleisurecomcheesepairinghtml (accessed October 242003)

NOBLE AC 1996 Tastendasharoma interactions Trends Food Sci Technol

7

439ndash444

PHILLIPS R 2000

A Short History of Wine

Penguin Books Londonhttpwwwintowinecomshop1282-0excerpthtml (accessed October24 2003)

SHEPHERD R SMITH K and FARLEIGH CA 1989 The relationshipbetween intensity hedonic and relative-to-ideal ratings Food QualPrefer

1

75ndash80SMITH J 1995

Cheesemaking in Scotland ndash A History

Scottish DairyAssociation Scotland httpwwwefrhwacukSDAbook1html(accessed October 24 2003)

Page 2: EVALUATION OF IDEAL WINE AND CHEESE PAIRS USING A DEVIATION-FROM-IDEAL SCALE WITH FOOD AND WINE EXPERTS

246 M KING and M CLIFF

into whey which was drunk and curds which were removed drained andsalted to create a high-protein food (Smith 1995) Wine also dates backthousands of years As a fermented beverage it was safer to drink than waterand provided some level of nutrition (Phillips 2000) Because fermentationwas one of the earliest forms of food preservation cheese and wine were anatural combination providing a safe source of a complete protein along witha thirst-quenching liquid that was safe to drink

The early production of both wine and cheese was very rudimentaryresulting in crude products that did not resemble the quality products we seetoday Wine was produced in dirty open vats fermented to a lower alcoholthan most modern wines and was generally unbalanced It was consumed asa daily commodity and was not expected to have a shelf life of more than ayear Cheese was consumed with wine that was produced in the same regionIt has been suggested that both cheese and wine styles are a reflection of their

terrior

so that wines and cheeses fermented in the same region have a naturalpairing (Moessner 2003)

With time wine and cheese moved away from being a daily staple to anincreased emphasis on quality and style People began to try products fromdifferent regions and experiment with different combinations of wine andcheese General guidelines for positive wine and cheese pairings can be foundon the Internet (EampJ Gallo Winery 2002) For example EampJ Gallo Winery(2002) suggests that Softer cheeses need wines with acidity mild cheeses thattaste slightly sweet make dry wines seem acidic and strong cheeses requirestrong wines

It is generally acknowledged that there are a multitude of cheese andwine combinations and the perfect match for any individual has an elementof personal preference There are several factors that contribute to personalchoice Within the human population there is a wide range in perceivedintensities of basic tastes based on taste-bud density so to some extent thereis a physiological basis for preference differences (Bartoshuk 2000) Addi-tionally both cheese and wines are highly aromatic with a wide range oftextures The overall impression of flavor is an integration of aroma and tastewith other mouth-feel sensations such as texture and pungency Psychologicalfactors also play a role in this integration with perceived quality being asso-ciated with past experiences with a particular odor or taste (Noble 1996)

Despite the many factors that contribute to the variability in good wineand cheese matches it is generally agreed that some wine and cheese pairingsare better than others Conventional wisdom dictates that the intensity of thewine must match with that of the cheese (Jackson 2002) Suitable pairings areoften determined by group discussion or by one or two individual ldquoexpertsrdquoHowever there has been very little research conducted under controlled con-ditions to evaluate under blind conditions if a group of food and wine

WINE AND CHEESE PAIRS 247

professionals can establish ldquoidealrdquo pairings for still wine However someinteresting work has been done by Korbel Champagne Cellars in the 1980s todetermine how well ldquochampagnerdquo goes with nongourmet foods resulting ina Champagne Compatibility Index (Baldy 1995) but no such index is avail-able for still wine and cheese This study was designed to determine if sucha group of experts can determine under controlled experimental conditionsthe suitability of a larger number of wine and cheese pairs

MATERIALS AND METHODS

Judges

Twenty-seven judges 8 restaurateurs and 19 wine-industry personnel(winemakers and Vintners Quality Alliance [VQA] judges) were recruited forthe evaluation All had extensive wine and food-matching experience Judgeswere given a brief training session on the technical aspects of cheese tastingand the criteria to use for evaluating the ldquoidealrdquo cheese and wine pair Theldquoideal pairrdquo was defined as a combination of wine and cheese where neitherthe wine nor cheese dominated and the pair together was better than eachindividually

Cheeses

Nine award-winning Canadian cheeses were provided by the Dairy Farm-ers of Canada They were selected to cover a broad spectrum of artisan cheesesthat are available in the Canadian marketplace Hard cheeses were cut into30-g (1 oz) cubes and soft cheeses were cut into 30-g wedges Rind-ripenedcheeses were cut so that a section of rind was left on the sample Cheeseswere randomly assigned an alphabetic code (AndashI) and placed in covered 60-mL plastic cups and served at room temperature A list of cheeses and theircharacteristics are presented in Table 1

Wines

Eighteen commercial British Columbia (BC) VQA wines were selectedfrom the Okanagan Wine Festival wine library located at the Pacific Agri-FoodResearch Center (PARC) Summerland BC Wines were selected by VQApersonnel who had extensive knowledge of the range of character and qualityassociated with BC wines six white six red and six specialty wines werechosen to be representative of their type All wines were labeled with three-digitrandom numbers and served in random order Table 2 lists the varietals winesand their known characteristics as summarized by a panel of VQA judges

248 M KING and M CLIFF

TAB

LE

1

DE

SCR

IPT

ION

OF

NIN

E C

HE

ESE

S U

SED

IN

ST

UD

Y

Nam

eC

ateg

ory

Tast

eA

rom

aTe

xtur

e

Blu

e B

ened

ictin

Sem

isof

t(i

nter

ior

ripe

ned)

Piqu

ant

to s

tron

g an

d sa

ltySl

ight

ly p

unge

ntSm

ooth

cre

amy

use

sldquop

enic

illiu

m r

oque

fort

irdquocu

lture

Boe

renk

aas

Firm

Shar

pPu

ngen

tFi

rm a

nd s

light

ly g

ranu

lar

Cam

embe

rtSo

ftC

ream

y b

utte

ry f

unga

l nu

tty t

angy

as

it ri

pens

But

tery

nut

tySm

ooth

cre

amy

whi

tebl

oom

y ri

ndC

hedd

ar ndash

med

ium

Firm

But

tery

lac

ticM

ilky

mild

Smoo

th s

omew

hat

spri

ngy

Che

ddar

whi

te 3

yea

rsFi

rmA

cidu

lous

sal

ty l

actic

cul

ture

Milk

y s

light

ly f

ruity

Smoo

th t

o gr

ainy

firm

ho

mog

eneo

us n

o op

enin

gsL

e M

igne

ron

de

Cha

rlev

oix

Sem

isof

t(s

urfa

ce r

ipen

ed)

Fully

rip

e nu

tty fl

avor

but

ter

with

a t

ouch

of

acid

ityL

ight

not

es o

f cr

eam

an

d a

hint

of

yogu

rtW

ashe

d ri

nd s

oft

supp

le

slig

htly

ora

nge

surf

ace

rind

an

d so

met

imes

slig

htly

stic

kyO

ka ndash

Cla

ssiq

ueSe

mis

oft

(sur

face

rip

ened

)C

ream

y b

utte

ry n

utty

slig

ht

flora

l sh

arpe

r ta

ste

due

to r

aw m

ilkQ

uite

pro

noun

ced

ndash ag

edSu

pple

ten

der

and

soft

was

hed

rind

Prov

olon

e se

tte f

ette

Fi

rmL

ight

tas

te o

f bu

tter

slig

htly

tar

t an

d ta

ngy

with

ful

l bo

dy w

ith a

geSl

ight

ly m

ilky

fru

ityPa

sta

filat

a fi

brou

s in

teri

or d

ue

to a

ging

gou

rd s

hape

d r

iche

r ye

llow

col

orSw

iss

Firm

Slig

htly

alm

ond

som

e sw

eetn

ess

Plea

sant

nut

ty m

ilky

Uni

form

ly d

istr

ibut

ed r

ound

ldquoe

yes

rdquo so

mew

hat

gritt

y

shin

y s

uppl

e

WINE AND CHEESE PAIRS 249

Experimental Design

Wine and cheese pairings were assigned using a partially balancedincomplete block design The treatment combinations were arranged in aLatin-square formation (Cochran and Cox 1957) In this design each cheeseand wine pair occurred in each tasting position (first second third) the same

TABLE 2DESCRIPTION OF 18 WINES USED IN STUDY

Wine Wine description

Sauvignon Blanc Slight herbaceous hints mixed with light pear and gooseberry fruit and a hint of oak Medium to light body with a spicy vanilla aftertaste

Chardonnay (unoaked) Apple pear and citrus aromas and flavors No oak or malolactic character Full body with firm acid and long fresh fruity aftertaste

Pinot Gris Fruity apple pear aromas and flavors Touch of honey Clean wine with good body and balance and a long smooth finish

Chardonnay Citrus apple fruit on aroma and palate Rich oak character with caramel and butternotes Firm acid with full body and a long aftertaste

Gewurztraminer Floral perfume spicy and lychee aromas Some pear pineapple and tropical characters Clean with good balance Long smooth finish

Riesling Fruity and fresh with apricot peach and apple aromas and flavor Crisp dry well balanced wine with medium body and length of finish

Pinot noir (light) Light clean berry cherry fruit Delicate character with some toasty vanilla oak Soft tannins with a relatively short finish

Pinot noir (oaked) Oaky with a toasty caramel spicy character Jammy cherry and strawberry flavors Soft round finish with a hint of residual sweetness

Merlot Lots of ripe jammy fruit raspberry and cherry on aroma and palate Spicy character A complex wine and toasty caramel vanilla notes

Meritage Toasted oak black pepper and ripe fruit (cherry plum and prune) aroma smokey character Full tannins and good body Long aftertaste

Foch Smokey spicy oaky aromas Some earthy cedar vegetative character on palate Clean with good body and balance Medium long finish

Sparkling Clean fresh fruit Ripe fruit with citrus apricot and pear characters on both aroma and palate Good body and balance with a long smooth finish Nice mousse on palate

Gamay blush Fruity light clean wine Light strawberry and raspberry character with a hint of black pepper Good body balance and long smooth finish

Gewuztraminer (late harvest)

Lychee character with citrus apricot peach and pear fruit character Some perfume floral and honey notes Good late harvest concentration of flavors Sweet with lower acid

Riesling Icewine Concentrated apple and citrus fruit flavors with a touch of honey High residual sugar with acid to balance with viscous mouth feel

Pinot noir Icewine Smokey baked-fruit character on nose Some light raspberry and strawberry on palate Concentrated fruit sweetness and acid with a full body and long finish

Port type Dried fruit plum and oak on aroma and palate Slightly hot with some residual sugar and sweet lingering aftertaste

250 M KING and M CLIFF

number of times allowing experimental error from ldquocarryoverrdquo or ldquoresidualrdquoto be balanced across all samples

The wines were divided into three blocks of six white red and spe-cialty wines respectively The white and red wines were tasted in a randomorder while the specialty wines had an assigned order of sparkling winefirst and Icewines last This was thought to be necessary because Icewinesare very sweet and intense with a potential for a strong carryover effect thatcould interfere with a judgersquos ability to assess the other cheese and winepairs

Using the incomplete block design each judge tasted three cheeses withsix of the wines The cheeses were systematically paired with the wines sothat after the three blocks of wines a judge had tasted all nine cheeses andeach cheese had been evaluated nine times with each of the wines Judgeswere asked to evaluate the suitability of the pairing for each cheese with allof the wines in the block before moving on to the next cheese

Tasting Process

Judges were instructed to take a small bite of cheese sip the wine andevaluate the appropriateness of the pair The judges evaluated the match usinga variation of the ldquojust right scalerdquo (Shepherd

et al

1989) Judges scored thedeviation from an ldquoideal matchrdquo on a 12-cm line scale where the ideal matchwas the midpoint of the line (6 cm) If the judges thought the cheese domi-nated the wine they marked the line scale to the left of the midpoint whereasif they thought the wine dominated the cheese the judges marked the line scaleto the right of the midpoint The scale was labeled with ldquocheese dominatesexcessivelyrdquo ldquocheese dominates moderatelyrdquo and ldquocheese dominates slightlyrdquoat 0 2 and 4 cm The other half of the scale was labeled with ldquowine dominatesslightlyrdquo ldquowine dominates moderatelyrdquo and ldquowine dominates excessivelyrdquo at8 10 12 cm respectively

Judges were given sparkling water bread and melba toast in order toclear their palates between evaluations They were also instructed to take ashort break between cheeses After the tasting the scorecards were collectedand the judgesrsquo scores for each wine and cheese pairing were measured(0ndash12) for data analysis and interpretation

Statistical Analysis

A single factor analysis of variance (ANOVA) (Microsoft Excel Seat-tle WA) was used to determine if there were significant differences in thesuitability of the match between each cheese with each wine A Fisherrsquosleast significant difference (LSD) was calculated to determine if the ldquoidealrdquoscores were significantly different from each other The average standard

WINE AND CHEESE PAIRS 251

deviation from the mean was calculated for each cheese and wine combina-tion This was then averaged across all wines to give an indication of judgevariability

RESULTS AND DISCUSSION

Results of the ANOVA are shown in Table 3 As might be expected mostcheeses showed significant differences (

P

lt

0001) in the suitability of theirpairing with the 18 wines More surprisingly two cheeses (Swiss and mediumCheddar) showed no differences in their suitability with any of the winesTable 3 tabulates the mean score for each of the wine and cheese pairs Anideal pairing would have a mean score of six a mean score greater than sixwould indicate that the wine dominated the cheese a score less than six wouldindicate that the cheese dominated

Overall the relatively small deviations-from-ideal confirmed that wineand cheese are compatible food pairings In the three categories of wine whitewines tended to be a better match with these cheeses This was indicated byan average deviation-from-ideal of 085 compared to a deviation of 12 and17 for the red and specialty wines respectively The most versatile white winewas the Riesling followed by the Sauvignon blanc and the Pinot Gris asreflected by the smallest deviations The Pinot noir wine was the versatile redwine in that it paired best with the most number of cheeses The specialtysparkling and blush wines also paired well with many of the cheeses whilethe late harvest ice and port wines were the most difficult to pair with manyof the cheeses The Blue Benedictine Oka and Provolone cheeses were themost suitable match with these wines thus supporting the general rule thatstronger flavored cheeses tend to be better matched with stronger flavoredwines In general in this study the specialty wines (late harvest icewine porttype) dominated the cheeses However the addition of other foods or a dif-ference in the tasting method might change the perception of the pairing andlikely make these wines more compatible

Four cheeses (aged white Cheddar Migerone de Charlevoix Swiss andmedium Cheddar) had mean scores greater than six indicating that the wineswere more strongly flavored than the cheese Although the mean scores forthe 3-year old Cheddar were greater than six the means for the white winesthe Pinot noir and Merlot wines were not significantly different than six Thisindicated that there would be a range of wines that would be a suitable matchwith this cheese In general the stronger flavored cheeses were the mostvariable in their matches with the wines This indicated that more care wouldbe required in selecting wines to pair with these cheeses Figure 1 shows thedeviation-from-ideal for the match of the Benedictine Blue with each of the

252 M KING and M CLIFF

TAB

LE

3

ME

AN

DE

VIA

TIO

N-F

RO

M-I

DE

AL

SC

OR

ES

FOR

9 C

HE

ESE

S PA

IRE

D W

ITH

18

WIN

ES

(

n

=

9)

Win

ety

peW

ine

no

Win

eL

a pr

ovid

ence

d

cent

Oka

Blu

e B

ened

ictin

Prov

olon

eA

ged

3 ye

arC

hedd

ar

Mig

nero

n de

C

harl

evoi

x

Boe

renk

ass

Cam

embe

rtSw

iss

Med

ium

Che

ddar

Mea

n(

n

=

81)

acro

ss a

llch

eese

s

Whi

te1

Pino

t G

ris

50

61

62

68

76

65

55

75

77

654

Whi

te2

Rie

slin

g5

05

15

56

97

35

84

66

57

96

07W

hite

3Sa

uvig

non

blan

c4

04

45

16

86

15

76

46

77

25

82W

hite

4C

hard

onna

y (o

aked

)5

87

08

57

56

36

53

77

77

26

69W

hite

5G

ewur

ztra

min

er5

86

05

67

47

45

45

89

47

96

74W

hite

6C

hard

onna

y (u

noak

ed)

60

76

82

79

76

70

57

72

82

727

Red

7Pi

not

noir

(oa

ked)

52

59

58

65

77

71

68

65

67

647

Red

8R

ed M

erita

ge5

66

85

48

48

77

06

98

97

57

24R

ed9

Foch

(oa

ked)

48

65

80

83

90

76

70

87

86

761

Red

10Sy

rah

51

65

64

73

84

63

71

86

71

698

Red

11Pi

not

noir

(lig

ht)

40

40

57

52

67

66

57

62

68

566

Red

12M

erlo

t5

25

86

77

69

28

66

38

58

27

34Sp

ecia

lty13

Spar

klin

g6

35

45

86

57

06

15

37

67

36

37Sp

ecia

lty14

Blu

sh4

74

74

06

06

26

46

36

96

85

78Sp

ecia

lty15

LH

Gew

urzt

ram

iner

80

70

96

78

86

75

70

75

83

792

Spec

ialty

16R

iesl

ing

Icew

ine

79

73

87

83

97

82

83

90

100

860

Spec

ialty

17Pi

not

noir

Ice

win

e7

78

08

38

77

27

88

19

29

78

30Sp

ecia

lty18

Port

typ

e8

37

88

010

29

410

110

49

18

69

10St

anda

rd D

evia

tion

26

24

31

24

26

25

31

27

22

naL

SD1

891

752

241

751

961

852

19ns

nsna

P

pound

000

040

004

9E-0

40

003

002

000

60

0006

nsns

naR

ange

40ndash

79

40ndash

78

40ndash

87

52ndash

102

61ndash

97

54ndash

101

46ndash

104

65ndash

94

67ndash

10

WINE AND CHEESE PAIRS 253

wines In each of the wine categories there were wines where the cheesedominated and others where the wine dominated

The average standard deviation for each cheese across all wines rangedfrom 31 to 22 (Table 3) This relatively high deviation indicated the judgeswere not in agreement on their evaluation of the match For a trained descrip-tive panel using a 12-cm scale the standard deviation is generally 1ndash15Figure 2 showed the judge variation in assessing the match of Provolonecheese for three of the wines Although the mean score was not significantlydifferent from an ideal match of six the wide range of judgesrsquo assessmentswas evident For each of the three wines some of the judges thought thecheese dominated the wine while others thought the wine dominated thecheese Figure 3 showed that although there was some variation in the judgeassessment of how the medium Cheddar paired with the various wines therewas better agreement that the wine tended to dominate the cheese The dif-ference in variability is most likely due to the cheese with Provolone beinga stronger cheese and more difficult to match

It was very important to recognize the extent of the judge variation incheese and wine assessments Although all of the judges were individually

FIG 1 MEAN DEVIATION-FROM-IDEAL FOR BLUE CHEESE PAIRED WITH 18 WINES (

n

=

9)

ndash2

ndash15

ndash1

ndash05

0

50

1

51

2

52

Idea

l mat

ch

1

2

3

5

4 0196

21

11

87

41

7161 8151

31

winesetihWGristoniP1gnilseiR2

cnalbnongivuaS3)dekao(yannodrahC4

renimartzruweG5)dekaonu(yannodrahC6

Cheese dominates

Wine dominates

winesdeR)dekao(riontoniP7

egatireM8hcoF9haryS01

)thgil(riontoniP11tolreM21

winesytlaicepSeniwgnilkrapS31

hsulB-esoR41HLrenimartzruweG51

eniwecIgnilseiR61eniwecIriontoniP71

epyttroP81

254 M KING and M CLIFF

FIG 2 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING PROVOLONE CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

renimartzruweG

)thgil(riontoniPeniwgnilkrapS

STD = 315

STD = 361

STD = 417

Cheese dominates

Wine dominates

FIG 3 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING MEDIUM CHEDDAR CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

)thgil( rion toniPrenimartzruweG eniw gnilkrapS

STD = 266 STD = 247STD = 269

Wine dominates

Cheese dominates

WINE AND CHEESE PAIRS 255

experienced in food and wine pairing as a group their results under experi-mental conditions were not always consistent ie they did not always agreeTherefore the results obtained in this study are only broad recommendationsfor possible cheese and wine pairings In the light of the range of assessmentsobtained by knowledgeable judges it would be important to empower con-sumers to have the confidence to make their own assessment of how muchthey enjoy the cheese and wine pair

In conclusion white wines were easier to pair with a broader range ofcheeses than red or specialty wines The stronger more flavorful cheeses weremore difficult to pair with the wines but were more likely to be a good matchfor the late harvest and ice wines than other milder cheeses The results ofthis study were determined with a single tasting protocol and the wine andcheese were tasted without other foods The addition of other foods or agreater time period between the assessments would change the perception ofthe pairing Finally the variation in judge assessment indicates a high degreeof personal expectation or preference associated with the suitability of wineand cheese pairing Round-table discussion and consensus developmentmight produce ldquotrainedrdquo judges with less variation in the assessments althoughthis might be difficult given the ldquorespectedrdquo opinions of acknowledgedexperts Individuals should be encouraged to experiment to determine theirown preferences

ACKNOWLEDGMENTS

The authors would like to thank Reg Hendrickson of the Dairy Farmersof Canada for donating the cheese Christa-Lee Bond of the Okanagan WineFestival Society for allowing this research to be conducted as part of theIcewine Festival events Kimberly Dever for her technical assistance and theBC Winemakers and restaurateurs who participated in the study

REFERENCES

BALDY MW 1995

The University Wine Course

p 426 Wine AppreciationGuild San Francisco CA

BARTOSHUK LM 2000 Comparing sensory experiences across individu-als Recent psychophysical advances illuminate genetic variation in tasteperception Chem Senses

25

447ndash460COCHRAN WG amp COX GM 1957

Experimental Designs

p 611 JohnWiley amp Sons New York

256 M KING and M CLIFF

EampJ GALLO WINERY 2002 Gallo of sonoma wine and cheese pairingguide ABCrsquos of wine and cheese httpwwwwinewebcentralcomawineandcheeseabchtm (accessed October 24 2003)

JACKSON RS 2002

Wine Tasting A Professional Handbook

pp 261ndash262Academic Press San Diego CA

MOESSNER M Inniskillinrsquos guide to wine and cheese pairings httpwwwwineandleisurecomcheesepairinghtml (accessed October 242003)

NOBLE AC 1996 Tastendasharoma interactions Trends Food Sci Technol

7

439ndash444

PHILLIPS R 2000

A Short History of Wine

Penguin Books Londonhttpwwwintowinecomshop1282-0excerpthtml (accessed October24 2003)

SHEPHERD R SMITH K and FARLEIGH CA 1989 The relationshipbetween intensity hedonic and relative-to-ideal ratings Food QualPrefer

1

75ndash80SMITH J 1995

Cheesemaking in Scotland ndash A History

Scottish DairyAssociation Scotland httpwwwefrhwacukSDAbook1html(accessed October 24 2003)

Page 3: EVALUATION OF IDEAL WINE AND CHEESE PAIRS USING A DEVIATION-FROM-IDEAL SCALE WITH FOOD AND WINE EXPERTS

WINE AND CHEESE PAIRS 247

professionals can establish ldquoidealrdquo pairings for still wine However someinteresting work has been done by Korbel Champagne Cellars in the 1980s todetermine how well ldquochampagnerdquo goes with nongourmet foods resulting ina Champagne Compatibility Index (Baldy 1995) but no such index is avail-able for still wine and cheese This study was designed to determine if sucha group of experts can determine under controlled experimental conditionsthe suitability of a larger number of wine and cheese pairs

MATERIALS AND METHODS

Judges

Twenty-seven judges 8 restaurateurs and 19 wine-industry personnel(winemakers and Vintners Quality Alliance [VQA] judges) were recruited forthe evaluation All had extensive wine and food-matching experience Judgeswere given a brief training session on the technical aspects of cheese tastingand the criteria to use for evaluating the ldquoidealrdquo cheese and wine pair Theldquoideal pairrdquo was defined as a combination of wine and cheese where neitherthe wine nor cheese dominated and the pair together was better than eachindividually

Cheeses

Nine award-winning Canadian cheeses were provided by the Dairy Farm-ers of Canada They were selected to cover a broad spectrum of artisan cheesesthat are available in the Canadian marketplace Hard cheeses were cut into30-g (1 oz) cubes and soft cheeses were cut into 30-g wedges Rind-ripenedcheeses were cut so that a section of rind was left on the sample Cheeseswere randomly assigned an alphabetic code (AndashI) and placed in covered 60-mL plastic cups and served at room temperature A list of cheeses and theircharacteristics are presented in Table 1

Wines

Eighteen commercial British Columbia (BC) VQA wines were selectedfrom the Okanagan Wine Festival wine library located at the Pacific Agri-FoodResearch Center (PARC) Summerland BC Wines were selected by VQApersonnel who had extensive knowledge of the range of character and qualityassociated with BC wines six white six red and six specialty wines werechosen to be representative of their type All wines were labeled with three-digitrandom numbers and served in random order Table 2 lists the varietals winesand their known characteristics as summarized by a panel of VQA judges

248 M KING and M CLIFF

TAB

LE

1

DE

SCR

IPT

ION

OF

NIN

E C

HE

ESE

S U

SED

IN

ST

UD

Y

Nam

eC

ateg

ory

Tast

eA

rom

aTe

xtur

e

Blu

e B

ened

ictin

Sem

isof

t(i

nter

ior

ripe

ned)

Piqu

ant

to s

tron

g an

d sa

ltySl

ight

ly p

unge

ntSm

ooth

cre

amy

use

sldquop

enic

illiu

m r

oque

fort

irdquocu

lture

Boe

renk

aas

Firm

Shar

pPu

ngen

tFi

rm a

nd s

light

ly g

ranu

lar

Cam

embe

rtSo

ftC

ream

y b

utte

ry f

unga

l nu

tty t

angy

as

it ri

pens

But

tery

nut

tySm

ooth

cre

amy

whi

tebl

oom

y ri

ndC

hedd

ar ndash

med

ium

Firm

But

tery

lac

ticM

ilky

mild

Smoo

th s

omew

hat

spri

ngy

Che

ddar

whi

te 3

yea

rsFi

rmA

cidu

lous

sal

ty l

actic

cul

ture

Milk

y s

light

ly f

ruity

Smoo

th t

o gr

ainy

firm

ho

mog

eneo

us n

o op

enin

gsL

e M

igne

ron

de

Cha

rlev

oix

Sem

isof

t(s

urfa

ce r

ipen

ed)

Fully

rip

e nu

tty fl

avor

but

ter

with

a t

ouch

of

acid

ityL

ight

not

es o

f cr

eam

an

d a

hint

of

yogu

rtW

ashe

d ri

nd s

oft

supp

le

slig

htly

ora

nge

surf

ace

rind

an

d so

met

imes

slig

htly

stic

kyO

ka ndash

Cla

ssiq

ueSe

mis

oft

(sur

face

rip

ened

)C

ream

y b

utte

ry n

utty

slig

ht

flora

l sh

arpe

r ta

ste

due

to r

aw m

ilkQ

uite

pro

noun

ced

ndash ag

edSu

pple

ten

der

and

soft

was

hed

rind

Prov

olon

e se

tte f

ette

Fi

rmL

ight

tas

te o

f bu

tter

slig

htly

tar

t an

d ta

ngy

with

ful

l bo

dy w

ith a

geSl

ight

ly m

ilky

fru

ityPa

sta

filat

a fi

brou

s in

teri

or d

ue

to a

ging

gou

rd s

hape

d r

iche

r ye

llow

col

orSw

iss

Firm

Slig

htly

alm

ond

som

e sw

eetn

ess

Plea

sant

nut

ty m

ilky

Uni

form

ly d

istr

ibut

ed r

ound

ldquoe

yes

rdquo so

mew

hat

gritt

y

shin

y s

uppl

e

WINE AND CHEESE PAIRS 249

Experimental Design

Wine and cheese pairings were assigned using a partially balancedincomplete block design The treatment combinations were arranged in aLatin-square formation (Cochran and Cox 1957) In this design each cheeseand wine pair occurred in each tasting position (first second third) the same

TABLE 2DESCRIPTION OF 18 WINES USED IN STUDY

Wine Wine description

Sauvignon Blanc Slight herbaceous hints mixed with light pear and gooseberry fruit and a hint of oak Medium to light body with a spicy vanilla aftertaste

Chardonnay (unoaked) Apple pear and citrus aromas and flavors No oak or malolactic character Full body with firm acid and long fresh fruity aftertaste

Pinot Gris Fruity apple pear aromas and flavors Touch of honey Clean wine with good body and balance and a long smooth finish

Chardonnay Citrus apple fruit on aroma and palate Rich oak character with caramel and butternotes Firm acid with full body and a long aftertaste

Gewurztraminer Floral perfume spicy and lychee aromas Some pear pineapple and tropical characters Clean with good balance Long smooth finish

Riesling Fruity and fresh with apricot peach and apple aromas and flavor Crisp dry well balanced wine with medium body and length of finish

Pinot noir (light) Light clean berry cherry fruit Delicate character with some toasty vanilla oak Soft tannins with a relatively short finish

Pinot noir (oaked) Oaky with a toasty caramel spicy character Jammy cherry and strawberry flavors Soft round finish with a hint of residual sweetness

Merlot Lots of ripe jammy fruit raspberry and cherry on aroma and palate Spicy character A complex wine and toasty caramel vanilla notes

Meritage Toasted oak black pepper and ripe fruit (cherry plum and prune) aroma smokey character Full tannins and good body Long aftertaste

Foch Smokey spicy oaky aromas Some earthy cedar vegetative character on palate Clean with good body and balance Medium long finish

Sparkling Clean fresh fruit Ripe fruit with citrus apricot and pear characters on both aroma and palate Good body and balance with a long smooth finish Nice mousse on palate

Gamay blush Fruity light clean wine Light strawberry and raspberry character with a hint of black pepper Good body balance and long smooth finish

Gewuztraminer (late harvest)

Lychee character with citrus apricot peach and pear fruit character Some perfume floral and honey notes Good late harvest concentration of flavors Sweet with lower acid

Riesling Icewine Concentrated apple and citrus fruit flavors with a touch of honey High residual sugar with acid to balance with viscous mouth feel

Pinot noir Icewine Smokey baked-fruit character on nose Some light raspberry and strawberry on palate Concentrated fruit sweetness and acid with a full body and long finish

Port type Dried fruit plum and oak on aroma and palate Slightly hot with some residual sugar and sweet lingering aftertaste

250 M KING and M CLIFF

number of times allowing experimental error from ldquocarryoverrdquo or ldquoresidualrdquoto be balanced across all samples

The wines were divided into three blocks of six white red and spe-cialty wines respectively The white and red wines were tasted in a randomorder while the specialty wines had an assigned order of sparkling winefirst and Icewines last This was thought to be necessary because Icewinesare very sweet and intense with a potential for a strong carryover effect thatcould interfere with a judgersquos ability to assess the other cheese and winepairs

Using the incomplete block design each judge tasted three cheeses withsix of the wines The cheeses were systematically paired with the wines sothat after the three blocks of wines a judge had tasted all nine cheeses andeach cheese had been evaluated nine times with each of the wines Judgeswere asked to evaluate the suitability of the pairing for each cheese with allof the wines in the block before moving on to the next cheese

Tasting Process

Judges were instructed to take a small bite of cheese sip the wine andevaluate the appropriateness of the pair The judges evaluated the match usinga variation of the ldquojust right scalerdquo (Shepherd

et al

1989) Judges scored thedeviation from an ldquoideal matchrdquo on a 12-cm line scale where the ideal matchwas the midpoint of the line (6 cm) If the judges thought the cheese domi-nated the wine they marked the line scale to the left of the midpoint whereasif they thought the wine dominated the cheese the judges marked the line scaleto the right of the midpoint The scale was labeled with ldquocheese dominatesexcessivelyrdquo ldquocheese dominates moderatelyrdquo and ldquocheese dominates slightlyrdquoat 0 2 and 4 cm The other half of the scale was labeled with ldquowine dominatesslightlyrdquo ldquowine dominates moderatelyrdquo and ldquowine dominates excessivelyrdquo at8 10 12 cm respectively

Judges were given sparkling water bread and melba toast in order toclear their palates between evaluations They were also instructed to take ashort break between cheeses After the tasting the scorecards were collectedand the judgesrsquo scores for each wine and cheese pairing were measured(0ndash12) for data analysis and interpretation

Statistical Analysis

A single factor analysis of variance (ANOVA) (Microsoft Excel Seat-tle WA) was used to determine if there were significant differences in thesuitability of the match between each cheese with each wine A Fisherrsquosleast significant difference (LSD) was calculated to determine if the ldquoidealrdquoscores were significantly different from each other The average standard

WINE AND CHEESE PAIRS 251

deviation from the mean was calculated for each cheese and wine combina-tion This was then averaged across all wines to give an indication of judgevariability

RESULTS AND DISCUSSION

Results of the ANOVA are shown in Table 3 As might be expected mostcheeses showed significant differences (

P

lt

0001) in the suitability of theirpairing with the 18 wines More surprisingly two cheeses (Swiss and mediumCheddar) showed no differences in their suitability with any of the winesTable 3 tabulates the mean score for each of the wine and cheese pairs Anideal pairing would have a mean score of six a mean score greater than sixwould indicate that the wine dominated the cheese a score less than six wouldindicate that the cheese dominated

Overall the relatively small deviations-from-ideal confirmed that wineand cheese are compatible food pairings In the three categories of wine whitewines tended to be a better match with these cheeses This was indicated byan average deviation-from-ideal of 085 compared to a deviation of 12 and17 for the red and specialty wines respectively The most versatile white winewas the Riesling followed by the Sauvignon blanc and the Pinot Gris asreflected by the smallest deviations The Pinot noir wine was the versatile redwine in that it paired best with the most number of cheeses The specialtysparkling and blush wines also paired well with many of the cheeses whilethe late harvest ice and port wines were the most difficult to pair with manyof the cheeses The Blue Benedictine Oka and Provolone cheeses were themost suitable match with these wines thus supporting the general rule thatstronger flavored cheeses tend to be better matched with stronger flavoredwines In general in this study the specialty wines (late harvest icewine porttype) dominated the cheeses However the addition of other foods or a dif-ference in the tasting method might change the perception of the pairing andlikely make these wines more compatible

Four cheeses (aged white Cheddar Migerone de Charlevoix Swiss andmedium Cheddar) had mean scores greater than six indicating that the wineswere more strongly flavored than the cheese Although the mean scores forthe 3-year old Cheddar were greater than six the means for the white winesthe Pinot noir and Merlot wines were not significantly different than six Thisindicated that there would be a range of wines that would be a suitable matchwith this cheese In general the stronger flavored cheeses were the mostvariable in their matches with the wines This indicated that more care wouldbe required in selecting wines to pair with these cheeses Figure 1 shows thedeviation-from-ideal for the match of the Benedictine Blue with each of the

252 M KING and M CLIFF

TAB

LE

3

ME

AN

DE

VIA

TIO

N-F

RO

M-I

DE

AL

SC

OR

ES

FOR

9 C

HE

ESE

S PA

IRE

D W

ITH

18

WIN

ES

(

n

=

9)

Win

ety

peW

ine

no

Win

eL

a pr

ovid

ence

d

cent

Oka

Blu

e B

ened

ictin

Prov

olon

eA

ged

3 ye

arC

hedd

ar

Mig

nero

n de

C

harl

evoi

x

Boe

renk

ass

Cam

embe

rtSw

iss

Med

ium

Che

ddar

Mea

n(

n

=

81)

acro

ss a

llch

eese

s

Whi

te1

Pino

t G

ris

50

61

62

68

76

65

55

75

77

654

Whi

te2

Rie

slin

g5

05

15

56

97

35

84

66

57

96

07W

hite

3Sa

uvig

non

blan

c4

04

45

16

86

15

76

46

77

25

82W

hite

4C

hard

onna

y (o

aked

)5

87

08

57

56

36

53

77

77

26

69W

hite

5G

ewur

ztra

min

er5

86

05

67

47

45

45

89

47

96

74W

hite

6C

hard

onna

y (u

noak

ed)

60

76

82

79

76

70

57

72

82

727

Red

7Pi

not

noir

(oa

ked)

52

59

58

65

77

71

68

65

67

647

Red

8R

ed M

erita

ge5

66

85

48

48

77

06

98

97

57

24R

ed9

Foch

(oa

ked)

48

65

80

83

90

76

70

87

86

761

Red

10Sy

rah

51

65

64

73

84

63

71

86

71

698

Red

11Pi

not

noir

(lig

ht)

40

40

57

52

67

66

57

62

68

566

Red

12M

erlo

t5

25

86

77

69

28

66

38

58

27

34Sp

ecia

lty13

Spar

klin

g6

35

45

86

57

06

15

37

67

36

37Sp

ecia

lty14

Blu

sh4

74

74

06

06

26

46

36

96

85

78Sp

ecia

lty15

LH

Gew

urzt

ram

iner

80

70

96

78

86

75

70

75

83

792

Spec

ialty

16R

iesl

ing

Icew

ine

79

73

87

83

97

82

83

90

100

860

Spec

ialty

17Pi

not

noir

Ice

win

e7

78

08

38

77

27

88

19

29

78

30Sp

ecia

lty18

Port

typ

e8

37

88

010

29

410

110

49

18

69

10St

anda

rd D

evia

tion

26

24

31

24

26

25

31

27

22

naL

SD1

891

752

241

751

961

852

19ns

nsna

P

pound

000

040

004

9E-0

40

003

002

000

60

0006

nsns

naR

ange

40ndash

79

40ndash

78

40ndash

87

52ndash

102

61ndash

97

54ndash

101

46ndash

104

65ndash

94

67ndash

10

WINE AND CHEESE PAIRS 253

wines In each of the wine categories there were wines where the cheesedominated and others where the wine dominated

The average standard deviation for each cheese across all wines rangedfrom 31 to 22 (Table 3) This relatively high deviation indicated the judgeswere not in agreement on their evaluation of the match For a trained descrip-tive panel using a 12-cm scale the standard deviation is generally 1ndash15Figure 2 showed the judge variation in assessing the match of Provolonecheese for three of the wines Although the mean score was not significantlydifferent from an ideal match of six the wide range of judgesrsquo assessmentswas evident For each of the three wines some of the judges thought thecheese dominated the wine while others thought the wine dominated thecheese Figure 3 showed that although there was some variation in the judgeassessment of how the medium Cheddar paired with the various wines therewas better agreement that the wine tended to dominate the cheese The dif-ference in variability is most likely due to the cheese with Provolone beinga stronger cheese and more difficult to match

It was very important to recognize the extent of the judge variation incheese and wine assessments Although all of the judges were individually

FIG 1 MEAN DEVIATION-FROM-IDEAL FOR BLUE CHEESE PAIRED WITH 18 WINES (

n

=

9)

ndash2

ndash15

ndash1

ndash05

0

50

1

51

2

52

Idea

l mat

ch

1

2

3

5

4 0196

21

11

87

41

7161 8151

31

winesetihWGristoniP1gnilseiR2

cnalbnongivuaS3)dekao(yannodrahC4

renimartzruweG5)dekaonu(yannodrahC6

Cheese dominates

Wine dominates

winesdeR)dekao(riontoniP7

egatireM8hcoF9haryS01

)thgil(riontoniP11tolreM21

winesytlaicepSeniwgnilkrapS31

hsulB-esoR41HLrenimartzruweG51

eniwecIgnilseiR61eniwecIriontoniP71

epyttroP81

254 M KING and M CLIFF

FIG 2 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING PROVOLONE CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

renimartzruweG

)thgil(riontoniPeniwgnilkrapS

STD = 315

STD = 361

STD = 417

Cheese dominates

Wine dominates

FIG 3 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING MEDIUM CHEDDAR CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

)thgil( rion toniPrenimartzruweG eniw gnilkrapS

STD = 266 STD = 247STD = 269

Wine dominates

Cheese dominates

WINE AND CHEESE PAIRS 255

experienced in food and wine pairing as a group their results under experi-mental conditions were not always consistent ie they did not always agreeTherefore the results obtained in this study are only broad recommendationsfor possible cheese and wine pairings In the light of the range of assessmentsobtained by knowledgeable judges it would be important to empower con-sumers to have the confidence to make their own assessment of how muchthey enjoy the cheese and wine pair

In conclusion white wines were easier to pair with a broader range ofcheeses than red or specialty wines The stronger more flavorful cheeses weremore difficult to pair with the wines but were more likely to be a good matchfor the late harvest and ice wines than other milder cheeses The results ofthis study were determined with a single tasting protocol and the wine andcheese were tasted without other foods The addition of other foods or agreater time period between the assessments would change the perception ofthe pairing Finally the variation in judge assessment indicates a high degreeof personal expectation or preference associated with the suitability of wineand cheese pairing Round-table discussion and consensus developmentmight produce ldquotrainedrdquo judges with less variation in the assessments althoughthis might be difficult given the ldquorespectedrdquo opinions of acknowledgedexperts Individuals should be encouraged to experiment to determine theirown preferences

ACKNOWLEDGMENTS

The authors would like to thank Reg Hendrickson of the Dairy Farmersof Canada for donating the cheese Christa-Lee Bond of the Okanagan WineFestival Society for allowing this research to be conducted as part of theIcewine Festival events Kimberly Dever for her technical assistance and theBC Winemakers and restaurateurs who participated in the study

REFERENCES

BALDY MW 1995

The University Wine Course

p 426 Wine AppreciationGuild San Francisco CA

BARTOSHUK LM 2000 Comparing sensory experiences across individu-als Recent psychophysical advances illuminate genetic variation in tasteperception Chem Senses

25

447ndash460COCHRAN WG amp COX GM 1957

Experimental Designs

p 611 JohnWiley amp Sons New York

256 M KING and M CLIFF

EampJ GALLO WINERY 2002 Gallo of sonoma wine and cheese pairingguide ABCrsquos of wine and cheese httpwwwwinewebcentralcomawineandcheeseabchtm (accessed October 24 2003)

JACKSON RS 2002

Wine Tasting A Professional Handbook

pp 261ndash262Academic Press San Diego CA

MOESSNER M Inniskillinrsquos guide to wine and cheese pairings httpwwwwineandleisurecomcheesepairinghtml (accessed October 242003)

NOBLE AC 1996 Tastendasharoma interactions Trends Food Sci Technol

7

439ndash444

PHILLIPS R 2000

A Short History of Wine

Penguin Books Londonhttpwwwintowinecomshop1282-0excerpthtml (accessed October24 2003)

SHEPHERD R SMITH K and FARLEIGH CA 1989 The relationshipbetween intensity hedonic and relative-to-ideal ratings Food QualPrefer

1

75ndash80SMITH J 1995

Cheesemaking in Scotland ndash A History

Scottish DairyAssociation Scotland httpwwwefrhwacukSDAbook1html(accessed October 24 2003)

Page 4: EVALUATION OF IDEAL WINE AND CHEESE PAIRS USING A DEVIATION-FROM-IDEAL SCALE WITH FOOD AND WINE EXPERTS

248 M KING and M CLIFF

TAB

LE

1

DE

SCR

IPT

ION

OF

NIN

E C

HE

ESE

S U

SED

IN

ST

UD

Y

Nam

eC

ateg

ory

Tast

eA

rom

aTe

xtur

e

Blu

e B

ened

ictin

Sem

isof

t(i

nter

ior

ripe

ned)

Piqu

ant

to s

tron

g an

d sa

ltySl

ight

ly p

unge

ntSm

ooth

cre

amy

use

sldquop

enic

illiu

m r

oque

fort

irdquocu

lture

Boe

renk

aas

Firm

Shar

pPu

ngen

tFi

rm a

nd s

light

ly g

ranu

lar

Cam

embe

rtSo

ftC

ream

y b

utte

ry f

unga

l nu

tty t

angy

as

it ri

pens

But

tery

nut

tySm

ooth

cre

amy

whi

tebl

oom

y ri

ndC

hedd

ar ndash

med

ium

Firm

But

tery

lac

ticM

ilky

mild

Smoo

th s

omew

hat

spri

ngy

Che

ddar

whi

te 3

yea

rsFi

rmA

cidu

lous

sal

ty l

actic

cul

ture

Milk

y s

light

ly f

ruity

Smoo

th t

o gr

ainy

firm

ho

mog

eneo

us n

o op

enin

gsL

e M

igne

ron

de

Cha

rlev

oix

Sem

isof

t(s

urfa

ce r

ipen

ed)

Fully

rip

e nu

tty fl

avor

but

ter

with

a t

ouch

of

acid

ityL

ight

not

es o

f cr

eam

an

d a

hint

of

yogu

rtW

ashe

d ri

nd s

oft

supp

le

slig

htly

ora

nge

surf

ace

rind

an

d so

met

imes

slig

htly

stic

kyO

ka ndash

Cla

ssiq

ueSe

mis

oft

(sur

face

rip

ened

)C

ream

y b

utte

ry n

utty

slig

ht

flora

l sh

arpe

r ta

ste

due

to r

aw m

ilkQ

uite

pro

noun

ced

ndash ag

edSu

pple

ten

der

and

soft

was

hed

rind

Prov

olon

e se

tte f

ette

Fi

rmL

ight

tas

te o

f bu

tter

slig

htly

tar

t an

d ta

ngy

with

ful

l bo

dy w

ith a

geSl

ight

ly m

ilky

fru

ityPa

sta

filat

a fi

brou

s in

teri

or d

ue

to a

ging

gou

rd s

hape

d r

iche

r ye

llow

col

orSw

iss

Firm

Slig

htly

alm

ond

som

e sw

eetn

ess

Plea

sant

nut

ty m

ilky

Uni

form

ly d

istr

ibut

ed r

ound

ldquoe

yes

rdquo so

mew

hat

gritt

y

shin

y s

uppl

e

WINE AND CHEESE PAIRS 249

Experimental Design

Wine and cheese pairings were assigned using a partially balancedincomplete block design The treatment combinations were arranged in aLatin-square formation (Cochran and Cox 1957) In this design each cheeseand wine pair occurred in each tasting position (first second third) the same

TABLE 2DESCRIPTION OF 18 WINES USED IN STUDY

Wine Wine description

Sauvignon Blanc Slight herbaceous hints mixed with light pear and gooseberry fruit and a hint of oak Medium to light body with a spicy vanilla aftertaste

Chardonnay (unoaked) Apple pear and citrus aromas and flavors No oak or malolactic character Full body with firm acid and long fresh fruity aftertaste

Pinot Gris Fruity apple pear aromas and flavors Touch of honey Clean wine with good body and balance and a long smooth finish

Chardonnay Citrus apple fruit on aroma and palate Rich oak character with caramel and butternotes Firm acid with full body and a long aftertaste

Gewurztraminer Floral perfume spicy and lychee aromas Some pear pineapple and tropical characters Clean with good balance Long smooth finish

Riesling Fruity and fresh with apricot peach and apple aromas and flavor Crisp dry well balanced wine with medium body and length of finish

Pinot noir (light) Light clean berry cherry fruit Delicate character with some toasty vanilla oak Soft tannins with a relatively short finish

Pinot noir (oaked) Oaky with a toasty caramel spicy character Jammy cherry and strawberry flavors Soft round finish with a hint of residual sweetness

Merlot Lots of ripe jammy fruit raspberry and cherry on aroma and palate Spicy character A complex wine and toasty caramel vanilla notes

Meritage Toasted oak black pepper and ripe fruit (cherry plum and prune) aroma smokey character Full tannins and good body Long aftertaste

Foch Smokey spicy oaky aromas Some earthy cedar vegetative character on palate Clean with good body and balance Medium long finish

Sparkling Clean fresh fruit Ripe fruit with citrus apricot and pear characters on both aroma and palate Good body and balance with a long smooth finish Nice mousse on palate

Gamay blush Fruity light clean wine Light strawberry and raspberry character with a hint of black pepper Good body balance and long smooth finish

Gewuztraminer (late harvest)

Lychee character with citrus apricot peach and pear fruit character Some perfume floral and honey notes Good late harvest concentration of flavors Sweet with lower acid

Riesling Icewine Concentrated apple and citrus fruit flavors with a touch of honey High residual sugar with acid to balance with viscous mouth feel

Pinot noir Icewine Smokey baked-fruit character on nose Some light raspberry and strawberry on palate Concentrated fruit sweetness and acid with a full body and long finish

Port type Dried fruit plum and oak on aroma and palate Slightly hot with some residual sugar and sweet lingering aftertaste

250 M KING and M CLIFF

number of times allowing experimental error from ldquocarryoverrdquo or ldquoresidualrdquoto be balanced across all samples

The wines were divided into three blocks of six white red and spe-cialty wines respectively The white and red wines were tasted in a randomorder while the specialty wines had an assigned order of sparkling winefirst and Icewines last This was thought to be necessary because Icewinesare very sweet and intense with a potential for a strong carryover effect thatcould interfere with a judgersquos ability to assess the other cheese and winepairs

Using the incomplete block design each judge tasted three cheeses withsix of the wines The cheeses were systematically paired with the wines sothat after the three blocks of wines a judge had tasted all nine cheeses andeach cheese had been evaluated nine times with each of the wines Judgeswere asked to evaluate the suitability of the pairing for each cheese with allof the wines in the block before moving on to the next cheese

Tasting Process

Judges were instructed to take a small bite of cheese sip the wine andevaluate the appropriateness of the pair The judges evaluated the match usinga variation of the ldquojust right scalerdquo (Shepherd

et al

1989) Judges scored thedeviation from an ldquoideal matchrdquo on a 12-cm line scale where the ideal matchwas the midpoint of the line (6 cm) If the judges thought the cheese domi-nated the wine they marked the line scale to the left of the midpoint whereasif they thought the wine dominated the cheese the judges marked the line scaleto the right of the midpoint The scale was labeled with ldquocheese dominatesexcessivelyrdquo ldquocheese dominates moderatelyrdquo and ldquocheese dominates slightlyrdquoat 0 2 and 4 cm The other half of the scale was labeled with ldquowine dominatesslightlyrdquo ldquowine dominates moderatelyrdquo and ldquowine dominates excessivelyrdquo at8 10 12 cm respectively

Judges were given sparkling water bread and melba toast in order toclear their palates between evaluations They were also instructed to take ashort break between cheeses After the tasting the scorecards were collectedand the judgesrsquo scores for each wine and cheese pairing were measured(0ndash12) for data analysis and interpretation

Statistical Analysis

A single factor analysis of variance (ANOVA) (Microsoft Excel Seat-tle WA) was used to determine if there were significant differences in thesuitability of the match between each cheese with each wine A Fisherrsquosleast significant difference (LSD) was calculated to determine if the ldquoidealrdquoscores were significantly different from each other The average standard

WINE AND CHEESE PAIRS 251

deviation from the mean was calculated for each cheese and wine combina-tion This was then averaged across all wines to give an indication of judgevariability

RESULTS AND DISCUSSION

Results of the ANOVA are shown in Table 3 As might be expected mostcheeses showed significant differences (

P

lt

0001) in the suitability of theirpairing with the 18 wines More surprisingly two cheeses (Swiss and mediumCheddar) showed no differences in their suitability with any of the winesTable 3 tabulates the mean score for each of the wine and cheese pairs Anideal pairing would have a mean score of six a mean score greater than sixwould indicate that the wine dominated the cheese a score less than six wouldindicate that the cheese dominated

Overall the relatively small deviations-from-ideal confirmed that wineand cheese are compatible food pairings In the three categories of wine whitewines tended to be a better match with these cheeses This was indicated byan average deviation-from-ideal of 085 compared to a deviation of 12 and17 for the red and specialty wines respectively The most versatile white winewas the Riesling followed by the Sauvignon blanc and the Pinot Gris asreflected by the smallest deviations The Pinot noir wine was the versatile redwine in that it paired best with the most number of cheeses The specialtysparkling and blush wines also paired well with many of the cheeses whilethe late harvest ice and port wines were the most difficult to pair with manyof the cheeses The Blue Benedictine Oka and Provolone cheeses were themost suitable match with these wines thus supporting the general rule thatstronger flavored cheeses tend to be better matched with stronger flavoredwines In general in this study the specialty wines (late harvest icewine porttype) dominated the cheeses However the addition of other foods or a dif-ference in the tasting method might change the perception of the pairing andlikely make these wines more compatible

Four cheeses (aged white Cheddar Migerone de Charlevoix Swiss andmedium Cheddar) had mean scores greater than six indicating that the wineswere more strongly flavored than the cheese Although the mean scores forthe 3-year old Cheddar were greater than six the means for the white winesthe Pinot noir and Merlot wines were not significantly different than six Thisindicated that there would be a range of wines that would be a suitable matchwith this cheese In general the stronger flavored cheeses were the mostvariable in their matches with the wines This indicated that more care wouldbe required in selecting wines to pair with these cheeses Figure 1 shows thedeviation-from-ideal for the match of the Benedictine Blue with each of the

252 M KING and M CLIFF

TAB

LE

3

ME

AN

DE

VIA

TIO

N-F

RO

M-I

DE

AL

SC

OR

ES

FOR

9 C

HE

ESE

S PA

IRE

D W

ITH

18

WIN

ES

(

n

=

9)

Win

ety

peW

ine

no

Win

eL

a pr

ovid

ence

d

cent

Oka

Blu

e B

ened

ictin

Prov

olon

eA

ged

3 ye

arC

hedd

ar

Mig

nero

n de

C

harl

evoi

x

Boe

renk

ass

Cam

embe

rtSw

iss

Med

ium

Che

ddar

Mea

n(

n

=

81)

acro

ss a

llch

eese

s

Whi

te1

Pino

t G

ris

50

61

62

68

76

65

55

75

77

654

Whi

te2

Rie

slin

g5

05

15

56

97

35

84

66

57

96

07W

hite

3Sa

uvig

non

blan

c4

04

45

16

86

15

76

46

77

25

82W

hite

4C

hard

onna

y (o

aked

)5

87

08

57

56

36

53

77

77

26

69W

hite

5G

ewur

ztra

min

er5

86

05

67

47

45

45

89

47

96

74W

hite

6C

hard

onna

y (u

noak

ed)

60

76

82

79

76

70

57

72

82

727

Red

7Pi

not

noir

(oa

ked)

52

59

58

65

77

71

68

65

67

647

Red

8R

ed M

erita

ge5

66

85

48

48

77

06

98

97

57

24R

ed9

Foch

(oa

ked)

48

65

80

83

90

76

70

87

86

761

Red

10Sy

rah

51

65

64

73

84

63

71

86

71

698

Red

11Pi

not

noir

(lig

ht)

40

40

57

52

67

66

57

62

68

566

Red

12M

erlo

t5

25

86

77

69

28

66

38

58

27

34Sp

ecia

lty13

Spar

klin

g6

35

45

86

57

06

15

37

67

36

37Sp

ecia

lty14

Blu

sh4

74

74

06

06

26

46

36

96

85

78Sp

ecia

lty15

LH

Gew

urzt

ram

iner

80

70

96

78

86

75

70

75

83

792

Spec

ialty

16R

iesl

ing

Icew

ine

79

73

87

83

97

82

83

90

100

860

Spec

ialty

17Pi

not

noir

Ice

win

e7

78

08

38

77

27

88

19

29

78

30Sp

ecia

lty18

Port

typ

e8

37

88

010

29

410

110

49

18

69

10St

anda

rd D

evia

tion

26

24

31

24

26

25

31

27

22

naL

SD1

891

752

241

751

961

852

19ns

nsna

P

pound

000

040

004

9E-0

40

003

002

000

60

0006

nsns

naR

ange

40ndash

79

40ndash

78

40ndash

87

52ndash

102

61ndash

97

54ndash

101

46ndash

104

65ndash

94

67ndash

10

WINE AND CHEESE PAIRS 253

wines In each of the wine categories there were wines where the cheesedominated and others where the wine dominated

The average standard deviation for each cheese across all wines rangedfrom 31 to 22 (Table 3) This relatively high deviation indicated the judgeswere not in agreement on their evaluation of the match For a trained descrip-tive panel using a 12-cm scale the standard deviation is generally 1ndash15Figure 2 showed the judge variation in assessing the match of Provolonecheese for three of the wines Although the mean score was not significantlydifferent from an ideal match of six the wide range of judgesrsquo assessmentswas evident For each of the three wines some of the judges thought thecheese dominated the wine while others thought the wine dominated thecheese Figure 3 showed that although there was some variation in the judgeassessment of how the medium Cheddar paired with the various wines therewas better agreement that the wine tended to dominate the cheese The dif-ference in variability is most likely due to the cheese with Provolone beinga stronger cheese and more difficult to match

It was very important to recognize the extent of the judge variation incheese and wine assessments Although all of the judges were individually

FIG 1 MEAN DEVIATION-FROM-IDEAL FOR BLUE CHEESE PAIRED WITH 18 WINES (

n

=

9)

ndash2

ndash15

ndash1

ndash05

0

50

1

51

2

52

Idea

l mat

ch

1

2

3

5

4 0196

21

11

87

41

7161 8151

31

winesetihWGristoniP1gnilseiR2

cnalbnongivuaS3)dekao(yannodrahC4

renimartzruweG5)dekaonu(yannodrahC6

Cheese dominates

Wine dominates

winesdeR)dekao(riontoniP7

egatireM8hcoF9haryS01

)thgil(riontoniP11tolreM21

winesytlaicepSeniwgnilkrapS31

hsulB-esoR41HLrenimartzruweG51

eniwecIgnilseiR61eniwecIriontoniP71

epyttroP81

254 M KING and M CLIFF

FIG 2 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING PROVOLONE CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

renimartzruweG

)thgil(riontoniPeniwgnilkrapS

STD = 315

STD = 361

STD = 417

Cheese dominates

Wine dominates

FIG 3 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING MEDIUM CHEDDAR CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

)thgil( rion toniPrenimartzruweG eniw gnilkrapS

STD = 266 STD = 247STD = 269

Wine dominates

Cheese dominates

WINE AND CHEESE PAIRS 255

experienced in food and wine pairing as a group their results under experi-mental conditions were not always consistent ie they did not always agreeTherefore the results obtained in this study are only broad recommendationsfor possible cheese and wine pairings In the light of the range of assessmentsobtained by knowledgeable judges it would be important to empower con-sumers to have the confidence to make their own assessment of how muchthey enjoy the cheese and wine pair

In conclusion white wines were easier to pair with a broader range ofcheeses than red or specialty wines The stronger more flavorful cheeses weremore difficult to pair with the wines but were more likely to be a good matchfor the late harvest and ice wines than other milder cheeses The results ofthis study were determined with a single tasting protocol and the wine andcheese were tasted without other foods The addition of other foods or agreater time period between the assessments would change the perception ofthe pairing Finally the variation in judge assessment indicates a high degreeof personal expectation or preference associated with the suitability of wineand cheese pairing Round-table discussion and consensus developmentmight produce ldquotrainedrdquo judges with less variation in the assessments althoughthis might be difficult given the ldquorespectedrdquo opinions of acknowledgedexperts Individuals should be encouraged to experiment to determine theirown preferences

ACKNOWLEDGMENTS

The authors would like to thank Reg Hendrickson of the Dairy Farmersof Canada for donating the cheese Christa-Lee Bond of the Okanagan WineFestival Society for allowing this research to be conducted as part of theIcewine Festival events Kimberly Dever for her technical assistance and theBC Winemakers and restaurateurs who participated in the study

REFERENCES

BALDY MW 1995

The University Wine Course

p 426 Wine AppreciationGuild San Francisco CA

BARTOSHUK LM 2000 Comparing sensory experiences across individu-als Recent psychophysical advances illuminate genetic variation in tasteperception Chem Senses

25

447ndash460COCHRAN WG amp COX GM 1957

Experimental Designs

p 611 JohnWiley amp Sons New York

256 M KING and M CLIFF

EampJ GALLO WINERY 2002 Gallo of sonoma wine and cheese pairingguide ABCrsquos of wine and cheese httpwwwwinewebcentralcomawineandcheeseabchtm (accessed October 24 2003)

JACKSON RS 2002

Wine Tasting A Professional Handbook

pp 261ndash262Academic Press San Diego CA

MOESSNER M Inniskillinrsquos guide to wine and cheese pairings httpwwwwineandleisurecomcheesepairinghtml (accessed October 242003)

NOBLE AC 1996 Tastendasharoma interactions Trends Food Sci Technol

7

439ndash444

PHILLIPS R 2000

A Short History of Wine

Penguin Books Londonhttpwwwintowinecomshop1282-0excerpthtml (accessed October24 2003)

SHEPHERD R SMITH K and FARLEIGH CA 1989 The relationshipbetween intensity hedonic and relative-to-ideal ratings Food QualPrefer

1

75ndash80SMITH J 1995

Cheesemaking in Scotland ndash A History

Scottish DairyAssociation Scotland httpwwwefrhwacukSDAbook1html(accessed October 24 2003)

Page 5: EVALUATION OF IDEAL WINE AND CHEESE PAIRS USING A DEVIATION-FROM-IDEAL SCALE WITH FOOD AND WINE EXPERTS

WINE AND CHEESE PAIRS 249

Experimental Design

Wine and cheese pairings were assigned using a partially balancedincomplete block design The treatment combinations were arranged in aLatin-square formation (Cochran and Cox 1957) In this design each cheeseand wine pair occurred in each tasting position (first second third) the same

TABLE 2DESCRIPTION OF 18 WINES USED IN STUDY

Wine Wine description

Sauvignon Blanc Slight herbaceous hints mixed with light pear and gooseberry fruit and a hint of oak Medium to light body with a spicy vanilla aftertaste

Chardonnay (unoaked) Apple pear and citrus aromas and flavors No oak or malolactic character Full body with firm acid and long fresh fruity aftertaste

Pinot Gris Fruity apple pear aromas and flavors Touch of honey Clean wine with good body and balance and a long smooth finish

Chardonnay Citrus apple fruit on aroma and palate Rich oak character with caramel and butternotes Firm acid with full body and a long aftertaste

Gewurztraminer Floral perfume spicy and lychee aromas Some pear pineapple and tropical characters Clean with good balance Long smooth finish

Riesling Fruity and fresh with apricot peach and apple aromas and flavor Crisp dry well balanced wine with medium body and length of finish

Pinot noir (light) Light clean berry cherry fruit Delicate character with some toasty vanilla oak Soft tannins with a relatively short finish

Pinot noir (oaked) Oaky with a toasty caramel spicy character Jammy cherry and strawberry flavors Soft round finish with a hint of residual sweetness

Merlot Lots of ripe jammy fruit raspberry and cherry on aroma and palate Spicy character A complex wine and toasty caramel vanilla notes

Meritage Toasted oak black pepper and ripe fruit (cherry plum and prune) aroma smokey character Full tannins and good body Long aftertaste

Foch Smokey spicy oaky aromas Some earthy cedar vegetative character on palate Clean with good body and balance Medium long finish

Sparkling Clean fresh fruit Ripe fruit with citrus apricot and pear characters on both aroma and palate Good body and balance with a long smooth finish Nice mousse on palate

Gamay blush Fruity light clean wine Light strawberry and raspberry character with a hint of black pepper Good body balance and long smooth finish

Gewuztraminer (late harvest)

Lychee character with citrus apricot peach and pear fruit character Some perfume floral and honey notes Good late harvest concentration of flavors Sweet with lower acid

Riesling Icewine Concentrated apple and citrus fruit flavors with a touch of honey High residual sugar with acid to balance with viscous mouth feel

Pinot noir Icewine Smokey baked-fruit character on nose Some light raspberry and strawberry on palate Concentrated fruit sweetness and acid with a full body and long finish

Port type Dried fruit plum and oak on aroma and palate Slightly hot with some residual sugar and sweet lingering aftertaste

250 M KING and M CLIFF

number of times allowing experimental error from ldquocarryoverrdquo or ldquoresidualrdquoto be balanced across all samples

The wines were divided into three blocks of six white red and spe-cialty wines respectively The white and red wines were tasted in a randomorder while the specialty wines had an assigned order of sparkling winefirst and Icewines last This was thought to be necessary because Icewinesare very sweet and intense with a potential for a strong carryover effect thatcould interfere with a judgersquos ability to assess the other cheese and winepairs

Using the incomplete block design each judge tasted three cheeses withsix of the wines The cheeses were systematically paired with the wines sothat after the three blocks of wines a judge had tasted all nine cheeses andeach cheese had been evaluated nine times with each of the wines Judgeswere asked to evaluate the suitability of the pairing for each cheese with allof the wines in the block before moving on to the next cheese

Tasting Process

Judges were instructed to take a small bite of cheese sip the wine andevaluate the appropriateness of the pair The judges evaluated the match usinga variation of the ldquojust right scalerdquo (Shepherd

et al

1989) Judges scored thedeviation from an ldquoideal matchrdquo on a 12-cm line scale where the ideal matchwas the midpoint of the line (6 cm) If the judges thought the cheese domi-nated the wine they marked the line scale to the left of the midpoint whereasif they thought the wine dominated the cheese the judges marked the line scaleto the right of the midpoint The scale was labeled with ldquocheese dominatesexcessivelyrdquo ldquocheese dominates moderatelyrdquo and ldquocheese dominates slightlyrdquoat 0 2 and 4 cm The other half of the scale was labeled with ldquowine dominatesslightlyrdquo ldquowine dominates moderatelyrdquo and ldquowine dominates excessivelyrdquo at8 10 12 cm respectively

Judges were given sparkling water bread and melba toast in order toclear their palates between evaluations They were also instructed to take ashort break between cheeses After the tasting the scorecards were collectedand the judgesrsquo scores for each wine and cheese pairing were measured(0ndash12) for data analysis and interpretation

Statistical Analysis

A single factor analysis of variance (ANOVA) (Microsoft Excel Seat-tle WA) was used to determine if there were significant differences in thesuitability of the match between each cheese with each wine A Fisherrsquosleast significant difference (LSD) was calculated to determine if the ldquoidealrdquoscores were significantly different from each other The average standard

WINE AND CHEESE PAIRS 251

deviation from the mean was calculated for each cheese and wine combina-tion This was then averaged across all wines to give an indication of judgevariability

RESULTS AND DISCUSSION

Results of the ANOVA are shown in Table 3 As might be expected mostcheeses showed significant differences (

P

lt

0001) in the suitability of theirpairing with the 18 wines More surprisingly two cheeses (Swiss and mediumCheddar) showed no differences in their suitability with any of the winesTable 3 tabulates the mean score for each of the wine and cheese pairs Anideal pairing would have a mean score of six a mean score greater than sixwould indicate that the wine dominated the cheese a score less than six wouldindicate that the cheese dominated

Overall the relatively small deviations-from-ideal confirmed that wineand cheese are compatible food pairings In the three categories of wine whitewines tended to be a better match with these cheeses This was indicated byan average deviation-from-ideal of 085 compared to a deviation of 12 and17 for the red and specialty wines respectively The most versatile white winewas the Riesling followed by the Sauvignon blanc and the Pinot Gris asreflected by the smallest deviations The Pinot noir wine was the versatile redwine in that it paired best with the most number of cheeses The specialtysparkling and blush wines also paired well with many of the cheeses whilethe late harvest ice and port wines were the most difficult to pair with manyof the cheeses The Blue Benedictine Oka and Provolone cheeses were themost suitable match with these wines thus supporting the general rule thatstronger flavored cheeses tend to be better matched with stronger flavoredwines In general in this study the specialty wines (late harvest icewine porttype) dominated the cheeses However the addition of other foods or a dif-ference in the tasting method might change the perception of the pairing andlikely make these wines more compatible

Four cheeses (aged white Cheddar Migerone de Charlevoix Swiss andmedium Cheddar) had mean scores greater than six indicating that the wineswere more strongly flavored than the cheese Although the mean scores forthe 3-year old Cheddar were greater than six the means for the white winesthe Pinot noir and Merlot wines were not significantly different than six Thisindicated that there would be a range of wines that would be a suitable matchwith this cheese In general the stronger flavored cheeses were the mostvariable in their matches with the wines This indicated that more care wouldbe required in selecting wines to pair with these cheeses Figure 1 shows thedeviation-from-ideal for the match of the Benedictine Blue with each of the

252 M KING and M CLIFF

TAB

LE

3

ME

AN

DE

VIA

TIO

N-F

RO

M-I

DE

AL

SC

OR

ES

FOR

9 C

HE

ESE

S PA

IRE

D W

ITH

18

WIN

ES

(

n

=

9)

Win

ety

peW

ine

no

Win

eL

a pr

ovid

ence

d

cent

Oka

Blu

e B

ened

ictin

Prov

olon

eA

ged

3 ye

arC

hedd

ar

Mig

nero

n de

C

harl

evoi

x

Boe

renk

ass

Cam

embe

rtSw

iss

Med

ium

Che

ddar

Mea

n(

n

=

81)

acro

ss a

llch

eese

s

Whi

te1

Pino

t G

ris

50

61

62

68

76

65

55

75

77

654

Whi

te2

Rie

slin

g5

05

15

56

97

35

84

66

57

96

07W

hite

3Sa

uvig

non

blan

c4

04

45

16

86

15

76

46

77

25

82W

hite

4C

hard

onna

y (o

aked

)5

87

08

57

56

36

53

77

77

26

69W

hite

5G

ewur

ztra

min

er5

86

05

67

47

45

45

89

47

96

74W

hite

6C

hard

onna

y (u

noak

ed)

60

76

82

79

76

70

57

72

82

727

Red

7Pi

not

noir

(oa

ked)

52

59

58

65

77

71

68

65

67

647

Red

8R

ed M

erita

ge5

66

85

48

48

77

06

98

97

57

24R

ed9

Foch

(oa

ked)

48

65

80

83

90

76

70

87

86

761

Red

10Sy

rah

51

65

64

73

84

63

71

86

71

698

Red

11Pi

not

noir

(lig

ht)

40

40

57

52

67

66

57

62

68

566

Red

12M

erlo

t5

25

86

77

69

28

66

38

58

27

34Sp

ecia

lty13

Spar

klin

g6

35

45

86

57

06

15

37

67

36

37Sp

ecia

lty14

Blu

sh4

74

74

06

06

26

46

36

96

85

78Sp

ecia

lty15

LH

Gew

urzt

ram

iner

80

70

96

78

86

75

70

75

83

792

Spec

ialty

16R

iesl

ing

Icew

ine

79

73

87

83

97

82

83

90

100

860

Spec

ialty

17Pi

not

noir

Ice

win

e7

78

08

38

77

27

88

19

29

78

30Sp

ecia

lty18

Port

typ

e8

37

88

010

29

410

110

49

18

69

10St

anda

rd D

evia

tion

26

24

31

24

26

25

31

27

22

naL

SD1

891

752

241

751

961

852

19ns

nsna

P

pound

000

040

004

9E-0

40

003

002

000

60

0006

nsns

naR

ange

40ndash

79

40ndash

78

40ndash

87

52ndash

102

61ndash

97

54ndash

101

46ndash

104

65ndash

94

67ndash

10

WINE AND CHEESE PAIRS 253

wines In each of the wine categories there were wines where the cheesedominated and others where the wine dominated

The average standard deviation for each cheese across all wines rangedfrom 31 to 22 (Table 3) This relatively high deviation indicated the judgeswere not in agreement on their evaluation of the match For a trained descrip-tive panel using a 12-cm scale the standard deviation is generally 1ndash15Figure 2 showed the judge variation in assessing the match of Provolonecheese for three of the wines Although the mean score was not significantlydifferent from an ideal match of six the wide range of judgesrsquo assessmentswas evident For each of the three wines some of the judges thought thecheese dominated the wine while others thought the wine dominated thecheese Figure 3 showed that although there was some variation in the judgeassessment of how the medium Cheddar paired with the various wines therewas better agreement that the wine tended to dominate the cheese The dif-ference in variability is most likely due to the cheese with Provolone beinga stronger cheese and more difficult to match

It was very important to recognize the extent of the judge variation incheese and wine assessments Although all of the judges were individually

FIG 1 MEAN DEVIATION-FROM-IDEAL FOR BLUE CHEESE PAIRED WITH 18 WINES (

n

=

9)

ndash2

ndash15

ndash1

ndash05

0

50

1

51

2

52

Idea

l mat

ch

1

2

3

5

4 0196

21

11

87

41

7161 8151

31

winesetihWGristoniP1gnilseiR2

cnalbnongivuaS3)dekao(yannodrahC4

renimartzruweG5)dekaonu(yannodrahC6

Cheese dominates

Wine dominates

winesdeR)dekao(riontoniP7

egatireM8hcoF9haryS01

)thgil(riontoniP11tolreM21

winesytlaicepSeniwgnilkrapS31

hsulB-esoR41HLrenimartzruweG51

eniwecIgnilseiR61eniwecIriontoniP71

epyttroP81

254 M KING and M CLIFF

FIG 2 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING PROVOLONE CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

renimartzruweG

)thgil(riontoniPeniwgnilkrapS

STD = 315

STD = 361

STD = 417

Cheese dominates

Wine dominates

FIG 3 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING MEDIUM CHEDDAR CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

)thgil( rion toniPrenimartzruweG eniw gnilkrapS

STD = 266 STD = 247STD = 269

Wine dominates

Cheese dominates

WINE AND CHEESE PAIRS 255

experienced in food and wine pairing as a group their results under experi-mental conditions were not always consistent ie they did not always agreeTherefore the results obtained in this study are only broad recommendationsfor possible cheese and wine pairings In the light of the range of assessmentsobtained by knowledgeable judges it would be important to empower con-sumers to have the confidence to make their own assessment of how muchthey enjoy the cheese and wine pair

In conclusion white wines were easier to pair with a broader range ofcheeses than red or specialty wines The stronger more flavorful cheeses weremore difficult to pair with the wines but were more likely to be a good matchfor the late harvest and ice wines than other milder cheeses The results ofthis study were determined with a single tasting protocol and the wine andcheese were tasted without other foods The addition of other foods or agreater time period between the assessments would change the perception ofthe pairing Finally the variation in judge assessment indicates a high degreeof personal expectation or preference associated with the suitability of wineand cheese pairing Round-table discussion and consensus developmentmight produce ldquotrainedrdquo judges with less variation in the assessments althoughthis might be difficult given the ldquorespectedrdquo opinions of acknowledgedexperts Individuals should be encouraged to experiment to determine theirown preferences

ACKNOWLEDGMENTS

The authors would like to thank Reg Hendrickson of the Dairy Farmersof Canada for donating the cheese Christa-Lee Bond of the Okanagan WineFestival Society for allowing this research to be conducted as part of theIcewine Festival events Kimberly Dever for her technical assistance and theBC Winemakers and restaurateurs who participated in the study

REFERENCES

BALDY MW 1995

The University Wine Course

p 426 Wine AppreciationGuild San Francisco CA

BARTOSHUK LM 2000 Comparing sensory experiences across individu-als Recent psychophysical advances illuminate genetic variation in tasteperception Chem Senses

25

447ndash460COCHRAN WG amp COX GM 1957

Experimental Designs

p 611 JohnWiley amp Sons New York

256 M KING and M CLIFF

EampJ GALLO WINERY 2002 Gallo of sonoma wine and cheese pairingguide ABCrsquos of wine and cheese httpwwwwinewebcentralcomawineandcheeseabchtm (accessed October 24 2003)

JACKSON RS 2002

Wine Tasting A Professional Handbook

pp 261ndash262Academic Press San Diego CA

MOESSNER M Inniskillinrsquos guide to wine and cheese pairings httpwwwwineandleisurecomcheesepairinghtml (accessed October 242003)

NOBLE AC 1996 Tastendasharoma interactions Trends Food Sci Technol

7

439ndash444

PHILLIPS R 2000

A Short History of Wine

Penguin Books Londonhttpwwwintowinecomshop1282-0excerpthtml (accessed October24 2003)

SHEPHERD R SMITH K and FARLEIGH CA 1989 The relationshipbetween intensity hedonic and relative-to-ideal ratings Food QualPrefer

1

75ndash80SMITH J 1995

Cheesemaking in Scotland ndash A History

Scottish DairyAssociation Scotland httpwwwefrhwacukSDAbook1html(accessed October 24 2003)

Page 6: EVALUATION OF IDEAL WINE AND CHEESE PAIRS USING A DEVIATION-FROM-IDEAL SCALE WITH FOOD AND WINE EXPERTS

250 M KING and M CLIFF

number of times allowing experimental error from ldquocarryoverrdquo or ldquoresidualrdquoto be balanced across all samples

The wines were divided into three blocks of six white red and spe-cialty wines respectively The white and red wines were tasted in a randomorder while the specialty wines had an assigned order of sparkling winefirst and Icewines last This was thought to be necessary because Icewinesare very sweet and intense with a potential for a strong carryover effect thatcould interfere with a judgersquos ability to assess the other cheese and winepairs

Using the incomplete block design each judge tasted three cheeses withsix of the wines The cheeses were systematically paired with the wines sothat after the three blocks of wines a judge had tasted all nine cheeses andeach cheese had been evaluated nine times with each of the wines Judgeswere asked to evaluate the suitability of the pairing for each cheese with allof the wines in the block before moving on to the next cheese

Tasting Process

Judges were instructed to take a small bite of cheese sip the wine andevaluate the appropriateness of the pair The judges evaluated the match usinga variation of the ldquojust right scalerdquo (Shepherd

et al

1989) Judges scored thedeviation from an ldquoideal matchrdquo on a 12-cm line scale where the ideal matchwas the midpoint of the line (6 cm) If the judges thought the cheese domi-nated the wine they marked the line scale to the left of the midpoint whereasif they thought the wine dominated the cheese the judges marked the line scaleto the right of the midpoint The scale was labeled with ldquocheese dominatesexcessivelyrdquo ldquocheese dominates moderatelyrdquo and ldquocheese dominates slightlyrdquoat 0 2 and 4 cm The other half of the scale was labeled with ldquowine dominatesslightlyrdquo ldquowine dominates moderatelyrdquo and ldquowine dominates excessivelyrdquo at8 10 12 cm respectively

Judges were given sparkling water bread and melba toast in order toclear their palates between evaluations They were also instructed to take ashort break between cheeses After the tasting the scorecards were collectedand the judgesrsquo scores for each wine and cheese pairing were measured(0ndash12) for data analysis and interpretation

Statistical Analysis

A single factor analysis of variance (ANOVA) (Microsoft Excel Seat-tle WA) was used to determine if there were significant differences in thesuitability of the match between each cheese with each wine A Fisherrsquosleast significant difference (LSD) was calculated to determine if the ldquoidealrdquoscores were significantly different from each other The average standard

WINE AND CHEESE PAIRS 251

deviation from the mean was calculated for each cheese and wine combina-tion This was then averaged across all wines to give an indication of judgevariability

RESULTS AND DISCUSSION

Results of the ANOVA are shown in Table 3 As might be expected mostcheeses showed significant differences (

P

lt

0001) in the suitability of theirpairing with the 18 wines More surprisingly two cheeses (Swiss and mediumCheddar) showed no differences in their suitability with any of the winesTable 3 tabulates the mean score for each of the wine and cheese pairs Anideal pairing would have a mean score of six a mean score greater than sixwould indicate that the wine dominated the cheese a score less than six wouldindicate that the cheese dominated

Overall the relatively small deviations-from-ideal confirmed that wineand cheese are compatible food pairings In the three categories of wine whitewines tended to be a better match with these cheeses This was indicated byan average deviation-from-ideal of 085 compared to a deviation of 12 and17 for the red and specialty wines respectively The most versatile white winewas the Riesling followed by the Sauvignon blanc and the Pinot Gris asreflected by the smallest deviations The Pinot noir wine was the versatile redwine in that it paired best with the most number of cheeses The specialtysparkling and blush wines also paired well with many of the cheeses whilethe late harvest ice and port wines were the most difficult to pair with manyof the cheeses The Blue Benedictine Oka and Provolone cheeses were themost suitable match with these wines thus supporting the general rule thatstronger flavored cheeses tend to be better matched with stronger flavoredwines In general in this study the specialty wines (late harvest icewine porttype) dominated the cheeses However the addition of other foods or a dif-ference in the tasting method might change the perception of the pairing andlikely make these wines more compatible

Four cheeses (aged white Cheddar Migerone de Charlevoix Swiss andmedium Cheddar) had mean scores greater than six indicating that the wineswere more strongly flavored than the cheese Although the mean scores forthe 3-year old Cheddar were greater than six the means for the white winesthe Pinot noir and Merlot wines were not significantly different than six Thisindicated that there would be a range of wines that would be a suitable matchwith this cheese In general the stronger flavored cheeses were the mostvariable in their matches with the wines This indicated that more care wouldbe required in selecting wines to pair with these cheeses Figure 1 shows thedeviation-from-ideal for the match of the Benedictine Blue with each of the

252 M KING and M CLIFF

TAB

LE

3

ME

AN

DE

VIA

TIO

N-F

RO

M-I

DE

AL

SC

OR

ES

FOR

9 C

HE

ESE

S PA

IRE

D W

ITH

18

WIN

ES

(

n

=

9)

Win

ety

peW

ine

no

Win

eL

a pr

ovid

ence

d

cent

Oka

Blu

e B

ened

ictin

Prov

olon

eA

ged

3 ye

arC

hedd

ar

Mig

nero

n de

C

harl

evoi

x

Boe

renk

ass

Cam

embe

rtSw

iss

Med

ium

Che

ddar

Mea

n(

n

=

81)

acro

ss a

llch

eese

s

Whi

te1

Pino

t G

ris

50

61

62

68

76

65

55

75

77

654

Whi

te2

Rie

slin

g5

05

15

56

97

35

84

66

57

96

07W

hite

3Sa

uvig

non

blan

c4

04

45

16

86

15

76

46

77

25

82W

hite

4C

hard

onna

y (o

aked

)5

87

08

57

56

36

53

77

77

26

69W

hite

5G

ewur

ztra

min

er5

86

05

67

47

45

45

89

47

96

74W

hite

6C

hard

onna

y (u

noak

ed)

60

76

82

79

76

70

57

72

82

727

Red

7Pi

not

noir

(oa

ked)

52

59

58

65

77

71

68

65

67

647

Red

8R

ed M

erita

ge5

66

85

48

48

77

06

98

97

57

24R

ed9

Foch

(oa

ked)

48

65

80

83

90

76

70

87

86

761

Red

10Sy

rah

51

65

64

73

84

63

71

86

71

698

Red

11Pi

not

noir

(lig

ht)

40

40

57

52

67

66

57

62

68

566

Red

12M

erlo

t5

25

86

77

69

28

66

38

58

27

34Sp

ecia

lty13

Spar

klin

g6

35

45

86

57

06

15

37

67

36

37Sp

ecia

lty14

Blu

sh4

74

74

06

06

26

46

36

96

85

78Sp

ecia

lty15

LH

Gew

urzt

ram

iner

80

70

96

78

86

75

70

75

83

792

Spec

ialty

16R

iesl

ing

Icew

ine

79

73

87

83

97

82

83

90

100

860

Spec

ialty

17Pi

not

noir

Ice

win

e7

78

08

38

77

27

88

19

29

78

30Sp

ecia

lty18

Port

typ

e8

37

88

010

29

410

110

49

18

69

10St

anda

rd D

evia

tion

26

24

31

24

26

25

31

27

22

naL

SD1

891

752

241

751

961

852

19ns

nsna

P

pound

000

040

004

9E-0

40

003

002

000

60

0006

nsns

naR

ange

40ndash

79

40ndash

78

40ndash

87

52ndash

102

61ndash

97

54ndash

101

46ndash

104

65ndash

94

67ndash

10

WINE AND CHEESE PAIRS 253

wines In each of the wine categories there were wines where the cheesedominated and others where the wine dominated

The average standard deviation for each cheese across all wines rangedfrom 31 to 22 (Table 3) This relatively high deviation indicated the judgeswere not in agreement on their evaluation of the match For a trained descrip-tive panel using a 12-cm scale the standard deviation is generally 1ndash15Figure 2 showed the judge variation in assessing the match of Provolonecheese for three of the wines Although the mean score was not significantlydifferent from an ideal match of six the wide range of judgesrsquo assessmentswas evident For each of the three wines some of the judges thought thecheese dominated the wine while others thought the wine dominated thecheese Figure 3 showed that although there was some variation in the judgeassessment of how the medium Cheddar paired with the various wines therewas better agreement that the wine tended to dominate the cheese The dif-ference in variability is most likely due to the cheese with Provolone beinga stronger cheese and more difficult to match

It was very important to recognize the extent of the judge variation incheese and wine assessments Although all of the judges were individually

FIG 1 MEAN DEVIATION-FROM-IDEAL FOR BLUE CHEESE PAIRED WITH 18 WINES (

n

=

9)

ndash2

ndash15

ndash1

ndash05

0

50

1

51

2

52

Idea

l mat

ch

1

2

3

5

4 0196

21

11

87

41

7161 8151

31

winesetihWGristoniP1gnilseiR2

cnalbnongivuaS3)dekao(yannodrahC4

renimartzruweG5)dekaonu(yannodrahC6

Cheese dominates

Wine dominates

winesdeR)dekao(riontoniP7

egatireM8hcoF9haryS01

)thgil(riontoniP11tolreM21

winesytlaicepSeniwgnilkrapS31

hsulB-esoR41HLrenimartzruweG51

eniwecIgnilseiR61eniwecIriontoniP71

epyttroP81

254 M KING and M CLIFF

FIG 2 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING PROVOLONE CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

renimartzruweG

)thgil(riontoniPeniwgnilkrapS

STD = 315

STD = 361

STD = 417

Cheese dominates

Wine dominates

FIG 3 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING MEDIUM CHEDDAR CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

)thgil( rion toniPrenimartzruweG eniw gnilkrapS

STD = 266 STD = 247STD = 269

Wine dominates

Cheese dominates

WINE AND CHEESE PAIRS 255

experienced in food and wine pairing as a group their results under experi-mental conditions were not always consistent ie they did not always agreeTherefore the results obtained in this study are only broad recommendationsfor possible cheese and wine pairings In the light of the range of assessmentsobtained by knowledgeable judges it would be important to empower con-sumers to have the confidence to make their own assessment of how muchthey enjoy the cheese and wine pair

In conclusion white wines were easier to pair with a broader range ofcheeses than red or specialty wines The stronger more flavorful cheeses weremore difficult to pair with the wines but were more likely to be a good matchfor the late harvest and ice wines than other milder cheeses The results ofthis study were determined with a single tasting protocol and the wine andcheese were tasted without other foods The addition of other foods or agreater time period between the assessments would change the perception ofthe pairing Finally the variation in judge assessment indicates a high degreeof personal expectation or preference associated with the suitability of wineand cheese pairing Round-table discussion and consensus developmentmight produce ldquotrainedrdquo judges with less variation in the assessments althoughthis might be difficult given the ldquorespectedrdquo opinions of acknowledgedexperts Individuals should be encouraged to experiment to determine theirown preferences

ACKNOWLEDGMENTS

The authors would like to thank Reg Hendrickson of the Dairy Farmersof Canada for donating the cheese Christa-Lee Bond of the Okanagan WineFestival Society for allowing this research to be conducted as part of theIcewine Festival events Kimberly Dever for her technical assistance and theBC Winemakers and restaurateurs who participated in the study

REFERENCES

BALDY MW 1995

The University Wine Course

p 426 Wine AppreciationGuild San Francisco CA

BARTOSHUK LM 2000 Comparing sensory experiences across individu-als Recent psychophysical advances illuminate genetic variation in tasteperception Chem Senses

25

447ndash460COCHRAN WG amp COX GM 1957

Experimental Designs

p 611 JohnWiley amp Sons New York

256 M KING and M CLIFF

EampJ GALLO WINERY 2002 Gallo of sonoma wine and cheese pairingguide ABCrsquos of wine and cheese httpwwwwinewebcentralcomawineandcheeseabchtm (accessed October 24 2003)

JACKSON RS 2002

Wine Tasting A Professional Handbook

pp 261ndash262Academic Press San Diego CA

MOESSNER M Inniskillinrsquos guide to wine and cheese pairings httpwwwwineandleisurecomcheesepairinghtml (accessed October 242003)

NOBLE AC 1996 Tastendasharoma interactions Trends Food Sci Technol

7

439ndash444

PHILLIPS R 2000

A Short History of Wine

Penguin Books Londonhttpwwwintowinecomshop1282-0excerpthtml (accessed October24 2003)

SHEPHERD R SMITH K and FARLEIGH CA 1989 The relationshipbetween intensity hedonic and relative-to-ideal ratings Food QualPrefer

1

75ndash80SMITH J 1995

Cheesemaking in Scotland ndash A History

Scottish DairyAssociation Scotland httpwwwefrhwacukSDAbook1html(accessed October 24 2003)

Page 7: EVALUATION OF IDEAL WINE AND CHEESE PAIRS USING A DEVIATION-FROM-IDEAL SCALE WITH FOOD AND WINE EXPERTS

WINE AND CHEESE PAIRS 251

deviation from the mean was calculated for each cheese and wine combina-tion This was then averaged across all wines to give an indication of judgevariability

RESULTS AND DISCUSSION

Results of the ANOVA are shown in Table 3 As might be expected mostcheeses showed significant differences (

P

lt

0001) in the suitability of theirpairing with the 18 wines More surprisingly two cheeses (Swiss and mediumCheddar) showed no differences in their suitability with any of the winesTable 3 tabulates the mean score for each of the wine and cheese pairs Anideal pairing would have a mean score of six a mean score greater than sixwould indicate that the wine dominated the cheese a score less than six wouldindicate that the cheese dominated

Overall the relatively small deviations-from-ideal confirmed that wineand cheese are compatible food pairings In the three categories of wine whitewines tended to be a better match with these cheeses This was indicated byan average deviation-from-ideal of 085 compared to a deviation of 12 and17 for the red and specialty wines respectively The most versatile white winewas the Riesling followed by the Sauvignon blanc and the Pinot Gris asreflected by the smallest deviations The Pinot noir wine was the versatile redwine in that it paired best with the most number of cheeses The specialtysparkling and blush wines also paired well with many of the cheeses whilethe late harvest ice and port wines were the most difficult to pair with manyof the cheeses The Blue Benedictine Oka and Provolone cheeses were themost suitable match with these wines thus supporting the general rule thatstronger flavored cheeses tend to be better matched with stronger flavoredwines In general in this study the specialty wines (late harvest icewine porttype) dominated the cheeses However the addition of other foods or a dif-ference in the tasting method might change the perception of the pairing andlikely make these wines more compatible

Four cheeses (aged white Cheddar Migerone de Charlevoix Swiss andmedium Cheddar) had mean scores greater than six indicating that the wineswere more strongly flavored than the cheese Although the mean scores forthe 3-year old Cheddar were greater than six the means for the white winesthe Pinot noir and Merlot wines were not significantly different than six Thisindicated that there would be a range of wines that would be a suitable matchwith this cheese In general the stronger flavored cheeses were the mostvariable in their matches with the wines This indicated that more care wouldbe required in selecting wines to pair with these cheeses Figure 1 shows thedeviation-from-ideal for the match of the Benedictine Blue with each of the

252 M KING and M CLIFF

TAB

LE

3

ME

AN

DE

VIA

TIO

N-F

RO

M-I

DE

AL

SC

OR

ES

FOR

9 C

HE

ESE

S PA

IRE

D W

ITH

18

WIN

ES

(

n

=

9)

Win

ety

peW

ine

no

Win

eL

a pr

ovid

ence

d

cent

Oka

Blu

e B

ened

ictin

Prov

olon

eA

ged

3 ye

arC

hedd

ar

Mig

nero

n de

C

harl

evoi

x

Boe

renk

ass

Cam

embe

rtSw

iss

Med

ium

Che

ddar

Mea

n(

n

=

81)

acro

ss a

llch

eese

s

Whi

te1

Pino

t G

ris

50

61

62

68

76

65

55

75

77

654

Whi

te2

Rie

slin

g5

05

15

56

97

35

84

66

57

96

07W

hite

3Sa

uvig

non

blan

c4

04

45

16

86

15

76

46

77

25

82W

hite

4C

hard

onna

y (o

aked

)5

87

08

57

56

36

53

77

77

26

69W

hite

5G

ewur

ztra

min

er5

86

05

67

47

45

45

89

47

96

74W

hite

6C

hard

onna

y (u

noak

ed)

60

76

82

79

76

70

57

72

82

727

Red

7Pi

not

noir

(oa

ked)

52

59

58

65

77

71

68

65

67

647

Red

8R

ed M

erita

ge5

66

85

48

48

77

06

98

97

57

24R

ed9

Foch

(oa

ked)

48

65

80

83

90

76

70

87

86

761

Red

10Sy

rah

51

65

64

73

84

63

71

86

71

698

Red

11Pi

not

noir

(lig

ht)

40

40

57

52

67

66

57

62

68

566

Red

12M

erlo

t5

25

86

77

69

28

66

38

58

27

34Sp

ecia

lty13

Spar

klin

g6

35

45

86

57

06

15

37

67

36

37Sp

ecia

lty14

Blu

sh4

74

74

06

06

26

46

36

96

85

78Sp

ecia

lty15

LH

Gew

urzt

ram

iner

80

70

96

78

86

75

70

75

83

792

Spec

ialty

16R

iesl

ing

Icew

ine

79

73

87

83

97

82

83

90

100

860

Spec

ialty

17Pi

not

noir

Ice

win

e7

78

08

38

77

27

88

19

29

78

30Sp

ecia

lty18

Port

typ

e8

37

88

010

29

410

110

49

18

69

10St

anda

rd D

evia

tion

26

24

31

24

26

25

31

27

22

naL

SD1

891

752

241

751

961

852

19ns

nsna

P

pound

000

040

004

9E-0

40

003

002

000

60

0006

nsns

naR

ange

40ndash

79

40ndash

78

40ndash

87

52ndash

102

61ndash

97

54ndash

101

46ndash

104

65ndash

94

67ndash

10

WINE AND CHEESE PAIRS 253

wines In each of the wine categories there were wines where the cheesedominated and others where the wine dominated

The average standard deviation for each cheese across all wines rangedfrom 31 to 22 (Table 3) This relatively high deviation indicated the judgeswere not in agreement on their evaluation of the match For a trained descrip-tive panel using a 12-cm scale the standard deviation is generally 1ndash15Figure 2 showed the judge variation in assessing the match of Provolonecheese for three of the wines Although the mean score was not significantlydifferent from an ideal match of six the wide range of judgesrsquo assessmentswas evident For each of the three wines some of the judges thought thecheese dominated the wine while others thought the wine dominated thecheese Figure 3 showed that although there was some variation in the judgeassessment of how the medium Cheddar paired with the various wines therewas better agreement that the wine tended to dominate the cheese The dif-ference in variability is most likely due to the cheese with Provolone beinga stronger cheese and more difficult to match

It was very important to recognize the extent of the judge variation incheese and wine assessments Although all of the judges were individually

FIG 1 MEAN DEVIATION-FROM-IDEAL FOR BLUE CHEESE PAIRED WITH 18 WINES (

n

=

9)

ndash2

ndash15

ndash1

ndash05

0

50

1

51

2

52

Idea

l mat

ch

1

2

3

5

4 0196

21

11

87

41

7161 8151

31

winesetihWGristoniP1gnilseiR2

cnalbnongivuaS3)dekao(yannodrahC4

renimartzruweG5)dekaonu(yannodrahC6

Cheese dominates

Wine dominates

winesdeR)dekao(riontoniP7

egatireM8hcoF9haryS01

)thgil(riontoniP11tolreM21

winesytlaicepSeniwgnilkrapS31

hsulB-esoR41HLrenimartzruweG51

eniwecIgnilseiR61eniwecIriontoniP71

epyttroP81

254 M KING and M CLIFF

FIG 2 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING PROVOLONE CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

renimartzruweG

)thgil(riontoniPeniwgnilkrapS

STD = 315

STD = 361

STD = 417

Cheese dominates

Wine dominates

FIG 3 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING MEDIUM CHEDDAR CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

)thgil( rion toniPrenimartzruweG eniw gnilkrapS

STD = 266 STD = 247STD = 269

Wine dominates

Cheese dominates

WINE AND CHEESE PAIRS 255

experienced in food and wine pairing as a group their results under experi-mental conditions were not always consistent ie they did not always agreeTherefore the results obtained in this study are only broad recommendationsfor possible cheese and wine pairings In the light of the range of assessmentsobtained by knowledgeable judges it would be important to empower con-sumers to have the confidence to make their own assessment of how muchthey enjoy the cheese and wine pair

In conclusion white wines were easier to pair with a broader range ofcheeses than red or specialty wines The stronger more flavorful cheeses weremore difficult to pair with the wines but were more likely to be a good matchfor the late harvest and ice wines than other milder cheeses The results ofthis study were determined with a single tasting protocol and the wine andcheese were tasted without other foods The addition of other foods or agreater time period between the assessments would change the perception ofthe pairing Finally the variation in judge assessment indicates a high degreeof personal expectation or preference associated with the suitability of wineand cheese pairing Round-table discussion and consensus developmentmight produce ldquotrainedrdquo judges with less variation in the assessments althoughthis might be difficult given the ldquorespectedrdquo opinions of acknowledgedexperts Individuals should be encouraged to experiment to determine theirown preferences

ACKNOWLEDGMENTS

The authors would like to thank Reg Hendrickson of the Dairy Farmersof Canada for donating the cheese Christa-Lee Bond of the Okanagan WineFestival Society for allowing this research to be conducted as part of theIcewine Festival events Kimberly Dever for her technical assistance and theBC Winemakers and restaurateurs who participated in the study

REFERENCES

BALDY MW 1995

The University Wine Course

p 426 Wine AppreciationGuild San Francisco CA

BARTOSHUK LM 2000 Comparing sensory experiences across individu-als Recent psychophysical advances illuminate genetic variation in tasteperception Chem Senses

25

447ndash460COCHRAN WG amp COX GM 1957

Experimental Designs

p 611 JohnWiley amp Sons New York

256 M KING and M CLIFF

EampJ GALLO WINERY 2002 Gallo of sonoma wine and cheese pairingguide ABCrsquos of wine and cheese httpwwwwinewebcentralcomawineandcheeseabchtm (accessed October 24 2003)

JACKSON RS 2002

Wine Tasting A Professional Handbook

pp 261ndash262Academic Press San Diego CA

MOESSNER M Inniskillinrsquos guide to wine and cheese pairings httpwwwwineandleisurecomcheesepairinghtml (accessed October 242003)

NOBLE AC 1996 Tastendasharoma interactions Trends Food Sci Technol

7

439ndash444

PHILLIPS R 2000

A Short History of Wine

Penguin Books Londonhttpwwwintowinecomshop1282-0excerpthtml (accessed October24 2003)

SHEPHERD R SMITH K and FARLEIGH CA 1989 The relationshipbetween intensity hedonic and relative-to-ideal ratings Food QualPrefer

1

75ndash80SMITH J 1995

Cheesemaking in Scotland ndash A History

Scottish DairyAssociation Scotland httpwwwefrhwacukSDAbook1html(accessed October 24 2003)

Page 8: EVALUATION OF IDEAL WINE AND CHEESE PAIRS USING A DEVIATION-FROM-IDEAL SCALE WITH FOOD AND WINE EXPERTS

252 M KING and M CLIFF

TAB

LE

3

ME

AN

DE

VIA

TIO

N-F

RO

M-I

DE

AL

SC

OR

ES

FOR

9 C

HE

ESE

S PA

IRE

D W

ITH

18

WIN

ES

(

n

=

9)

Win

ety

peW

ine

no

Win

eL

a pr

ovid

ence

d

cent

Oka

Blu

e B

ened

ictin

Prov

olon

eA

ged

3 ye

arC

hedd

ar

Mig

nero

n de

C

harl

evoi

x

Boe

renk

ass

Cam

embe

rtSw

iss

Med

ium

Che

ddar

Mea

n(

n

=

81)

acro

ss a

llch

eese

s

Whi

te1

Pino

t G

ris

50

61

62

68

76

65

55

75

77

654

Whi

te2

Rie

slin

g5

05

15

56

97

35

84

66

57

96

07W

hite

3Sa

uvig

non

blan

c4

04

45

16

86

15

76

46

77

25

82W

hite

4C

hard

onna

y (o

aked

)5

87

08

57

56

36

53

77

77

26

69W

hite

5G

ewur

ztra

min

er5

86

05

67

47

45

45

89

47

96

74W

hite

6C

hard

onna

y (u

noak

ed)

60

76

82

79

76

70

57

72

82

727

Red

7Pi

not

noir

(oa

ked)

52

59

58

65

77

71

68

65

67

647

Red

8R

ed M

erita

ge5

66

85

48

48

77

06

98

97

57

24R

ed9

Foch

(oa

ked)

48

65

80

83

90

76

70

87

86

761

Red

10Sy

rah

51

65

64

73

84

63

71

86

71

698

Red

11Pi

not

noir

(lig

ht)

40

40

57

52

67

66

57

62

68

566

Red

12M

erlo

t5

25

86

77

69

28

66

38

58

27

34Sp

ecia

lty13

Spar

klin

g6

35

45

86

57

06

15

37

67

36

37Sp

ecia

lty14

Blu

sh4

74

74

06

06

26

46

36

96

85

78Sp

ecia

lty15

LH

Gew

urzt

ram

iner

80

70

96

78

86

75

70

75

83

792

Spec

ialty

16R

iesl

ing

Icew

ine

79

73

87

83

97

82

83

90

100

860

Spec

ialty

17Pi

not

noir

Ice

win

e7

78

08

38

77

27

88

19

29

78

30Sp

ecia

lty18

Port

typ

e8

37

88

010

29

410

110

49

18

69

10St

anda

rd D

evia

tion

26

24

31

24

26

25

31

27

22

naL

SD1

891

752

241

751

961

852

19ns

nsna

P

pound

000

040

004

9E-0

40

003

002

000

60

0006

nsns

naR

ange

40ndash

79

40ndash

78

40ndash

87

52ndash

102

61ndash

97

54ndash

101

46ndash

104

65ndash

94

67ndash

10

WINE AND CHEESE PAIRS 253

wines In each of the wine categories there were wines where the cheesedominated and others where the wine dominated

The average standard deviation for each cheese across all wines rangedfrom 31 to 22 (Table 3) This relatively high deviation indicated the judgeswere not in agreement on their evaluation of the match For a trained descrip-tive panel using a 12-cm scale the standard deviation is generally 1ndash15Figure 2 showed the judge variation in assessing the match of Provolonecheese for three of the wines Although the mean score was not significantlydifferent from an ideal match of six the wide range of judgesrsquo assessmentswas evident For each of the three wines some of the judges thought thecheese dominated the wine while others thought the wine dominated thecheese Figure 3 showed that although there was some variation in the judgeassessment of how the medium Cheddar paired with the various wines therewas better agreement that the wine tended to dominate the cheese The dif-ference in variability is most likely due to the cheese with Provolone beinga stronger cheese and more difficult to match

It was very important to recognize the extent of the judge variation incheese and wine assessments Although all of the judges were individually

FIG 1 MEAN DEVIATION-FROM-IDEAL FOR BLUE CHEESE PAIRED WITH 18 WINES (

n

=

9)

ndash2

ndash15

ndash1

ndash05

0

50

1

51

2

52

Idea

l mat

ch

1

2

3

5

4 0196

21

11

87

41

7161 8151

31

winesetihWGristoniP1gnilseiR2

cnalbnongivuaS3)dekao(yannodrahC4

renimartzruweG5)dekaonu(yannodrahC6

Cheese dominates

Wine dominates

winesdeR)dekao(riontoniP7

egatireM8hcoF9haryS01

)thgil(riontoniP11tolreM21

winesytlaicepSeniwgnilkrapS31

hsulB-esoR41HLrenimartzruweG51

eniwecIgnilseiR61eniwecIriontoniP71

epyttroP81

254 M KING and M CLIFF

FIG 2 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING PROVOLONE CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

renimartzruweG

)thgil(riontoniPeniwgnilkrapS

STD = 315

STD = 361

STD = 417

Cheese dominates

Wine dominates

FIG 3 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING MEDIUM CHEDDAR CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

)thgil( rion toniPrenimartzruweG eniw gnilkrapS

STD = 266 STD = 247STD = 269

Wine dominates

Cheese dominates

WINE AND CHEESE PAIRS 255

experienced in food and wine pairing as a group their results under experi-mental conditions were not always consistent ie they did not always agreeTherefore the results obtained in this study are only broad recommendationsfor possible cheese and wine pairings In the light of the range of assessmentsobtained by knowledgeable judges it would be important to empower con-sumers to have the confidence to make their own assessment of how muchthey enjoy the cheese and wine pair

In conclusion white wines were easier to pair with a broader range ofcheeses than red or specialty wines The stronger more flavorful cheeses weremore difficult to pair with the wines but were more likely to be a good matchfor the late harvest and ice wines than other milder cheeses The results ofthis study were determined with a single tasting protocol and the wine andcheese were tasted without other foods The addition of other foods or agreater time period between the assessments would change the perception ofthe pairing Finally the variation in judge assessment indicates a high degreeof personal expectation or preference associated with the suitability of wineand cheese pairing Round-table discussion and consensus developmentmight produce ldquotrainedrdquo judges with less variation in the assessments althoughthis might be difficult given the ldquorespectedrdquo opinions of acknowledgedexperts Individuals should be encouraged to experiment to determine theirown preferences

ACKNOWLEDGMENTS

The authors would like to thank Reg Hendrickson of the Dairy Farmersof Canada for donating the cheese Christa-Lee Bond of the Okanagan WineFestival Society for allowing this research to be conducted as part of theIcewine Festival events Kimberly Dever for her technical assistance and theBC Winemakers and restaurateurs who participated in the study

REFERENCES

BALDY MW 1995

The University Wine Course

p 426 Wine AppreciationGuild San Francisco CA

BARTOSHUK LM 2000 Comparing sensory experiences across individu-als Recent psychophysical advances illuminate genetic variation in tasteperception Chem Senses

25

447ndash460COCHRAN WG amp COX GM 1957

Experimental Designs

p 611 JohnWiley amp Sons New York

256 M KING and M CLIFF

EampJ GALLO WINERY 2002 Gallo of sonoma wine and cheese pairingguide ABCrsquos of wine and cheese httpwwwwinewebcentralcomawineandcheeseabchtm (accessed October 24 2003)

JACKSON RS 2002

Wine Tasting A Professional Handbook

pp 261ndash262Academic Press San Diego CA

MOESSNER M Inniskillinrsquos guide to wine and cheese pairings httpwwwwineandleisurecomcheesepairinghtml (accessed October 242003)

NOBLE AC 1996 Tastendasharoma interactions Trends Food Sci Technol

7

439ndash444

PHILLIPS R 2000

A Short History of Wine

Penguin Books Londonhttpwwwintowinecomshop1282-0excerpthtml (accessed October24 2003)

SHEPHERD R SMITH K and FARLEIGH CA 1989 The relationshipbetween intensity hedonic and relative-to-ideal ratings Food QualPrefer

1

75ndash80SMITH J 1995

Cheesemaking in Scotland ndash A History

Scottish DairyAssociation Scotland httpwwwefrhwacukSDAbook1html(accessed October 24 2003)

Page 9: EVALUATION OF IDEAL WINE AND CHEESE PAIRS USING A DEVIATION-FROM-IDEAL SCALE WITH FOOD AND WINE EXPERTS

WINE AND CHEESE PAIRS 253

wines In each of the wine categories there were wines where the cheesedominated and others where the wine dominated

The average standard deviation for each cheese across all wines rangedfrom 31 to 22 (Table 3) This relatively high deviation indicated the judgeswere not in agreement on their evaluation of the match For a trained descrip-tive panel using a 12-cm scale the standard deviation is generally 1ndash15Figure 2 showed the judge variation in assessing the match of Provolonecheese for three of the wines Although the mean score was not significantlydifferent from an ideal match of six the wide range of judgesrsquo assessmentswas evident For each of the three wines some of the judges thought thecheese dominated the wine while others thought the wine dominated thecheese Figure 3 showed that although there was some variation in the judgeassessment of how the medium Cheddar paired with the various wines therewas better agreement that the wine tended to dominate the cheese The dif-ference in variability is most likely due to the cheese with Provolone beinga stronger cheese and more difficult to match

It was very important to recognize the extent of the judge variation incheese and wine assessments Although all of the judges were individually

FIG 1 MEAN DEVIATION-FROM-IDEAL FOR BLUE CHEESE PAIRED WITH 18 WINES (

n

=

9)

ndash2

ndash15

ndash1

ndash05

0

50

1

51

2

52

Idea

l mat

ch

1

2

3

5

4 0196

21

11

87

41

7161 8151

31

winesetihWGristoniP1gnilseiR2

cnalbnongivuaS3)dekao(yannodrahC4

renimartzruweG5)dekaonu(yannodrahC6

Cheese dominates

Wine dominates

winesdeR)dekao(riontoniP7

egatireM8hcoF9haryS01

)thgil(riontoniP11tolreM21

winesytlaicepSeniwgnilkrapS31

hsulB-esoR41HLrenimartzruweG51

eniwecIgnilseiR61eniwecIriontoniP71

epyttroP81

254 M KING and M CLIFF

FIG 2 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING PROVOLONE CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

renimartzruweG

)thgil(riontoniPeniwgnilkrapS

STD = 315

STD = 361

STD = 417

Cheese dominates

Wine dominates

FIG 3 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING MEDIUM CHEDDAR CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

)thgil( rion toniPrenimartzruweG eniw gnilkrapS

STD = 266 STD = 247STD = 269

Wine dominates

Cheese dominates

WINE AND CHEESE PAIRS 255

experienced in food and wine pairing as a group their results under experi-mental conditions were not always consistent ie they did not always agreeTherefore the results obtained in this study are only broad recommendationsfor possible cheese and wine pairings In the light of the range of assessmentsobtained by knowledgeable judges it would be important to empower con-sumers to have the confidence to make their own assessment of how muchthey enjoy the cheese and wine pair

In conclusion white wines were easier to pair with a broader range ofcheeses than red or specialty wines The stronger more flavorful cheeses weremore difficult to pair with the wines but were more likely to be a good matchfor the late harvest and ice wines than other milder cheeses The results ofthis study were determined with a single tasting protocol and the wine andcheese were tasted without other foods The addition of other foods or agreater time period between the assessments would change the perception ofthe pairing Finally the variation in judge assessment indicates a high degreeof personal expectation or preference associated with the suitability of wineand cheese pairing Round-table discussion and consensus developmentmight produce ldquotrainedrdquo judges with less variation in the assessments althoughthis might be difficult given the ldquorespectedrdquo opinions of acknowledgedexperts Individuals should be encouraged to experiment to determine theirown preferences

ACKNOWLEDGMENTS

The authors would like to thank Reg Hendrickson of the Dairy Farmersof Canada for donating the cheese Christa-Lee Bond of the Okanagan WineFestival Society for allowing this research to be conducted as part of theIcewine Festival events Kimberly Dever for her technical assistance and theBC Winemakers and restaurateurs who participated in the study

REFERENCES

BALDY MW 1995

The University Wine Course

p 426 Wine AppreciationGuild San Francisco CA

BARTOSHUK LM 2000 Comparing sensory experiences across individu-als Recent psychophysical advances illuminate genetic variation in tasteperception Chem Senses

25

447ndash460COCHRAN WG amp COX GM 1957

Experimental Designs

p 611 JohnWiley amp Sons New York

256 M KING and M CLIFF

EampJ GALLO WINERY 2002 Gallo of sonoma wine and cheese pairingguide ABCrsquos of wine and cheese httpwwwwinewebcentralcomawineandcheeseabchtm (accessed October 24 2003)

JACKSON RS 2002

Wine Tasting A Professional Handbook

pp 261ndash262Academic Press San Diego CA

MOESSNER M Inniskillinrsquos guide to wine and cheese pairings httpwwwwineandleisurecomcheesepairinghtml (accessed October 242003)

NOBLE AC 1996 Tastendasharoma interactions Trends Food Sci Technol

7

439ndash444

PHILLIPS R 2000

A Short History of Wine

Penguin Books Londonhttpwwwintowinecomshop1282-0excerpthtml (accessed October24 2003)

SHEPHERD R SMITH K and FARLEIGH CA 1989 The relationshipbetween intensity hedonic and relative-to-ideal ratings Food QualPrefer

1

75ndash80SMITH J 1995

Cheesemaking in Scotland ndash A History

Scottish DairyAssociation Scotland httpwwwefrhwacukSDAbook1html(accessed October 24 2003)

Page 10: EVALUATION OF IDEAL WINE AND CHEESE PAIRS USING A DEVIATION-FROM-IDEAL SCALE WITH FOOD AND WINE EXPERTS

254 M KING and M CLIFF

FIG 2 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING PROVOLONE CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

renimartzruweG

)thgil(riontoniPeniwgnilkrapS

STD = 315

STD = 361

STD = 417

Cheese dominates

Wine dominates

FIG 3 DEVIATION-FROM-IDEAL FOR NINE JUDGE ASSESSING MEDIUM CHEDDAR CHEESE WITH THREE WINES

ndash6

ndash4

ndash2

0

2

4

6

8

Idea

l mat

ch

)thgil( rion toniPrenimartzruweG eniw gnilkrapS

STD = 266 STD = 247STD = 269

Wine dominates

Cheese dominates

WINE AND CHEESE PAIRS 255

experienced in food and wine pairing as a group their results under experi-mental conditions were not always consistent ie they did not always agreeTherefore the results obtained in this study are only broad recommendationsfor possible cheese and wine pairings In the light of the range of assessmentsobtained by knowledgeable judges it would be important to empower con-sumers to have the confidence to make their own assessment of how muchthey enjoy the cheese and wine pair

In conclusion white wines were easier to pair with a broader range ofcheeses than red or specialty wines The stronger more flavorful cheeses weremore difficult to pair with the wines but were more likely to be a good matchfor the late harvest and ice wines than other milder cheeses The results ofthis study were determined with a single tasting protocol and the wine andcheese were tasted without other foods The addition of other foods or agreater time period between the assessments would change the perception ofthe pairing Finally the variation in judge assessment indicates a high degreeof personal expectation or preference associated with the suitability of wineand cheese pairing Round-table discussion and consensus developmentmight produce ldquotrainedrdquo judges with less variation in the assessments althoughthis might be difficult given the ldquorespectedrdquo opinions of acknowledgedexperts Individuals should be encouraged to experiment to determine theirown preferences

ACKNOWLEDGMENTS

The authors would like to thank Reg Hendrickson of the Dairy Farmersof Canada for donating the cheese Christa-Lee Bond of the Okanagan WineFestival Society for allowing this research to be conducted as part of theIcewine Festival events Kimberly Dever for her technical assistance and theBC Winemakers and restaurateurs who participated in the study

REFERENCES

BALDY MW 1995

The University Wine Course

p 426 Wine AppreciationGuild San Francisco CA

BARTOSHUK LM 2000 Comparing sensory experiences across individu-als Recent psychophysical advances illuminate genetic variation in tasteperception Chem Senses

25

447ndash460COCHRAN WG amp COX GM 1957

Experimental Designs

p 611 JohnWiley amp Sons New York

256 M KING and M CLIFF

EampJ GALLO WINERY 2002 Gallo of sonoma wine and cheese pairingguide ABCrsquos of wine and cheese httpwwwwinewebcentralcomawineandcheeseabchtm (accessed October 24 2003)

JACKSON RS 2002

Wine Tasting A Professional Handbook

pp 261ndash262Academic Press San Diego CA

MOESSNER M Inniskillinrsquos guide to wine and cheese pairings httpwwwwineandleisurecomcheesepairinghtml (accessed October 242003)

NOBLE AC 1996 Tastendasharoma interactions Trends Food Sci Technol

7

439ndash444

PHILLIPS R 2000

A Short History of Wine

Penguin Books Londonhttpwwwintowinecomshop1282-0excerpthtml (accessed October24 2003)

SHEPHERD R SMITH K and FARLEIGH CA 1989 The relationshipbetween intensity hedonic and relative-to-ideal ratings Food QualPrefer

1

75ndash80SMITH J 1995

Cheesemaking in Scotland ndash A History

Scottish DairyAssociation Scotland httpwwwefrhwacukSDAbook1html(accessed October 24 2003)

Page 11: EVALUATION OF IDEAL WINE AND CHEESE PAIRS USING A DEVIATION-FROM-IDEAL SCALE WITH FOOD AND WINE EXPERTS

WINE AND CHEESE PAIRS 255

experienced in food and wine pairing as a group their results under experi-mental conditions were not always consistent ie they did not always agreeTherefore the results obtained in this study are only broad recommendationsfor possible cheese and wine pairings In the light of the range of assessmentsobtained by knowledgeable judges it would be important to empower con-sumers to have the confidence to make their own assessment of how muchthey enjoy the cheese and wine pair

In conclusion white wines were easier to pair with a broader range ofcheeses than red or specialty wines The stronger more flavorful cheeses weremore difficult to pair with the wines but were more likely to be a good matchfor the late harvest and ice wines than other milder cheeses The results ofthis study were determined with a single tasting protocol and the wine andcheese were tasted without other foods The addition of other foods or agreater time period between the assessments would change the perception ofthe pairing Finally the variation in judge assessment indicates a high degreeof personal expectation or preference associated with the suitability of wineand cheese pairing Round-table discussion and consensus developmentmight produce ldquotrainedrdquo judges with less variation in the assessments althoughthis might be difficult given the ldquorespectedrdquo opinions of acknowledgedexperts Individuals should be encouraged to experiment to determine theirown preferences

ACKNOWLEDGMENTS

The authors would like to thank Reg Hendrickson of the Dairy Farmersof Canada for donating the cheese Christa-Lee Bond of the Okanagan WineFestival Society for allowing this research to be conducted as part of theIcewine Festival events Kimberly Dever for her technical assistance and theBC Winemakers and restaurateurs who participated in the study

REFERENCES

BALDY MW 1995

The University Wine Course

p 426 Wine AppreciationGuild San Francisco CA

BARTOSHUK LM 2000 Comparing sensory experiences across individu-als Recent psychophysical advances illuminate genetic variation in tasteperception Chem Senses

25

447ndash460COCHRAN WG amp COX GM 1957

Experimental Designs

p 611 JohnWiley amp Sons New York

256 M KING and M CLIFF

EampJ GALLO WINERY 2002 Gallo of sonoma wine and cheese pairingguide ABCrsquos of wine and cheese httpwwwwinewebcentralcomawineandcheeseabchtm (accessed October 24 2003)

JACKSON RS 2002

Wine Tasting A Professional Handbook

pp 261ndash262Academic Press San Diego CA

MOESSNER M Inniskillinrsquos guide to wine and cheese pairings httpwwwwineandleisurecomcheesepairinghtml (accessed October 242003)

NOBLE AC 1996 Tastendasharoma interactions Trends Food Sci Technol

7

439ndash444

PHILLIPS R 2000

A Short History of Wine

Penguin Books Londonhttpwwwintowinecomshop1282-0excerpthtml (accessed October24 2003)

SHEPHERD R SMITH K and FARLEIGH CA 1989 The relationshipbetween intensity hedonic and relative-to-ideal ratings Food QualPrefer

1

75ndash80SMITH J 1995

Cheesemaking in Scotland ndash A History

Scottish DairyAssociation Scotland httpwwwefrhwacukSDAbook1html(accessed October 24 2003)

Page 12: EVALUATION OF IDEAL WINE AND CHEESE PAIRS USING A DEVIATION-FROM-IDEAL SCALE WITH FOOD AND WINE EXPERTS

256 M KING and M CLIFF

EampJ GALLO WINERY 2002 Gallo of sonoma wine and cheese pairingguide ABCrsquos of wine and cheese httpwwwwinewebcentralcomawineandcheeseabchtm (accessed October 24 2003)

JACKSON RS 2002

Wine Tasting A Professional Handbook

pp 261ndash262Academic Press San Diego CA

MOESSNER M Inniskillinrsquos guide to wine and cheese pairings httpwwwwineandleisurecomcheesepairinghtml (accessed October 242003)

NOBLE AC 1996 Tastendasharoma interactions Trends Food Sci Technol

7

439ndash444

PHILLIPS R 2000

A Short History of Wine

Penguin Books Londonhttpwwwintowinecomshop1282-0excerpthtml (accessed October24 2003)

SHEPHERD R SMITH K and FARLEIGH CA 1989 The relationshipbetween intensity hedonic and relative-to-ideal ratings Food QualPrefer

1

75ndash80SMITH J 1995

Cheesemaking in Scotland ndash A History

Scottish DairyAssociation Scotland httpwwwefrhwacukSDAbook1html(accessed October 24 2003)