psychology 242, dr. mckirnan

38
1 Foundations of Research Psychology 242, Dr. McKirnan Defining your target population Probability & Non- Probability sampling methods. Research Sampling. Please run this as a PowerPoint Show Go to “slide show” and click “run show”. Click through it by pressing any key. Focus & think about each point; do not just passively click. Dr. David J. McKirnan, University of Illinois at Chicago, Psychology; [email protected]

Upload: josef

Post on 24-Feb-2016

47 views

Category:

Documents


2 download

DESCRIPTION

Psychology 242, Dr. McKirnan. Defining your target population Probability & Non-Probability sampling methods. Research Sampling. Please run this as a PowerPoint Show G o to “slide show” and click “run show”. Click through it by pressing any key. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Psychology 242, Dr. McKirnan

1Foundations of Research Psychology 242, Dr. McKirnan

Defining your target population

Probability & Non-Probability sampling methods.

Research Sampling.

Please run this as a PowerPoint Show Go to “slide show” and click

“run show”. Click through it by pressing

any key. Focus & think about each

point; do not just passively click.Dr. David J. McKirnan, University of Illinois at

Chicago, Psychology; [email protected]

Page 2: Psychology 242, Dr. McKirnan

2Foundations of Research The big picture: Research sampling

Define your target population What group do you want to generalize to? How is / is not a member of the group? What is your sampling frame?

Page 3: Psychology 242, Dr. McKirnan

3Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Sampling

Sampling: Who do you want to generalize to?

Any study assesses only a sample of the population.

There are many different ways we may collect a sample.

There are many different populations or sub-populations we may be interested in.

The size and breadth of a sample can affect the Internal or External validity of the study.

Page 4: Psychology 242, Dr. McKirnan

4Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Define the target population

MammalsHumans

All Western peopleAll Americans

Young AmericansCollege students

UIC StudentsThis class

Breadth of population to sample

from (i.e., size of sampling frame).

Represents increasing

external validity.

Specificity (and ease) of sampling

frame.Generally increases internal validity.

Who do you want to generalize to?

Page 5: Psychology 242, Dr. McKirnan

5Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Who do you want to generalize to?

Samples are often comprised of very targeted sub-populations

Demographic groups; Ethnicity Socio-economic status Geography; e./g., urban dwellers…

Behavioral groups Registered voters Home owners

Clinical or other groups Medical or psychiatric patients…

That Specificity increases Internal

validity by decreasing the

complexity of the sample.

Page 6: Psychology 242, Dr. McKirnan

6Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Research samples & validityClinical drug trials illustrate the

conflict between internal v. external validity in sampling.

People with diverse symptoms and backgrounds see physicians for depression.

To enhance internal validity drug researchers use exclusion criteria to select only participants who fit a specific definition of depression

Zimmerman et al. suggest that too many exclusion criteria compromises the validity of this research area. (click

image for article)

Zimmerman, M.l, Mattia, J.I., & Posternak, M.A. (2002). Are Subjects in Pharm-acological Treatment Trials of Depression Representative of Patients in Routine Clinical Practice? Am J Psychiatry, 159, 469–473.

E X

A M

P L

E

Page 7: Psychology 242, Dr. McKirnan

7Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Exclusion criteria & validity

The study begins with a large # of people self-referred for depression

They exclude those with serious mental illness, drug abuse or personality disorder…

…whose symptoms are not severe enough, are suicidal, or who have other affective disorders..

…whose symptoms are too recent OR too long-standing…

…and end up with a small, carefully selected sub-set of patients (8.4% of general depression patients).

E X

A M

P L

E

Page 8: Psychology 242, Dr. McKirnan

8Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

External vs. internal validity in sampling

Applying rigorous study selection criteria for drug trials excludes the great majority of routine depression patients.

Rigorous participant selection for internal validity seriously compromises external validity in these studies.

This leaves the actual usefulness of anti-depressant (and other) medications for the general population in doubt.

To be useful research must balance the need for careful subject selection with the need for representativeness

E X

A M

P L

E

Page 9: Psychology 242, Dr. McKirnan

9Foundations of Research

Psychology 242, Dr. McKirnan Week 3; Experimental designs

Do you use Facebook or other media 5 times a week or more?A = YesB = NoC = Not sure – lost count.

Who is a group member?

Page 10: Psychology 242, Dr. McKirnan

10Foundations of Research

Psychology 242, Dr. McKirnan Week 3; Experimental designs

Who is a group member?

Are you a “Facebooker”?

A = YesB = NoC = Not sure – let me

facebook that.

Page 11: Psychology 242, Dr. McKirnan

11Foundations of Research

Psychology 242, Dr. McKirnan Week 3; Experimental designs

Are you a Latino?

A = YesB = NoC = Maybe – I’m not sure

Who is a group member?

Page 12: Psychology 242, Dr. McKirnan

12Foundations of Research

Psychology 242, Dr. McKirnan Week 3; Experimental designs

Do you speak Spanish?

A = YesB = NoC = Maybe – I’m not sure

Who is a group member?

Page 13: Psychology 242, Dr. McKirnan

13Foundations of Research Define the target population

Who do you want to generalize to: who is in the group?

Once choosing our sampling group, we must decide on criteria for membership…

To sample “Facebook users”, do I use a … Behavioral criterion (which behavior?) Self-identification?

To sample “Latinos”… Is it enough to call oneself “Latino” Is Spanish language necessary…?

Using a behavioral criterion (amount of Facebook use) may yield a different sample than self-identification.

Clearer and narrower group

criteria increases Internal validity

by making the sample more

homogeneous.

Page 14: Psychology 242, Dr. McKirnan

14Foundations of Research

Criterion Demographic / Behavioral Self-Identification

“gay / lesbian” Sexual patterns Self-label

“Drug User” # of drugs used Perceived dependence

“Latino” Geographic origins. language Ethic identification

“Student” # Hours registered Occupational Choice

“Depression patient”

Specific profile of behaviors & symptoms

Self-referred for treatment

The criteria used to define the group will determine who specifically gets sampled.

Who do you want to generalize to?

Page 15: Psychology 242, Dr. McKirnan

15Foundations of Research

What is known about your larger population? Are there Census or survey data?

E.g., are there “population” data on depressed people? Do we know the demographic profiles of Facebook users?

Data about your target population will help you determine how well your sample represents that population.

What is its size, sub-groups, location…. Where / how can I best recruit members of the population Will different recruitment methods be biased in favor of some

sub-groups? E.g., internet surveys are biased against less computer-oriented

people.

Who do you want to generalize to: Your “Sampling Frame”.

Page 16: Psychology 242, Dr. McKirnan

16Foundations of Research

Psychology 242, Dr. McKirnan Week 3; Experimental designs

Research sampling

Defining your target population

Probability & Non-Probability sampling methods.

Page 17: Psychology 242, Dr. McKirnan

17Foundations of Research Major forms of sampling

Use available samples for convenience, or targeted outreach to unusual or small populations.

Selection may be either systematic or haphazard, but is not random.

Often the most externally valid approach to unusual, small, or extreme groups, or groups where little is known.

When used only for convenience it is the least externally valid.

Probability (Random) Sampling Recruit (or select) participants to maximize the

representativeness of the sample to a known population. Uses some form of random selection. Requires that each member of the population has a known (often

equal) probability of being selected. Most externally valid approach to sampling general populations

Non-Probability Sampling

Page 18: Psychology 242, Dr. McKirnan

18Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Probability / Random Sampling

• Core feature: all members of the study population have an equal chance of being sampled

• Procedure: Choose participants in a systematic, random fashion.

• e.g., every nth name from a list, every nth number in a phone exchange, etc.

• Advantages: eliminates obvious biases of convenience sampling

• Limitations: • May under-sample unusual / hard to reach

participants• Some may be unavailable in, e.g., telephone lists,

computer files.

Page 19: Psychology 242, Dr. McKirnan

19Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Basic Forms of random sampling

• Simple Random Sampling: Select a specific % of a target population; all members of population have about equal chance of selection.

• Multi-Stage: Randomly select population units (census tracts, households, schools..), then randomly select individuals within unit.

• Stratified: Random within population sub-blocks, e.g., gender (randomly select 50 women and randomly select 50 men), ethnicity, etc.

• Cluster: Random within (potentially convenience) clusters, e.g., specific locations or “venues”, events, times of day, etc.

Page 20: Psychology 242, Dr. McKirnan

20Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Simple Random sampling

Objective: Attempts to truly represent the general population; absolute minimal selection bias.

Procedure: Recruitment method where all members of the population have ~ chance of being selected:

Advantages: Most representative sampling frame for general (non-targeted) population

Disadvantages: Any recruitment method excludes some people (no telephone, no stable address, etc.).

“Long form” of the census to a small % of U.S. households

Gallup polls using random digit dialing surveys

Examples:

Page 21: Psychology 242, Dr. McKirnan

21Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Multi-Stage Random samplingObjective: Focused & efficient random sample. Procedure: Concentrate recruitment in specific locations

or venues.

Advantage: Much more efficient that simple randomDisadvantage: Same as simple random

“CITY” HIV study among youth:1) Randomly select bars, clubs,

other venues across the city2) Randomly approach every 4th

person who enters the venue to recruit for interview

NIDA household drug surveys: 1) Random select moderate # of

census tracts nationally2) randomly select small % of

households within each tract;3) Interview 1st adult who answers

phone in each household

Examples:

Page 22: Psychology 242, Dr. McKirnan

22Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Stratified or cluster sampling

Objective: Represent every key segment of the population.

Procedure: Decide which population segments are important (e.g.

ethnic groups, census tracts, geographic areas...), Randomly select from each segment.

• Proportionate: Same sampling fraction from each segment; approximates overall population

• (e.g., sample 1% of all African-Americans, 1% of all Latinos, etc…)

• Disproportionate: Unequal sampling fraction across segments, to over-represent smaller groups

• (e.g., select larger % of recent immigrants…)

Page 23: Psychology 242, Dr. McKirnan

23Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Non-Probability Sampling

Useful for populations that:Cannot be randomly sampled; “hidden” or difficult to

reach

No sampling frame available, such as census data, describing its size, composition, etc.

Examples: drug users, gay men, homeless, etc.

Likely to misrepresent the populationMay be difficult or impossible to detect this misrepresentationOften over-sensitive to incentives: paying

participants attracts more poor people “Respondent Driven” sampling (RDS) allows for “targeted”

population estimates

Page 24: Psychology 242, Dr. McKirnan

24Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Problem: No evidence for representativenessAdvantage: availability of participants

Non-Probability methods (1)

Modal Instance Sampling; “Typical” case Typical New Yorker describing trade tower tragedy Typical voter.

Problem: May not represent the modal group.Advantage: Describe simple, “typical case”

Haphazard / Modal instance often used by journalists or qualitative-descriptive studies; see NYT “down low” article.

Haphazard Sampling; “Man on the street” College psychology majors Available medical / therapy clients Volunteer samples

Page 25: Psychology 242, Dr. McKirnan

25Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Non-Probability methods, 2

Sample a specific, well-defined, often hard to reach group Assume group members are well represented at

specific locations or settings (“venues”). Use “Intercept” methods for reaching participants

Use indigenous outreach workers from the population Develop a standard recruitment script Collect / distribute contact information for later participation

Time / Space randomization: Lessen bias due to choice of venue:

Randomly approach different venues at different times Randomly select participants within the venue (e.g., every 4th person…)

Strategy must be based on a clear epidemiological or theory question.

Examples: Shopping mall intercepts, gay recruitment

Venue & time / space Sampling

Page 26: Psychology 242, Dr. McKirnan

26Foundations of Research

Psychology 242, Dr. McKirnan Week 3; Experimental designs

Outreach / venue sampling: examples of palm cards

Page 27: Psychology 242, Dr. McKirnan

27Foundations of Research

Psychology 242, Dr. McKirnan Week 3; Experimental designs

Outreach lead sheet

Page 28: Psychology 242, Dr. McKirnan

28Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Non-Probability methods, 3

Sample a specific, hard to reach group No census or similar data for sampling frame. Uses multiple (convenience) sampling “frames”:

Direct outreach to places where population members are available (venue sampling).

Newsletters, internet lists & chat rooms Organizations or meeting places

Strategy must be based on a clear epidemiological or theory question.

Most common & valid convenience sample

People who have risky sex Homeless people…

“MTV” Market segments Shoplifters

Examples:

Targeted Multi-Frame Sampling

Page 29: Psychology 242, Dr. McKirnan

29Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Non-Probability methods (3)

Early participants are paid to recruit others, who recruit others, etc.

Form of targeted sampling: Recruit network of “linked” people tracked by referrals Problem: Advantage: Access unusual or “hidden” people related

by a common behavior. With enough “generations” of links can well represent a

target population.

With RDS can show “chain” of referrals / links. Often part of multi-frame approach. Useful for people who mistrust research or where personal

contact is necessary for recruitment (HIV, drug use). Portrays “chain” of influence or, e.g., infectious disease.

Snowball / “Respondent Driven” Sampling (RDS)

Choice of seeds.Eligibility criteria

Sensitive to incentives!

Page 30: Psychology 242, Dr. McKirnan

30Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

RDS coupon examples

Heckathorn, D.D. & Magnani, R. (2004). Snowball and Respondent-Driven Sampling. In: Behavioral Surveillance Surveys: Guidelines for Repeated Behavioral Surveys in Populations at Risk of HIV

Page 31: Psychology 242, Dr. McKirnan

31Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

RDS; chain description

Heckathorn, D.D. & Magnani, R. (2004). Snowball and Respondent-Driven Sampling. In: Behavioral Surveillance Surveys: Guidelines for Repeated Behavioral Surveys in Populations at Risk of HIV.

Page 32: Psychology 242, Dr. McKirnan

32Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Example of social network sampling:Bearman et al., Romantic ties among adolescents

A substantial majority of students are in an extended, linked chain of relationships.

With a number of smaller chains

And a small % in 2 to 4 person chains From sampling

perspective, several “seeds” access most of the population

Findings suggest a clear potential for STI transmission.

Page 33: Psychology 242, Dr. McKirnan

33Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Non-Probability methods Quota Sampling

Select people non-randomly according to quotas Must have clear theory / research question to pick

relevant population characteristic(s). Proportional quota sampling

• Represent major characteristics of a population. If gender is important, and the proportion of women :: men in your population = 65% :: 35%, the sample must meet that quota.

Non-proportional quota sampling• Sample enough members of each group to test hypothesis,

even if the sample is not proportional. (e.g., recruit 50 women & 50 men, even though the real proportion is 65::35).

• Helps assure that you have good representation of smaller population groups.

Similar to cluster sampling, except you cannot randomly sample each population segment.

Page 34: Psychology 242, Dr. McKirnan

34Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Non-Probability methods Web sampling Typically highly targeted samples

Gay / bisexual men… Adolescents… “Gamers”…

Typically access through existing venues: Users of specific web sites List-serves, e-mail lists Active recruitment in “chat rooms”

Problem: Inherent bias in computer literacy(?)Advantage: Cheap large national sample

Access unusual or “hidden” people who reach others via internet

Page 35: Psychology 242, Dr. McKirnan

35Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Non-Probability methods; Heterogeneity Sampling

• Sample every sector of a population -- at least several of everyone -- without worrying about proportions.

• At least some members of each geographic area• …ethnic group• …behavioral group (voters & non-voters…)

• Assume that a few people are a good proxy for the group.

Problem; Cannot be sure a few people really represent their sub-group.

Advantage: At least some representation of all sub-groups.

Examples: focus groups or qualitative interviews about products, social issues...

Page 36: Psychology 242, Dr. McKirnan

36Foundations of Research

Psychology 242, Dr. McKirnan Week 6; Sampling

Sampling overviewWho do you want to generalize to?

Who is the target population? broad – external validity narrow – internal validity

How do you decide who is a member? demographic / behavioral criteria? subjective / attitudinal?

What do you know about the population already – what is the “sampling frame”.

Is a Probability or random sample possible? “Hidden” population? Socially undesirable research topic? Easily available via telephone, door-to-door? Sampling frame adequate to choose selection method?

Sum

mar

y

Page 37: Psychology 242, Dr. McKirnan

37Foundations of Research

Psychology 242, Dr. McKirnan Week 3; Experimental designs

Types of Non-probability Samples Haphazard Modal instance Venue – time / space Multi-frame Snowball / Respondent driven Web Quota Heterogeneity

Overview, 2Su

mm

ary

Page 38: Psychology 242, Dr. McKirnan

38Foundations of Research Overview, 3

Non-probability sampling targeted / multi-frame snowball quota, etc.

Probability sampling simple multi-stage cluster or stratified

Most externally valid Assumes:

Clear sampling frame Population is available

Less externally valid for hidden groups.

Less externally valid High “convenience” Best when:

No clear sampling frame Hidden / avoidant

population.

Sum

mar

y