j m laidlaw uantitative research

Forward

Designed and produced by theEducation Development UnitNHS Education for Scotland

Initial development sponsored by theScottish Higher Education Funding Council

QQQQQ uantitative Research

Authors

E A Hesketh

J M Laidlaw

Instructional Design

J M Laidlaw

Desktop Publishing

L E BellM C Gunn

Computing

N K McManus

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units SSSSS nippets on quantitative research

Quantitative research tends to focus onmeasurement and proof. This researchadopts a ‘scientific approach’. It is basedon the premise that something ismeaningful only if it can be observed andcounted. Its key characteristics are numerical data thatpermits a range of statistical analysis. There are severalapproaches to quantitative research which includeexperimental, descriptive, correlational and causalcomparison. Inferential statistics are frequently used togeneralise what is found about the study sample to thepopulation as a whole.

Quantitative Research... Snippets: 1 of 2

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units

Quantitative Research... Snippets: 2 of 2

””

““

The Problem The Solution

All these different research methods –experimental, descriptive, correlationaland more – which do I choose?

I’m put off doing research – myknowledge of statistics isn’t that hot!

Research methodology is a big subjectarea. If your skills need refreshed fromyour undergraduate days, this unit willgive you an overview of quantitativeresearch and start you off in the rightdirection.

Quantitative research concentrates onthe collection and analysis of numericaldata. Statistics, although not an integralpart of a quantitative study, are usedto confirm or contradict theconclusions you draw from youranalysed data. Getting the choice ofstudy type right, then getting advicefrom a statistician at the planning stagewill help.

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units IIIII ntroducing quantitative research

Research is often described as either quantitative or qualitative.Studies which involve the collection and analysis of numericaldata are known as quantitative research. Quantitative researchadopts the scientific method and focuses on controllingvariables, gathering measurable evidence and coming togeneralisable conclusions or providing new explanations.

A typical quantitative study will specify the number ofparticipants from a defined population and present data on, forexample, the number of events occurring or some biologicalvariable. It will include a statistical section which may showcause and effect relationships. Many research studies have anelement of both quantitative and qualitative data. Read the‘Qualitative research’ unit to find out when this approach is moreappropriate.

Quantitative Research... Core: 1 of 13

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units QQQQQ uantitative research in action

The aspects listed below take you through the various steps of theresearch process.

Selecting and definingthe research question

Deciding on study type

Deciding datacollection tools

Selecting the sample

Sample size

Analysing, interpretingand validating findings


Click the appropriateoption if you wish tofind out more

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units


Read up on the subject to find out if yourquestion has already been answered. Jotdown the key concepts and concerns thenform these into a meaningful question.

Ask yourself:

Can I answer this through datacollection and analysis?

Will this contribute to knowledge inthe subject area?

Show this to colleagues, if their reactionis ‘Who cares’, then think again.





Sample size


??

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

UnitsSome research problems suggest onetype of strategy rather than another. Forexample:

If you want to know the effect of a drug

or treatment

then a tightly controlled experimentalmethodology is appropriate.

If, however, you want to know whethersome disease or symptom is related toage, smoking or some other life style

factor

then a correlational approach isappropriate.

Four different strategies are now brieflydescribed. These are Experimental,Descriptive, Correlational and CausalComparison.

Experimental...





Sample size



ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

UnitsExperimental

In an experimental study an intervention is made under controlledconditions with the sole purpose of evaluating its effect. The goldstandard is the ‘Randomised Controlled Trial’ or RCT, mostcommonly used in clinical trials to assess whether a therapeuticintervention is effective.

A randomised selection of the population is split into the followingtwo groups:

and

Blinding, where the allocation of the intervention isconcealed, is often used. Valid outcome measures areidentified and assessed.

Descriptive...


one group being giventhe new treatment

or intervention

the other the controltreatment or no

intervention

Note...

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

UnitsDescriptive

Descriptive research involvescollecting data to test a hypothesis. Itis commonly used in epidemiologicalstudies to look at the incidence ordistribution of disease, or assessingcurrent health problems and needs.Data is usually collected throughquestionnaires.

Correlation...


ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units

Note...

Correlational

This method is used to discoverrelationships between two variables, eg isthere a link between birth weight and thedevelopment of disease over time?

The down side of this approach is that although links canbe made and insight gained into the strength of the link,direct cause and effect relationships cannot beconsidered proven.

Causal comparative research...


ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units Causal comparative research

If a cause and effect relationship is beinginvestigated, eg truancy and exam results; familyincome and health etc. the researcher canobserve the status quo without intervention. Theindependent variable (truancy, income) havingalready occurred, cannot be manipulated.


ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units

Your chosen tool will be determined by theinformation you want to collect.

Typical data for a quantitative study maybe morbidity and mortality trends, hospitaladmissions or prescribing patterns.

Data sources will be:

! existing documents

! patient records

! surveys and questionnaires.

Questionnaire design is anart in itself. Use an existingdesign or take advice.





Sample size



Note...

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units

Define the population you are interestedin. The sample should be representativewith minimum bias. To learn more aboutcommonly used sampling procedures,cllck on the following:

Random

Non-random





Sample size



ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units Random

Simple : using published random number tables everyone in thepopulation has an equal chance of being selected.

Systematic: a sampling interval is calculated based on the populationsize and the required sample size. Then every nth subjectis selected.

Stratified: sub-groups which may influence the outcome of a study,eg sex, age, ethnicity, are identified and a random sampletaken.

Non-random

Purposive: subjects typical of the characteristics under investigationare chosen, eg patients from practices with a highincidence of a particular health problem.

Volunteers: there is a danger of creating a biased sample but may be anecessary option where legal or ethical issues arise, egdrug testing.


ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units

Sample size is important, too big increasescosts, too small and you don’t havesufficient data to reach any meaningfulconclusions. Take advice from astatistician who will help you decide thenumbers required to give validity to yourresults.





Sample size



ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units

Decide at the planning stage whatstatistical tests will be used and thus thestatistical database package best suitedto your needs. Look for patterns,correlation’s and cause and effectrelationships in the data.





Sample size



ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units EEEEE xamples in practice

Here are examples in practice of two of the approachescommonly used.

Correlational

Experimental

Quantitative Research... Examples: 1 of 3

Click the appropriateoption if you wish tofind out more

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units


All entrants fromfour consecutive years

Academic performanceprior to admission recorded

Students annual Grade Point Average recordedGPA based on exam (with OSCE year 5) and written

coursework (years 1-5) and ward reports (year 6)

Affirmative action scheme forethnic minorities excluded

Data analysisSome of the correlations examined and findings:

! relationship between overall achievement marksin school leaving exams and performance inmedical school - but the level of correlationdecreased as students progressed through thecourse.

! the lack of correlation between overall marks inschool leaving exam and the OSCE results.

! no correlation between total school leavingmarks and clinical competence as measuredduring the ward clerkship.

▼

▼

Correlational

A study on the relationship betweenacademic requirements andsubsequent performance in medicalschool.

Taken from Collins JP, White GR &Kennedy JA (1995). Entry to medicalschool: an audit of traditionalselection requirements. MedicalEducation 29: 22-28.

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units


Subjects excluded:!!!!! exclusion criteria, eg those with viral

illness, already had bed rest etc.!!!!! consent not given to join study.

▼

Experimental

A Randomised Control Trial on BedRest for Acute Lower Back Pain

Taken from Wilkinson MJB (1995).Does 48 hrs bed rest influence theoutcome of acute low back pain?British Journal of General Practice45: 481-4. Control group

no day time restIntervention group48 hours bed rest

Subjects enter trial▼

Nine general practitioners selected for study.Patients present with acute low back pain.

7 and 28 day assessment:!!!!! change in straight leg raise and lumber flexion (7 days)!!!!! disability scores!!!!! days off work.

▼

Data analysis:!!!!! differences between groups detected at 7 days etc.!!!!! both groups reached similar disability scores by 28

days etc!!!!! no advantage of 48 hours rest over no bed rest shown.

▼

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units

If you can’t think of a research idea ask forsuggestions or advice from your line manager.They may have ideas that would be useful toyour unit/organisation.Alternatively look at previous research work.Many research articles suggest areas thatneed further work. Remember most research isabout building on the work of others.Let your idea mature for a little while and then

talk it through with a colleague. They may helpyou to see some of the pitfalls and assist youto hone in on your research question.Make sure you get approval from any relevant

Ethics Committee and that the data is storedaccording to the Data Protection Act.

PPPPP ractical tips

continued. . .

Quantitative Research... Practical Tips: 1 of 3

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units

Take advice on statistics, sampling etc at anearly stage of planning.Always pilot your data collection tools and tryto analyse the data. This gives you theopportunity to improve on their effectivenessor, if necessary, reduce the scope of what youhad planned to do.Questionnaires - Avoid ambiguous words likeregular, normal, recently – they mean somethingdifferent to every one and need to be defined egtwice a week, within the last two weeksCheck the consistency of answers in your

questionnaires through asking one or twoquestions again with different wording.


continued. . .

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units


Always use the largest sample possible withinyour budget.For research studies that cover a fairly longperiod of time, make allowances for drop out inyour plans. Minimise this drop out throughmaking minimum demands on your subjects,keeping them informed and reminding them ofthe important part they play in your study. Forquestionnaires be prepared to do up to threemailings to get a good response rate.

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units OOOOO ther learning opportunities

There are many books and other resources on Quantitative Research.Here is a short resume of our suggestions.

Click the appropriatebutton if you wish tofind out more

Quantitative Research... Opportunities: 1 of 5

Books

Journals

Web

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units

Kumar R (1999) Research Methodology - A step by step guide forbeginner. Sage Publications, London

Blaxter L, Hughes C & Tight M (1996). How to research. OpenUniversity Press, Milton Keynes

Carter Y & Thomas C (eds) (1997). Research methods in primarycare. Radcliffe, Oxford

All these books take you through the research process fromdesigning your question to study design, analysis, interpretationand dissemination.


Books

Journals

Web

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units

Oxman AD, Sackett DL & Guyatt GH (1993). Users’ guides to themedical literature 1 How to get started. JAMA 270 2093-2095

The first of a series of articles dealing with critical appraisal skills.

Greenhalgh T (1998). Randomised controlled trials. British Journalof General Practice 48 1448-1449

A short interesting account of the history of RCTs and anintroduction to the shorthand terms commonly used in suchresearch. The article also stresses the importance of truerandomisation without which the benefits of an intervention maybe overestimated.


Books

Journals

Web

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units

The international database TimeLIT (Topics in Medical EducationLiterature) www.timelit.org gives free access to articles relatingto education in medicine, dentistry, nursing, patient health andthe professions allied to medicine.

Other useful websites for those interested in Medical Educationare that of the Association for the Study of Medical Educationwww.asme.org.uk and that of the Association for MedicalEducation in Europe www.amee.org

Quantititative Research... Opportunities: 4 of 5

Books

Study

Web

http://www.timelit.org/

http://www.asme.org.uk/

http://www.amee.org/

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units

You might now wish to look at the related unit on QualitativeResearch.

Qualitative ResearchClick the option totake you to a summaryof the unit

Qualitative Research... Opportunities: 5 of 5

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units GGGGG roup activities

Get together with about five colleagues andhave a go at investigating whether BodyMass Index (BMI) is related to the amountof junk food you eat. You will require scalesand a stadiometer. Take measurements ofeach group member’s height and weight.Then assess the frequency of their junkfood intake per week, where for example a bagof crisps or one small portion of cake could be counted as oneunit of junk food. Analyse your data by calculating each subject’sBMI. Plot BMI against junk food frequency and discuss the spreadof results on the graph and/or perform a correlational analysisusing a statistical package such as SPSS. Afterwards considerwhat other variables/factors could/should have been taken intoaccount to improve the study design.

ForwardBack

Snippets

Core

Examples

Practical Tips

Opportunities

Activities

Don’t Forget

Home

Units

Make sure your research aims andquestions are clear from the outset.Then ask an expert or experiencedperson for advice to ensure the rightmethods, tools and statistics areemployed.

j m laidlaw uantitative research

Documents