Skip to main content
Search form
  • 00:00

    [MUSIC PLAYING][ShortCuts -- making life easier][Sampling]

  • 00:11

    DR. STEVE TAYLOR: One of the ways doctors find outabout the state of our health is to take some of our bloodand just a small quantity is usually enough to tell themwhat's going on.Well, it's much the same in psychological research.If psychologists are studying autism, for example,from the hundreds and thousands of autistic children,they'll select a relatively small number to study.

  • 00:32

    DR. STEVE TAYLOR [continued]: This is called sampling.

  • 00:34

    HARRIET CARMICHAEL: But first, make sureyou're not going to get tripped upby any of the technical terms.The target population is all the peoplewho might have been studyied.The sample is the segment of that populationwho were actually studied.Representativeness is about the extentto which the sample reflects the target population,

  • 00:56

    HARRIET CARMICHAEL [continued]: and generalizability is about the extent to whichthe data from the sample can be applied to the targetpopulation.So the more representative the sample,the more researchers feel they can generalize.It's like if you're judging a bake-off competition,sampling all the ingredients is enough to tell youabout the cake.

  • 01:17

    HARRIET CARMICHAEL [continued]: You don't have to eat it all.So we've said what sampling is.The next thing is to check out how it's done.There are different sampling techniques,and it's important not to confuse them.And a key distinction that will help you hereis between probability and non-probability sampling.

  • 01:35

    DR. STEVE TAYLOR: Probability samplesare selected in such a way as to try to make them representativeof the target population.One way of doing this you'll probably knowis random sampling, where every member of the target populationhas an equal chance of being selected.So it's the same principle as the lottery.

  • 01:53

    HARRIET CARMICHAEL: Another type of probabilitysampling you need to make sure youknow is stratified sampling.

  • 01:59

    DR. STEVE TAYLOR: In stratified sampling,selection isn't just left to chance.It's manipulated to try to ensure a better reflectionof the target population.For example, boys with autism outnumber girlsby almost five to one.So a stratified sample would have five times as many boysas girls.

  • 02:17

    HARRIET CARMICHAEL: The great strengthsof sampling methods like random and stratified samplingis that it's more likely that the sample willbe representative of the target population.[Probability Sampling Strengths] Thismeans you can generalize from it, [generalisation]and it gives scope for measurement. [measurement]For example, researchers can contact samples of a target

  • 02:38

    HARRIET CARMICHAEL [continued]: population in an area and calculatethe prevalence of certain types of behavior, illnesses,who's been a victim of crime and so on.And because the sampling is done consistentlyor systematically, there is reliability. [reliability]For example, you can resample and seeif a rate has risen or fallen.

  • 03:00

    DR. STEVE TAYLOR: These methods have a lot going for themin principle, but in the real world,they're often difficult or impossible to do in practice.[Probability Sampling Limitations] You need verylarge numbers, and this means it can take a long time and bevery expensive. [large numbers needed] Asparticipants have not volunteered,there's often a large rate of attrition,which can compromise the representativeness of the data.

  • 03:21

    DR. STEVE TAYLOR [continued]: [high rate of attrition] And for many research problems,probability sampling is neither practical nor possible.[may not be practical/possible] So researches have to use otherways of sampling, which brings us to non-probability sampling.

  • 03:34

    HARRIET CARMICHAEL: This simply means sampling whererepresentativeness cannot be inferred.There are a number of ways researchers can do this,and here are three.Opportunity or convenience samplingis where the researcher simply askspeople if they'd take part. [opportunity/conveniencesampling] This is why about 75% of participants

  • 03:55

    HARRIET CARMICHAEL [continued]: in psychological research are students.

  • 03:58

    DR. STEVE TAYLOR: Another method you could bring inis purposive sampling. [purposive sampling] Here,the research just wants a certain categoryof participants.

  • 04:06

    HARRIET CARMICHAEL: A third method is self-selectedsampling, where the researcher simply advertisesfor volunteers, as Philip Zimbardo did for his StanfordPrison Experiment. [self-selected sampling]

  • 04:16

    DR. STEVE TAYLOR: There's a numberof reasons why psychologists might use non-random sampling.First, it's generally much quicker and cheaper,[quicker and cheaper] and if participants have volunteered,they're likely to be more committed,and providing their consent is informed,there's less likely to be ethical issues,[greater commitment] and in most cases,the psychologist has no choice anyway but to usenon-probability sampling. [no choice]

  • 04:39

    HARRIET CARMICHAEL: Of course, the major limitationsof non-probability sampling, like convenienceand self-selected sampling is that it'sunlikely to be representative.[Non Probability Sampling Limitations] [unlikely to berepresentative] And also, that researcher or participant biascould creep into the sample. [researcher or participantbias] And therefore, it's more difficult to generalizethe data to a wider population. [harder to generalise from]

  • 05:01

    CLAIRE PARSONS: So that's the back story of sampling,and it's really important to knowbecause it's a great way to evaluate any study that'sbased on sampling.[Sampling-- Evaluation] First, what was being measured?[concept being measured?] Differently defined conceptsand different questions will give different measurements.Second, what was the sample size?[sample size?] Was is big enough to justify the conclusions

  • 05:22

    CLAIRE PARSONS [continued]: being drawn?And third, was it a representativeor non-representative sample? [representative or nonrepresentative] A major error here, even madeby some researchers, is to try to generalize findingsfrom non-representative samples to wider target populations.

Video Info

Series Name: Core Concepts in Research

Publisher: ShortCutstv

Publication Year: 2017

Video Type:Tutorial

Methods: Sampling, Polling, Survey research

Keywords: practices, strategies, and tools

Segment Info

Segment Num.: 1

Persons Discussed:

Events Discussed:



Sampling is selecting a small number of of the population to study. The selection needs to be a representative sample for accurate findings.

Looks like you do not have access to this content.


Sampling is selecting a small number of of the population to study. The selection needs to be a representative sample for accurate findings.

Copy and paste the following HTML into your website