Skip to main content
SAGE
Search form
  • 00:00

    [MUSIC PLAYING][Experimental Design]

  • 00:11

    STEVE TAYLOR: When you're lookingat psychological statistics, it'simportant to distinguish between those that are just identifyingrelationships between things with variables and those thatare claiming more by suggesting that one variable is actuallycausing changes in another.And if the statistics are telling usthis bigger story of causality, then we

  • 00:32

    STEVE TAYLOR [continued]: need to know how the researchers movefrom correlation to causation.[Dr. Steve Taylor]

  • 00:36

    CLAIRE PARSONS: To try to establish causality meansresearchers using an experimental method thatenables them to test the effect of onething, an independent variable, on another,a dependent variable.But like most things in life, experiments don't just happen.Just as we have made choices about whatto wear for a special occasion, for example, so researchers

  • 00:60

    CLAIRE PARSONS [continued]: have to make choices when they'replanning an experiment--choices about how to organize your participants.[Claire Parsons]What do you want them to do?How are you going to allocate them to different conditions?And most important, how to controlfor their individual characteristicsinfluencing the research.The various options available are called research designs,

  • 01:20

    CLAIRE PARSONS [continued]: and that's what we're looking at here.

  • 01:22

    STEVE TAYLOR: We're going to illustrate experimental designswith a memory question we hope will be relevant to youstudents.What's the best time of day to revise for exams--morning or evening?So let's imagine we're doing an experiment to find out.Here, the independent variable is the time of daythe work's done, and the dependent variable

  • 01:42

    STEVE TAYLOR [continued]: is the learning outcome--how well you remember what you were looking at earlier.Well, one way we could do this is with a repeated measuresdesign.[Repeated Measures Design] Here, all participantsdo all the conditions.So [INAUDIBLE] give them a learning task in the morningand test them in the evening.Then in a second condition, they doa similar task in the evening and be tested next morning.

  • 02:05

    STEVE TAYLOR [continued]: So this would be a really good test of whether time of dayinfluences learning and remembering.The great strength of the repeated measures designis that because all participants do all the conditions,it controls the individual differences in participants.And the bonus is you'll need fewer of them,not only saving time and money, but also makingfollow up studies easier.

  • 02:25

    CLAIRE PARSONS: It does have those strengths,but because participants are doing a second test,this can lead to practice and order effects.They'll be more used to what's expected,and this could compromise and maybe confoundthe time of day variable.Participants may do better in a second testsimply because it is a second test.And with repeated measures there's

  • 02:47

    CLAIRE PARSONS [continued]: also more possibility of demand characteristics, whateverthe order of the tests, as participants doingmore than one test may begin to get wise to whatthe experiment's about and act accordingly--do what they think we want.But a way we could get round these problemsis to use a different design, an independent measures design.[Independent Measures Design] Here,

  • 03:07

    CLAIRE PARSONS [continued]: we'd randomly allocate our participantsinto two different conditions.So one group does the learning in the morningand gets tested in the evening while the other groupgets exactly the same material in the eveningand then gets tested in the morning.So all participants do the same test once,and it's the same for everyone.[Independent Measures Design Strengths]So the independent measures design

  • 03:27

    CLAIRE PARSONS [continued]: controls for practice and order effectsand also reduces the likelihood of demand characteristics.

  • 03:33

    STEVE TAYLOR: So the independent measures designdoes solve some problems, but it may lead us to others.For example, we'll have different participantsdoing the same test.[Independent Measures Design Weaknesses]And even if we allocate them randomly,there's still a chance of participant variablesinfluencing the results.And then there's time and cost implications.We'll need twice as many participants

  • 03:53

    STEVE TAYLOR [continued]: to get the same amount of data.

  • 03:55

    CLAIRE PARSONS: But there's another option we could use,a matched pairs design--matching up participants in key variables such as ageand gender, for example.[Matched Pairs Design] And then we could also do somepretests to try to match them in other ways,like learning abilities and IQ.Then one half of each pair would dothe test in the morning and the other half in the evening.

  • 04:16

    STEVE TAYLOR: [Matched Pairs Design Strengths]So using a matched pairs design, youwouldn't have the practice and order effectsproblem you could get in a repeated measures design,and it would also help control for individual differencesthat you might get in an independent measures design.

  • 04:29

    CLAIRE PARSONS: Controlling for individual differencesin advance doesn't make the effect of subject variablesgo away, but it can dilute their effect.

  • 04:36

    STEVE TAYLOR: [Matched Pairs Design Weaknesses]So in a matched pairs design, individual differencesmay be reduced, but not necessarily eliminated.And not only will we need more participants,but having to match and pretest themmeans we'd need to spend a lot more time with them,so there are time cost implications.

  • 04:54

    CLAIRE PARSONS: So all these experimental designshave their strengths and potential weaknesses.

  • 04:58

    STEVE TAYLOR: And what researchers have to dois work out what design works best for their experiment,and this will give us another wayof evaluating many of the studiesthat we read about in psychology.For example, in the famous Stroop experiments,the speed of identifying the color of the ink the word iswritten varies so much between individualsthat repeated measures is the only possible design that

  • 05:21

    STEVE TAYLOR [continued]: controls for these differences.

  • 05:22

    CLAIRE PARSONS: But Loftus and Palmer's famous experimenton eyewitness memory required different groupsof participants being given slightly differently wordedquestions, so this had to use an independent measures design.

  • 05:34

    STEVE TAYLOR: While Bandura's famous Bobo dollexperiments on aggression used a matched pairs design.Here, children were pretested and gradedfor levels of natural aggression from one to fivein advance of the experiment, and then theywere matched to control for individual differencesin aggression.

  • 05:51

    CLAIRE PARSONS: So what about our experiment?

  • 05:53

    STEVE TAYLOR: Well, in this study like Bandura's, a matchedpairs will probably be the best design becauseof the crucial importance of controllingfor subject variables, especiallydifferences in ability.But remember, all choices have potential limitations.

  • 06:09

    CLAIRE PARSONS: And if you're wondering, as a student, whatmight really be the best time to try and learn new stuff, well,real research has actually been done and foundthat learning seemed better when it was done in the evening.

  • 06:21

    STEVE TAYLOR: Psychologists have shownthat a lot of memory consolidationtakes place when we're asleep, so it seems that you reallycan learn when you're asleep.[MUSIC PLAYING]

Video Info

Series Name: Experimental Research Methods

Publisher: ShortCutstv

Publication Year: 2017

Video Type:Tutorial

Methods: Experimental design, Quasi-experimental designs

Keywords: practices, strategies, and tools

Segment Info

Segment Num.: 1

Persons Discussed:

Events Discussed:

Keywords:

Abstract

Learn about potential strengths and weaknesses of experiments and experimental design.

Looks like you do not have access to this content.

Experimental Design

Learn about potential strengths and weaknesses of experiments and experimental design.

Copy and paste the following HTML into your website