PYSC 224 Introduction To Experimental Psychology

PYSC 224Introduction to ExperimentalPsychologySession 6 – Quasi Experiments andFaulty Experimental Designs Part 1 & 2Lecturer: Dr. Margaret Amankwah-Poku, Dept. of PsychologyContact Information: mamankwah-poku@ug.edu.ghCollege of EducationSchool of Continuing and Distance Education2015/2016 – 2016/2017

Session Overview Two other kinds of experimental research design exist; quasiexperimental design and faulty experimental design Quasi experimental designs are similar to true experimentaldesigns but in quasi experiments, the experimenter lacks thedegree of control over the conditions that is possible in a trueexperiment Some research studies may necessitate the use of quasidesigns rather than true experimental designs Faulty experimental design on the other hand are designswhich you should avoid when doing research. They either donot include a control group and/ a pretestDr. Margaret Amankwah-PokuSlide 2

Session Goals and ObjectivesAt the end of this session, you should be able to Define a quasi experiment and distinguish it from atrue experiment Determine when to use a quasi design Outline the types of quasi experimental design Define what a faulty experimental design is anddistinguish it from a true experiment Outline the types of a faulty experimental designsDr. Margaret Amankwah-PokuSlide 3

Session OutlineThe key topics to be covered in the session are as follows: Quasi Experimental Designs Faulty Experimental DesignsDr. Margaret Amankwah-PokuSlide 4

Reading List Christensen, B.L. (2007). Experimental Methodology(10th ed.). Boston: Allyn & Bacon. (p. 330-335) Kantowitz, B. H., Roediger III, H. L., Elmes, D. G.(2015). Experimental Psychology. Stamford: CengageLearning (Pp 81)Dr. Margaret Amankwah-PokuSlide 5

Topic OneQUASI EXPERIMENTAL DESIGNDr. Margaret Amankwah-PokuSlide 6

Quasi Experiments Quasi means “as if “or “to a degree”(McBurney (2001) Quasi experiment- ‘an experiment that doesnot meet all the requirements necessary forcontrolling the influence of extraneousvariable’ (Christensen, 2007, p. 330)Dr. Margaret Amankwah-PokuSlide 7

Quasi Experiment Participant variables such as age, sex, etc.cannot be experimentally manipulated Participants cannot be randomly assigned to aparticular condition Instead they are exposed to a conditionbecause they already qualify for that conditionDr. Margaret Amankwah-PokuSlide 8

Quasi Experiment The experimenter can only select instancesthat satisfy different categories and studythem Use natural manipulations rather thanmanipulations by the researcher ‘Manipulation’ could be naturally occurringattribute or events- e.g. ethnic group, age,sex, intelligence etc.Dr. Margaret Amankwah-PokuSlide 9

Quasi Experiment Sometimes called ‘ex post facto’ or ‘after thefact’ experiments The experiment is conducted after thegroups have been formed The inability to control confounding variablesreduces the internal validity of a quasiexperiment but does not make it invalidDr. Margaret Amankwah-PokuSlide 10

Quasi Experiment Note that it is possible to have oneexperimental variable and one quasi variable inone experiment Example- investigating the effect of practice onperformance will require a true experimentaldesign The effect of practice and gender onperformance will require a combination of atrue and quasi experimentDr. Margaret Amankwah-PokuSlide 11

Quasi ExperimentsDifference between true experiment and quasiexperiments Major difference- random assignment ofparticipant to experimental groups in trueexperiments but not quasi experiments Experimenter manipulates variables in trueexperiments but observes categories ofparticipants in quasi experimentsDr. Margaret Amankwah-PokuSlide 12

Quasi Experiments True experiments- Control permits theestablishment of cause effect relationship Quasi experiment- cannot be determinedwhether difference in behaviour is caused bythe independent variable or the differencebetween the two groups Generally, true experiments are preferred to aquasi experimentDr. Margaret Amankwah-PokuSlide 13

Topic TwoTYPES OF QUASI EXPERIMENTALDESIGNSDr. Margaret Amankwah-PokuSlide 14

Quasi Experimental Designs Two types of quasi experimental designs 1. Non-equivalent control group design 2. Non-equivalent Pretest-posttest designDr. Margaret Amankwah-PokuSlide 15

1. Non-equivalent control groupdesign The performance of an experimental groupis compared with that of a non-equivalentcontrol group There is no random assignment of subjectsso the groups are not equivalentDr. Margaret Amankwah-PokuSlide 16

1. Non-equivalent control groupdesign Thus, difference between the groupsbecomes a confounding variable This provides an alternative explanation forthe results obtained Testing non-equivalent groups of participantscreates selection differenceDr. Margaret Amankwah-PokuSlide 17

1. Non-equivalent controlgroup design Effect of stress management on hypertensive pDr. Margaret bleMeasureDependent VariableMeasureDependent VariableSlide 18

1. Non-equivalent control groupdesign Example-The effect of stress management onhypertensive patients An experimenter may recruit ‘willing’hypertensive patients who serve as theexperimental group and are administeredstress management The experimenter later recruits morehypertensive patients to serve as the controlgroupDr. Margaret Amankwah-PokuSlide 19

1. Non-equivalent control groupdesign Thus, the two groups are not equated Differences found between the two groups,may be as a result of the stress managementor difference between the two groups This design is similar to the true experimentposttest-onlyDr. Margaret Amankwah-PokuSlide 20

2. Non-equivalent pretestposttest design An experimental and control group arepretested and posttested However, the two groups are not equivalent Differences between the groups, may be as aresult of the initial difference or the effect ofthe independent variableDr. Margaret Amankwah-PokuSlide 21

2. Non-equivalent pretest- posttestdesign Effect of stress management on diabetes Dr. Margaret riablePosttest: MeasureDependent VariableNoIndependentVariablePosttest: MeasureDependent VariableSlide 22

2. Non-equivalent pretestposttest designExperimentalGroupControlGroupDr. Margaret Amankwah-PokuPretest Score70%Pretest Score40%StressManagementNo StressManagementSlide 23PosttestScore30%PosttestScore35%

Topic ThreeFAULTY EXPERIMENTAL DESIGNSDr. Margaret Amankwah-PokuSlide 24

Faulty Experiment Faulty experimental designs usually do nothave a control group or do not include apretest There is internal invalidity as a result of notminimizing extraneous variables They should therefore be avoided byresearchersDr. Margaret Amankwah-PokuSlide 25

Faulty Experiment Two types of faulty research designs 1. One group posttest-only design 2. One group pretest-posttest designDr. Margaret Amankwah-PokuSlide 26

1. One group posttest-only design An experimental group is administered theindependent variable and then tested on thedependent variable There is no comparison group or pretestmeasure At a minimum, participants should bepretested Also, an equivalent control group must beincludedDr. Margaret Amankwah-PokuSlide 27

1. One group posttest-onlydesign The effect of type of Music on MoodExperimentalGroupDr. Margaret Amankwah-PokuIndependentVariable(listen tomusic)Posttest:Measure DependentVariable(Mood)Slide 28

2. One-group pretest-posttestdesign The effect of an independent variable isinferred from the pretest-posttestdifferences in a single group This design is faulty because it does not takeinto account extraneous variables such ashistory, maturation, etc. Also, there is no control group to comparethe experimental group withDr. Margaret Amankwah-PokuSlide 29

2. One group pretest-posttestdesignEffect of stress management on diabetes controlExperimentalGroupDr. Margaret isterIndependentVariableSlide 30Posttest:MeasureDependentVariable

One-Group pretest-posttest design Example- A block manufacturing factory decidesto change their work schedule from 8 hours a dayto 10 hours a day If there is improvement in productivity it doesnot definitely mean this is because of theincrease in the number of hours Other potential causes such as favourableweather condition, attention paid to employeeswhen the change was initiated etc. can alsoincrease productivityDr. Margaret Amankwah-PokuSlide 31

END OF SESSION 6 Part 1 & 2

References Christensen, B.L. (2007). ExperimentalMethodology (10th ed.). Boston: Allyn & Bacon. (p.83-90, 305-309, 321-322)Dr. Margaret Amankwah-PokuSlide 33

Quasi experimental designs are similar to true experimental designs but in quasi experiments, the experimenter lacks the degree of control over the conditions that is possible in a true experiment Some research studies may necessitate the use of quasi designs rather than true experimental designs Faulty experimental design on the .