Chapter 8: Experimental Design

Front 1

researchers manipulates _________ to create groups that differ _______________ in levels of that variable and then compares the groups in terms of ____________________

Back 1

(1) the independent variable

(2) in the levels of that variable

(3) their scores on dependent variable

Front 2

All other variables are kept constant by

Back 2

(1) direct experimental control

(2) Random assignment

Front 3

if the scores of the two groups are different

Back 3

Researcher can conclude that independent variable caused results

Front 4

Confounding Variable

Back 4

varies along with the independent variable

- occurs when effects of independent variable and an uncontrolled variable are intertwined

- with confounding variable cannot determine which variable is responsible for the effect

Front 5

Internal Validity

Back 5

When results of an experiment can confidently be attributed to the effect of the independent variable

- to achieve must design and conduct experiment so that only independent variable can be cause of results

Front 6

Simple experimental design =

Back 6

2 variables

- independent variable

- dependent variable

Front 7

To make sure only difference between the 2 groups is manipulated variable (3 steps)

Back 7

(1) Obtain two equivalent groups of participants

(2) Introduce the independent variable

(3) Measure effect of the independent variable on the dependent variable `

Front 8

Steps for experiment (for high internal validity)

Back 8

(1) Choose participants and assign them to the levels of the independent variable

(2) Operationally define the independent variable to create at least 2 levels

(3) Determine a way to measure the effect of the independent variable by operationally defining the dependent variable

Front 9

(1) Choose the participants and assign them to the levels of the independent variable

Back 9

- must achieve equivalent groups to eliminate selection differences

Front 10

3 ways to assign participants to experimental conditions

Back 10

(1) independent groups design: groups are made equivalent by randomly assigning participants to experience only one condition of the independent variable

(2) Repeated measures design: participants are assigned to participate in all levels of the independent variable

(3) Matched pairs design: makes group equivalent by first selecting pairs of participants who score the same on some variable of interest

Front 11

(3) Determine a way to measure the effect of the independent variable by operationally defining the dependent variable

Back 11

same measurement procedure is used for both groups so that comparison of the two groups is possible

Front 12

Independent groups design

Back 12

different participants are assigned to each level of the independent variable

- through random assignment

- as # of participants in study increases, random assignment prevents any systematic biases

- with sufficiently large sample of participants, random assignment will produce groups that are virtually identical in all respects

Front 13

pretest

Back 13

given each group prior to introduction of the experimental manipulation

- to ensure that groups are equivalent at the beginning of the experiment

- usually not necessary if participants have been randomly assigned to the two groups

Front 14

posttest - only design

Back 14

when no pretest is given

Front 15

3 main reasons why researcher may add a pretest

Back 15

(1) sample size is small => increases the likelihood that groups actually will not be equal on all variables

(2) to select the participants to include in the experiment

(3) when there is a possibility that participants will drop out of the experiment

Front 16

mortality

Back 16

dropout factor in an experiment

Front 17

Disadvantages of a pretest (3)

Back 17

(1) time-consuming and awkward to administer

(2) Can sensitizes participant to what is being studied thus creating demand characteristics

(3) can reduce external validity

Front 18

Solutions for pretest disadvantages (2)

Back 18

(1) pretest can be disguised using deception

(2) Embed the pretest in a set of irrelevant measures

Front 19

Solomon four-group design

Back 19

- possible to assess the impact of the pretest directly by treating the presence of a pretest as a 2nd independent variable

- Participants randomly assigned to one of four possible combinations

Front 20

Repeated Measures Design

Back 20

participants are repeatedly measured on the dependent variable after being in each condition of the experiment

Front 21

Advantages of Repeated Measures Design

Back 21

-Fewer participants needed

-Extremely sensitive to statistical differences

- Error variance is reduced b/c people serve as their own control group

Front 22

Disadvantages of Repeated Measures Design

Back 22

Order effect: the order of presenting the treatments may affect the dependent variable

Front 23

3 types of order effect

Back 23

(1) Practice effect: improvement in performance as a result of repeated practice with a task

(2) Fatigue Effect: deterioration in performance as the research participant becomes tired, bored or distracted

(3) Contrast Effect: when the response to the 2nd condition in the experiment is altered b/c the 2 conditions are contrasted to one another

Front 24

Complete Counterbalancing

Back 24

all possible orders of presentation are included in the experiment

- 2 conditions: 2 X 1 = 2 orders

- 3 conditions: 3 X 2 X 1 = 6 orders

Front 25

Partial Counterbalancing

Back 25

-Latin Square: limited set of orders constructed to ensure...

(1) each condition appears at each ordinal pattern

(2) each condition precedes and follows each condition one time

Front 26

Time Interval between Treatments

Back 26

need to determine the time interval between presentation of treatments and possible activities between them

Front 27

Matched Pairs Design

Back 27

more complicated method of assigning participants to conditions in an experiment

- ensure groups are equivalent on the matching variable prior to the intro of the independent variable manipulation

- can be costly and time-consuming

Front 28

3 steps of Matched Pairs Design

Back 28

(1) Obtain measure of the matching variable from each individual

(2) Form matched pairs that are approx. equal on the characteristics

(3) One member from each pair is to participate in one of the two conditions in the experiment