Measuring Effects Between Subjects

How can 'cause and effect' be established in between measures design?

Researcher can manipulate just the treatment while holding extraneous variables constant. Differences are thought to be caused by treatment.

1 of 22

Define a true experiment

Experimenter controls which condition participants experience.

2 of 22

What are the advantages of between measures design?

1. Less risk of crossover effects (practice, fatigue etc) 2. Multiple conditions can be tested simultaneously which means there are no contrast effects e.g. time of day and room temperature.

3 of 22

What are the disadvantages of a between measures design?

1. Need a larger sample (>30 for each condition) 2. Individual differences can confound results and increase error variance.

4 of 22

How can we overcome individual differences?

1. Random assignment. 2. Use a matched pairs design.

5 of 22

How can we increase the variability between groups?

Increase strength of manipulation.

6 of 22

Do we want to increase variability within the conditions?

NO!!! So make sure you carefully define population and standardise the test setting.

7 of 22

List the 7 threats to internal validity

1. Assignment bias 2. Differential Attrition 3. Experimenter bias 4. Diffusion/imitation of treatment 5. Compensatory equalization 6. Compensatory Rivalry 7. Resentful demoralization

8 of 22

Assignment bias?

Groups may be different prior to start of experiment, so treatment effect may be caused by individual differences.

9 of 22

Differential Attrition?

Participant withdraws from study before completion - If there's a difference in attrition rates, groups made up of different people, so defeats purpose of random sampling.

10 of 22

Experimenter bias?

Experimenter inadvertently treats participants differently based on expected outcome.

11 of 22

Diffusion/imitation of treatment?

Treatment effects can spread between groups, participants may notice subtle changes in experimenter.

12 of 22

Compensatory Equalization?

Control group demands to have the same treatment as experimental group.

13 of 22

Compensatory Rivaly?

Participants in the control group improve efforts to achieve better/same scores than experimental group.

14 of 22

Resentful Demoralization?

Control group participants 'give up' on efforts to obtain good scores as they resent the presumed treatment advantage.

15 of 22

What is total variance?

Deviation of scores from overall (grand mean) - treatment effects + error variance

16 of 22

Residual Variance?

Deviation of scored from group means.

17 of 22

Model effects?

Deviations of group means from the grand mean.

18 of 22

When and why do we use a post-hoc analysis?

If F is significant (there is a difference between groups). Used to indicate where the differences lie by comparing all possible combinations of groups.

19 of 22

What is Bonferroni?

Type of correction for type 1 error prior to post-hoc analysis - most stringent form.

20 of 22

What are Planned Contrasts?

Compare only means expected to differ at start of test given the hypothesis.

21 of 22

What are the Assumptions of a One-way ANOVA?

1. Homegenity of variance (Scores within groups have similar variances) 2. Scores are normally distributed about means 3. Independence of errors (random, independent sampling and treatment effects are the only similarities within groups)

22 of 22

Get Revising

Measuring Effects Between Subjects

Other cards in this set

Card 2

Front

Back

Card 3

Front

Back

Card 4

Front

Back

Card 5

Front

Back

Comments

Similar Psychology resources:

Other cards in this set

Card 2

Front

Back

Card 3

Front

Back

Card 4

Front

Back

Card 5

Front

Back

Comments

Related discussions on The Student Room

Similar Psychology resources: