Critical thinking

Question 1

What are quasi-experiments?

Answer 1

Studies in which we don’t have complete control over certain elements - circumstances lack at least one of the defining characteristics of a true experiment e.g. subjects are SELECTED rather than randomly assigned to conditions, so we have NON-EQUIVALENT GROUPS, experimenter does not have full control over the IV so we can’t be sure what is causing changes in DV

Question 2

What can we do in non-equivalent control group studies?

Answer 2

Employ pre-testing and post-testing of participants - the DV becomes the CHANGE over a period rather than simply a final score, and we can control for any pre-existing differences between groups
We can also try to test for the significance of any factors we think could influence results

Question 3

What are two interpretable patterns of results in non-equivalent control group studies?

Answer 3

No diff in pre-test but diff in post-test - difference is very likely to have been caused by the manipulation rather than by any existing differences
Pre-existing differences but go in opposite directions following manipulation

Question 4

What is a mixed factorial design with one non-manipulated variable?

Answer 4

A design which attempts to control for possible confounds resulting from non-equivalent control groups
We have more than one IV e.g. one may be participant variable of gender while the other is an experimental manipulation (to which subjects can be randomly allocated)

Question 5

What is an interrupted time series design?

Answer 5

Use publicly recorded data on large numbers of people (archival) in a time-series design on smaller samples
Performance is tracked over several measurements, and in the interrupted design this leads up to introduction of a treatment and then measurements afterwards
Effect is highlighted by a DETECTABLE CHANGE in the DV of interest, and no change in other variables we don’t expect to be affected

Question 6

What is a repeated treatments design?

Answer 6

Baseline measures taken and a treatment then administered
Treatment then withdrawn and participants return to baseline, before treatment administered for a second time
Group acts as own control

Question 7

When would we conduct a comparison study?

Answer 7

To provide information on changes in a psychological variable over time, which may provide evidence for developmental theories

Question 8

What are cross-sectional studies?

Answer 8

Compare samples drawn from distinct sub-groups of a population e.g. different birth years and compare them on a variable at one moment in time

Question 9

What are some disadvantages of cross-sectional studies?

Answer 9

Non-equivalent groups –> confounding
Can’t detect maturational changes that can confound
Cannot observe changes in SAME individuals
Cohort effects

Question 10

What are longitudinal studies?

Answer 10

Surmounts issue of group equivalence and employs repeated measures on the same group of people over a substantial period
Can observe genuine changes/stability in characteristics and if check at regular intervals may be able to identify major points of change

Question 11

What are panel designs?

Answer 11

Longitudinal studies which consider changes in people over time
If samples are taken from the whole population to give ideas of national trends, the sample will be called a cohort

Question 12

What are key disadvantages of longitudinal studies?

Answer 12

Risk of attrition
Decisions made at start cannot be reversed
Participants become wise to tests (reactivity)
Time consuming and expensive
Cross-generational problems (confounding by secular trends) i.e. change taking place in population over time

Question 13

What are cross-sequential studies?

Answer 13

Combine both longitudinal and cross-sectional studies
Begins cross-sectional - data collected at one point in time from two or more groups
A follow up is conducted on the groups two or more times, hence the longitudinal aspect
Can help to separate cohort, developmental, and secular effects

Question 14

What are meta-analyses?

Answer 14

Employ statistical techniques to combine results of hundreds of studies of the same/similar hypotheses/constructs
Studies thoroughly reviewed and sorted and the collated set of acceptable results forms the new data set
Result of each study is treated like a participant result in a single study

Classify and code important characteristics of the studies in relation to research question, and aggregate findings, generating weighted pooled estimates of effect size

Question 15

What are advantages of single-subject/small n designs?

Answer 15

Don’t distort behaviour as means do
Effects must be large to reach clinical significance
Practical/ethical problems addressed more easily
Flexibility in design

Question 16

What are disadvantages of single-subject/small n designs?

Answer 16

Sometimes subtle effects can be important
Some questions simply require groups
Retraining required if any participants leave

Question 17

What are baseline measures?

Answer 17

Control strategy in single subject designs with both a descriptive and predictive purpose
Need to be stable and showing a trend opposite to that of the predicted baseline
Under an intervention we should see a change from the descriptive to the predictive

Question 18

What is the AB design?

Answer 18

Simplest design to use - baseline followed by treatment

But cannot tell if any observed DV changes are due to IV or coincidental confounds

Question 19

What is the ABA design

Answer 19

Baseline –> treatment –> remove treatment

However treatment may not be reversible, or we may actually want to leave participant with effects of treatment

Question 20

What is an ABAB design?

Answer 20

Repeated treatments design - leaves learner with benefit of the treatment and if we see systematic changes we can be fairly confident the treatment caused them

Question 21

What is the alternating treatments design?

Answer 21

Allows for comparison of more than one treatment
Baseline –> treatment 1 –> treatment 2 –> treatment 1 –> treatment 2 etc
Can be used to assess which treatment is most effective

Question 22

What is the multiple baseline design?

Answer 22

Tries to see if a behaviour change corresponds to onset of treatment in different scenarios i.e. does the treatment effect generalise
Different behaviours, subjects and settings can be investigated

Question 23

What is the changing criterion design?

Answer 23

Often used in shaping behaviour, this introduces successively more stringent criteria for reinforcement to see if behaviour changes with criteria
Useful when behaviour change irreversible e.g. learning