Specialised research designs

Question 1

Developmental designs

Answer 1

Cross section
Longitudinal
Cohort sequential

Question 2

What is a cross-sectional design

Answer 2

• Create ‘cohort’ groups based on different ages of PPs
• Collect developmental data in a short period of time
• Generation effects may be a problem
o E.g. you might measure substantial differences in IQ with age (20-70 years old) because of education differences across different generations

Question 3

Longitudinal designs

Answer 3

o longitudinal research is used to study individuals at different stages of their lives, or during different stages of development of a variables of interest.
• Same set of PPs measured several times
E.g. over months, years etc.
It is observational, meaning that researchers do not manipulate or introduce any variables. Similarly, this also means that there are a lot unmeasured variables that will have an unknown extraneous influences on the variables of interest. Causal inferences can therefore not be made based on longitudinal data.
• Avoids possible generation effects that are present in cross sectional designs.
• Problems:
• Possible cross-generational problem- Results from one generation may not generalise to another
o PP attrition
o Multiple observation effects- doing the same survey multiple times. How they answered last time might influence how they answered this time.

Question 4

Cohort-sequential designs

Answer 4

• Combines cross-sectional and longitudinal components in the same design
research method in which a cross-section of the population is chosen and then each cohort is followed for a short period of time.
• Allows some disentangling of age, cohort and time
• Possible to test for generation effects, though does not eliminate them
Much less susceptible to bias (generational effects) therefore yields more accurate data than a cross sectional study.

Question 5

2 single case designs

Answer 5

Case study approach

Single-case experimental design

Question 6

Describe the case study approach

Answer 6

o Observational technique of observing a single PP
o Often used in clinical settings
o Descriptive or observational approach
o No manipulation or control of variables, simply recording of observations

Question 7

Describe single-case experimental design

Answer 7

o Consists, in the limit, of one PP but typically 3-6 PPs used to establish generalisability and external validity
o Includes stringent control
o Systematic effect of treatments/conditions over multiple observations establishes cause and effect

Question 8

2 types of single-case experimental designs

Answer 8

Baseline designs

Discrete trials designs

Question 9

single-factor baseline experimental design

Answer 9

 Single cae design
 Most common is ABA (baseline, treatment, baseline)
 Demonstration of treatment effect requires return to baseline
 To establish intra-participant replication and avoid ending on baseline in clinical contexts, can extend it to ABAB (ending with treatment)
 ABAB can be extended to accommodate multiple levels of the IV
 Need to counterbalance to deal with carryover effects

Question 10

Multifactor baseline experimental designs

Answer 10

Single case design
 More than IV
 Assesses main effects and interactions without stats

Question 11

Multiple baseline designs

Answer 11

Single case design
 Used when treatment produces irreversible change in behaviour
 Simultaneously measures multiple behaviours, each with its own baseline
 Behaviours must be independent of each other
 E.g. smoking and excessive coffee drinking

Question 12

Discrete trials design

Answer 12

Single case design
• Used when you cannot block treatments/conditions
o E.g. consider a detection experiment using target-present and target-absent trials; if target-present and target-absent trials are blocked, the bias to respond will increase in target-present blocks and decrease in target-absent blocks, and response bias will be measured NOT detection
• Multiple treatment/condition trials randomly intermixed and treatment/condition means are calculated by averaging across trials
• This approach can be combined with group testing and the group results analyses using inferential statistics

Question 13

Single case design, advantages and siadvantages

Answer 13

Single-case designs: advantages
• Focuses on tightly controlling error variance
• Focuses on individual behaviour:
o Makes identifying and controlling sources of error variance easy
o May reveal critical mechanistic details lost with group approach
 E.g. attention and transfer of learning about compound stimuli
• Causal relationships can be established with very few PPs
Single-case designs: disadvantages
• Making multiple observations is time consuming and can be tedious
• Not appropriate for all research questions
• Results may be of limited generality
• All variables that can cause error variance cannot be identified and controlled

Question 14

What is a meta-analysis

Answer 14

• Specific forms of statistical analysis have been developed to combine findings from different studies = meta-analysis
• Meta-analysis can be less conservative than subjective review
• To conduct:
o Set strict criteria for inclusion of studies
o Take account of sample size
o Assign different weights to different studies according to the quality of outlet/method
• Types of meta-analytic techniques:
o Determine if overall significance exists
o Determine size of treatment effect
 Take care if different paradigms are associated with different effect sizes
• Especially useful in trying to make more definitive conclusions where findings from different studies have been mixed