Lecture 3 - Designing Experiments

Question 1

What are the 3 criteria for causation?

Answer 1

• Covariation
• Temporal precedence
• Internal validity – to rule out other possibilities

Question 2

Define control condition.

Answer 2

Control condition -

Must keep everything the same except the independent variable.

e. g.
- Give an inactive tablet (placebo)
- ‘Sham’ operational procedure (on animals expts)
- Unrelated filler task
- Waiting list control

Experimental design is all about keeping the control as similar to the experimental condition as possible.

Question 3

Define internal validity.

Answer 3

Interna validity -

An evaluation of an experiment asking the question: ‘Is the independent variable the only possible explanation of the results shown for your dependent variable?’

• Need control on extraneous variables.

Question 4

Define extraneous variables (noise).

Answer 4

Unwanted variables that cause the variability of scores within groups to increase.

E.g., effects of time of day, noisy lift, mood of participant, initial skill level of participant.

Question 5

Define extraneous variables (confounders).

Answer 5

Something (not of research interest) that varies consistently with the independent variable.

Careful choice of comparative (control) task can minimise confounds.

Question 6

What are blind studies?

Answer 6

Blind studies -

Participants may have behaved in desired ways if they know they are in experimental condition. Participants should not know which condition they are in.

Question 7

Define demand characteristics.

Answer 7

Demand characteristics -

Features of the experiment that make the participant behave in a particular way.

People often react to the measures being taken

• If they know or guess the hypothesis
• Use funnel debriefing to assess this.
• If they wish to disguise undesirable traits (e.g., racism, paedophilia)
• Use implicit as well as explicit measures.

Question 8

What are expectancy effects?

Answer 8

If a teacher expects a child to do well then they will.
Self-fulfilling prophecy.

Similar effects can happen with researchers and participants through subtle clues.

(not sure I understand this)

Question 9

What’s one way that can help to avoid expectancy effects?

Answer 9

Double-blind studies

• Experimenter expectancy is were the researcher can (possibly subconsciously) affect the outcome. Double-blind procedures ensure that researcher does not know which condition as participant is in.

Question 10

Define HARKing.

Answer 10

Hypothesising After the Results are Known.

• An experiment might throw up many results.
• One happens to be significant.
• A hypothesis is generated for that significant finding post hoc but is claimed to be a priori.

Hypotheses should be clear before the experiment is carried out.

Question 11

List 4 ways of controlling extraneous variables.

Answer 11

• Elimination.
  Removing any known extraneous variables.
• Constancy.
  Ensuring that an extraneous variable is kept constant.
• Matching.
  Ensuring that the extraneous variable is matched over conditions.
• Counterbalancing.
  Balancing to control extraneous variables by using different treatment sequences.

Question 12

Define randomisation.

Answer 12

Randomisation -

Participants (and any other potentially confounding variables) randomly assigned to conditions.

Randomisation within the experiment should make sure that they any nuisance variable are not confounded with independent variables.

If an effect is real and large enough, it should be ‘heard’ over the noise.

Question 13

Define external validity.

Answer 13

External validity -

An evaluation of an experiment that asks whether the observed results generalise to populations or situations different from those in the experiment.

Things that prevent the observations being generalisable are threats to external validity.

Question 14

What are the 4 types of generalisation?

Define them.

Answer 14

Population generalisation -
> Is there something unusual about the population tested?

Stimulus generalisation -
> Is there something unusual about the experimental stimulus?

Environmental generalisation -
> Do the results generalise to the real world situation?
> Ecological validity

Temporal generalisation -
> Are the findings still valid given the changes in society since the experiment was conducted?