PSYC1040 Week 2

Question 1

Theories in research

Answer 1

• one study doesn’t tell us much - results could be a statistical fluke, caused by another factor, limited to details of that study’s methods or could even be fraudulent
• usually we need many studies to rule out these alternative explanations - lots of evidence and varied evidence
• if things go well from multiple studies we can get

Question 2

What qualities make for a better theory

Answer 2

• falsifiability, testability
• scope, breadth
• simplicity
  predictive power, accuracy
• fruitfulness

Question 3

What qualities make for a better theory? Falsifiability

Answer 3

• if we can always be wrong so we need to know when we are
• if our beliefs are untestable in principle, they may be irrelevant
• if they’re untestable in practise, we can get trapped with false beliefs
• many researchers view falsifiability as the defining feature of science

Question 4

What qualities make for a better theory? Predictive powers, accuracy

Answer 4

• how closely the theory predicts or fits observations
• ideally predictions are both precise and very similar observations

Question 5

What qualities make for a better theory? Scope, breadth

Answer 5

• the range of things the theory applies to

Question 6

What qualities make for a better theory? Simplicity

Answer 6

• relying on few assumptions or ideas; makes a theory easier to use
• of two otherwise equal theories, the simpler one is more likely to be time

Question 7

What qualities make for a better theory? Fruitfulness

Answer 7

• how many new testable ideas the theory leads to

Question 8

Desirable properties of studies: Reliability

Answer 8

• the likely similarity of results if the study were repeated
• the extent to which chance can be ruled out as an explanation for the results
• a study with no reliability is nearly useless

Question 9

Desirable properties of studies: Realism or relevance

Answer 9

• how well the study represents the topics and contexts of interest
• includes the validity of measures of realism of tasks and participants

Question 10

Desirable properties of studies: casual interpretability

Answer 10

• how well the study supports conclusions about cause and effect
• depends on ruling out alternative causes of the outcome variable

Question 11

Explorations for associations between variables

Answer 11

• A causes B - smoking causes long problems
• B causes A - long problem causes smoking
• C cause both A & B - genetic profile causes both
• coincidence (chance) - the studies were unreliable

Question 12

Basic designs - experiments

Answer 12

• a study in which researchers intervene to change the state if the variable in order to see its effects on another variable and can make reverse-cause and ‘third variables’ implausible explanations for results
• often intervention involves giving different treatments to different groups of participants

Question 13

Assigning participants to groups (between groups experiments)

Answer 13

• ideally, we would roughly match the groups on all characteristics
• the only way to do that is to randomly assign the participants to groups. Problem: groups will still differ by change
• solution: statistical analysis to assess how often chance alone would produce a result a result at least as strong as ours (also, repeating the study).

Question 14

Controlling a variable without comparing groups

Answer 14

• instead of giving different treatments to different groups, we can give each person all the treatments (for example, at different times)

Question 15

The basic logic of a simple idealised experiment

Answer 15

• step up two situations that are identical
• change the state of the supposed causal variable in one of the situations
• measure the outcome in both situations
• if the outcome differs between situations, the supposed causal variable must be a cause; if the outcome isn’t different, it must not be

Question 16

Other problems with experiments

Answer 16

• realism - setting up an intervention often reduces study realism
• feasibility - intervention maybe too costly or time consuming to run
• ethics - interventions may violate consent, cause sufferent and unfairness

Question 17

Observational (correlational) studies

Answer 17

• a study where researchers measure variable of interest, without intervening to change their states and look for associations between them
• problem: in principle, cannot rule out reverse-cause or ‘third variable’
• however, the right kinds of observational evidence can together give strong support to cause and effect conclusions. Experiments are perfect for determining cause and effect

Question 18

Interpreting observational studies

Answer 18

• like always, consider the plausibility of various alternative explanations for the results and consider all the study’s properties
• be especially aware of reverse-cause and third variable explanations
• check the most relevant experimental evidence, even if its not closely related
• also, consider may studies with different methods where possible