Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
52 Cards in this Set
- Front
- Back
why do formal research in psych?
|
so we can test and justify our claims and its what distinguishes psych from other social sciences.
|
|
why publish method, analyses and results/
|
so conclusions can be scrutinised by others, we can look at the strength of effects, makes us accountable.
|
|
What does the truthfulness of research results depend upon?
|
quality of design, quality of measures, nature of the effects and match between the results and the conclusions.
|
|
which studies are the only ones that determine causality/
|
experiments
|
|
problem with correlational studies..
|
reverse causality and that another factor was responsible for the relationship between A and B. A spurious relationship.
|
|
to establish causation you must have three things..
|
covariation - change in IV leads to change in DV
temporal order - cause must come before effect no alternative explainations. |
|
Designing an experiment so that there are no alternative explanations is otherwise known as..
|
increasing internal validity
|
|
What are history effects?
|
large scale events that affect your sample alongside the IV.
|
|
What are instrumentation effects?
|
changes in measurement throughout the study because of unfamiliarity with the assessment tool. or changes with the way measurement is interpreted.
|
|
What is the affect of selection on internal validity
|
can cause nonequivalence of groups ie self selection
|
|
Testing effects?
|
Differences in one set of measurements due to an earlier one. ie practice effects or might be more accurate second time around.
|
|
Experimenter bias
|
Where the researcher treats the groups differently either intentionally or unintentionally. or have to use subjective judgement.
|
|
regression towards the mean
|
where scores second time around are closer to the mean score. can be up or down.
|
|
confounds
|
something that varies systematically with the IV rther than an extraneous variable.
|
|
Attrition
|
Can cause inequality in groups. If its even across groups its OK.
|
|
Nonspecific treatment effects
|
Where other things such as jut having a treatment affect the DV rather than the treatment itself.
|
|
Observee reactivity
|
where people act differently because they are being watched.
|
|
Social desirability
|
giving the 'right' answer because you think thats whats needed. Still a prob even when anonymous.
|
|
maturation effects
|
natural change over time.
|
|
Demand effects
|
influence by the presence of the researcher. Wanting to report an improvement.
|
|
How do you control these threats to internal validity?
|
by using random allocation and an appropriate control condition.
|
|
what is a quasi experimental design? What are the two types?
|
Where Ps are not randomly allocated which means that your groups may not be even, but sometimes this is the only way to do it because of ethics. Two types are static comparisons and pretest postest nonequiv groups design.
|
|
what is a static group comparison?
|
where you have two nonrandomly allocated groups and you look at the effect of something on each group, where one gets the something and one does not. ie antenatal classes
|
|
Problem with static group comparisons/
|
self selection. Groups might differ on something else as well.
|
|
What is a factorial design/
|
Where there are two or more IVs manipulated and all combinations of the IVs are tested
|
|
How are factorial designs described?
|
2x2= two IVs with two levels each
2x3=two IVs, one with two levels one with 3 2x3x2=three IVs. two with 2 levels and one with 3 levels. |
|
What are psychometric characteristics/
|
the reliability and validity of a measure.
|
|
Construct validity is..
|
the degree to which the test actually tests what you want it to. the truthfulness of the measures
|
|
Face validity
|
the degree to which the test appears to test something.
|
|
content validity
|
the degree to which the test tests the whole contruct and not just part of it.
|
|
criterion validity
|
the relationship between the measure and an external construct that relies on it.
|
|
concurrent and predictive validity are two examples of what type of validity?
|
criterion.
|
|
convergent validity
|
how much the measurement converges/correlates with measurements of similar constructs.
|
|
discriminant or divergent validity
|
the opposite to convergent validity.
|
|
The quality of study relies upon..
|
construct validity, internal and external validity
|
|
What is reliability?
|
the consistency of a measure.
|
|
What does Cronbach's alpha measure?
|
the internal consistency -reliability! items that test the same thing should have a high correlation.
|
|
What is item-total correlation/
|
where individual items are correlated with the total score.
|
|
Does reliability guarantee validity?
|
No. it can be reliably wrong.
|
|
What are the standards for ethical research?
|
informed consent, voluntary, ethics committee consent, anonymous, debriefed afterwards, can withdraw at any time, do no harm, confidential
|
|
What are some of the probs with questiionaire questions?
|
SURDIQUALVS
|
|
what is regression?
|
clever correlations that can handle more than two continous variables or a categorical IV.
|
|
types of regressions
|
simple regression -just a correlation
multiple regression |
|
types of multiple regression
|
simultaneous multiple regression
hierarchical multiple regression |
|
simultaneous regressions...
|
all variables put in together and SPSS works out which ones are most important.
|
|
multiple regressions are good for answering what types of qns?
|
can look at combined influence and separate influence of IVs on the DV.
Can control for one IV while looking at another can look at interactions. |
|
an example of an interaction
|
when the affect of self esteem on body sat is diffferent for men than for women.
|
|
What is a mediation?
|
the effects of an intervening variable on the outcome. Where A causes B and b causes C. A causal relationship
|
|
What does the test of mediation do?
|
it allows us to get rid of the mediating factor and look at the effect of one IV on the DV.
|
|
What does the test of an interaction tell us?
|
the main effect of each Iv on the DV and the sig of the interaction itself.
|
|
A mediation exists when..
|
there is less effect on the DV of one IV when the other DV has been removed.
|
|
What the three steps to testing a mediation hypothesis?
|
1 check preconditions for a mediation exist by a correlation.
2do a simple regression to test whether A affects C without B for comparison purposes. R square will tell the variance explained by A. 3. Controlling for b, the second regression will tell how much additional variance is explained by A. This is evident in R square change. 3 |