Session 2 | Research Design Methods Flashcards
(14 cards)
What is the main difference in terms of reliability between lab experiments and field experiments?
lab > lot of internal validity and less external validity
field > lot of external validity and less internal validity
What is the difference between a Cross-sectional study and a longitudinal study?
Cross-sectional studies : Looking at one element/individual/family/conglomerate across different poiints (age, gender, performance). (focus: snapshot)
Longitudinal study: Looking at their work performance across time. (development)
How do I test causality?
I ask myself if changing the cause will change the effect while I keep other alternative explanations constant and controlled for.
Why should I care about internal validity? What does it mean?
Internal validity means that I am consistently measuring the right thing. Thus, I obtain the same results over and over and I measure the correct parameter. In the case of causality, it means that my effect is caused by the factor that I am measuring.
What can reinforce causality?
- Time sequencing > cause has to precede effect
- strong correlation
- non-spuriousness (no alternative explanations influencing results or possible)
- plausibility and consistency can help
What is group conformity and how is it explained by the Asch experiment?
In the Asch experiment a participant is asked to evaluate the size of an object among others. In reality all others are actor and he is being evaluated on his group conformity. Group conformity is the tendency of individuals to adapt to what others are doing to avoid the creation of bad reactions etc.
Which methods can be usedwithin Longitudinal studies? explain each one of them.
Survey > analysed the variation of one or more variables.
Lagged design > studies the influence of variable A from T1 on variable B in T2.
Cross-lagged design > studies the influence of variable A and B in T 1 and T2 and how they mutually influence each other.
What are latent variables? In which context are they used?
A latent variable is an average score across a series of indicators. It is used to measure psychometric status when something is not directly observable but only through a series of events. (eg. Intelligence, Attitude, Motivation)
Which questions do I need to ask myself when measuring a latent variable?
1) Validity (is it measuring what it should measure?)
2) It the measurement reliable (which means consistent over time? )
3) Responsiveness > is the scale that I am using detecting the changes that are taking place?
What is the difference between criterion validity and content validity?
Criterion validity : does the score of my latent variable rappresenting fully what I want to measure eg. intelligence of a person?
Content validity: Am I measuring the whole variable or only one part of it? eg. Is depression only mental health or does it have chemical components that I did not analyze?
What are the differences between these three types of reliability?
- Inter-rater reliability > is my measure independent of who is rating?
- Test-retest > If I run it twice, do I get the same results?
- Internal consistency > are multiple questions results correctly reporting the effect Y? (we can use Cronbach’s Alpha to verify)
What is Cronbach’s Alpha?
Cronbach’s Alpha calculates the INTERNAL CONSISTENCY which means it evaluates if a set of items can be defined as part of the same group/category.
If I need to test validity, what can I do?
- check correlation between what theory says and my results
- compare your scores with scores from other tests that measure the same thing.
If I need to test reliability, what can I do? How should I proceed?
- test multiple times (same score?)
- use multiple test takers/observers
- calculate internal consistency (Cronbach’s Alpha)
