Research Methods and Statistics Chapter 3, 6, 7, 8 and 10 (Different Research Claims) Flashcards
What are the three types of research claims
- Frequency claims (proportions of variable of population)
- Association claim (association between variables)8
- Causal claim (one variable causes another)
What are the 4 big validities
- External validity (can we generalize the research to other population and operationalizations)
- Construct validity (does the measure really measure what it is supposed to measure?)
- Internal validity (what is the causation –> only in experiments)
- Statistical validity (comparing the values from both conditions)
Is there an effext?
–>
Whats a thematic apperception test
Letting a person write a story about a picture and valuing the emotions
What is the most ideal design for frequency claim
Surveys and also observational research
What’s the most ideal research design for association claims
Correlational research
What are the 6 types of probability sampling
- Simple random sample (everybody of population has same chance to participate) (very hard but most important)
- Cluster sample (everybody in a category like schools or organisations)
- Multistage sampling (sample random subjects of clusters)
- Stratford random sampling (sample according to percentages like blood types)
- Oversampling (sampling more of a specific minority)
- Systematic sampling (every nth person)
–> is a representative sample which can be generalised
What are the types of non probability sampling
- Convenience sampling (accessible samples)
- Purposive sampling (search for samples with specific requirements)
- Snowball sampling (ask subjects for other fitting subjects)
- Quota sampling (set a quota and fulfill them)
–> can not be generalised to population
Where is good sampling important
In frequency claims
It is not that important in causal and association claims
What are the 3 challenges in observational research
- Observer effect (observer affects subjects behaviour)
- Observer bias (your expectations affect your observations)
- Reactivity (subjects behave differently when they know that they are being observed)
What are solutions against observer effect and observer bias
- Unambiguous codebook
- Multiple observers
- Masked design (observers are unaware of studies purpose)
What are solutions to reactivity problem
- Blend in (making yourself less noticable)
- Wait it out (subjects will get used to presence after some time)
- Unobstrusive (study evidence for behaviour)
What is the challenge in correlational research
Correlation doesn’t say anything HOW the association works
- A causes B
- B causes A
- C causes A and B
What are the three criterias for a causal claim
- Covariance (there is an effect)
- Temporal precedence (the causal came before the effect)
- Internal validity (there are no confounds)
What is a bivarate correlation
A association that involves exactly 2 variables
How do you interrogate statistical validity
- Effect size: (How big is effect size)
- Measure statistical significance (is it surely not just by chance?)
- Measure potential outliers (are there any values affecting the results)
- Restriction of range (is the range not big enough?)
- Curvilinear (is the association maybe a curve?)