Choosing Appropriate Stats Test Flashcards
What is the research process?
- Data and initial observation that leads to research question
- Generate theory
- Generate hypothesis and identify variables
- Collect data and measure variables
- Analyse data and graph it and fit model
Where is the research process do you choose your stats test?
In the planning phase before data collection as you need to select sample size that will provide high power which is determined by the type of test you do
What are the factors that determine statistical test choice?
- Purpose of study
- Number of DVs
- Type/nature of IVs and DVs
- Number of groups and/or conditions
- Relatedness of scores between conditions
What are the options for purpose of study?
Descriptive - summarising a single sample’s characteristics
Inferential - testing group differences or exploring relationships between variables
What are the options for number of DVs?
Univariate - 1 DV
Multivariate - >1 DV (related)
What are the types of variables?
Binary - 2 categories, categorical, non-parametric
Nominal - >2 categories, categorical, non-parametric
Ordinal - same as nominal but have a meaningful order or ranking, categorical, non-parametric
Interval - equal intervals between values, no true zero, continuous, parametric
Ratio - has absolute zero, can use ratios e.g a is twice as big as b, continuous, parametric
Why should you never alter the nature of variables to use a different test?
You can lose information about individual differences
Risks of overlooking non-linear effects
You can lose effect size and power
Spurious significant main effects and interactions
What are the options for number of groups/conditions?
None —> correlations, regressions etc.
2 —> t tests etc
> 2 —> ANOVAs etc
What are the options for relatedness of scores between conditions?
Independent- BS
Dependent - WS, Mixed design, matched pairs (couples, best friends, relatives)
