statistical tests Flashcards
1
Q
difference or association
A
- data investigating a difference will usually have 2 conditions, a control and an experimental. this is usually a typical experiment
- or a psychologist may be establishing an association/ relationship between 2 co-variables. this is usually a correlational study
2
Q
experimental design
A
- should only consider this when looking for a difference
- related data refers to when participants in each condition are somehow related: repeated measures or matched pairs
- unrelated data refers to having 2 separate groups of people in each condition: independent groups
3
Q
nominal data
A
- categorical data
- the data is discrete, it will only appear in one category
- e.g. if a researcher wanted to know if more students doing psych went to a school or college, the data would be categorised as ‘school’ or ‘college’
4
Q
ordinal data
A
- data which is ordered somehow and the intervals between data aren’t equal
- ranking data where values assigned have no meaning besides stating where one scores in relation to another
- e.g. someone ranking their preference of local restaurants from 1 to 10. the difference between 1st and 2nd will probably be different to the difference between 8th and 9th
5
Q
interval data
A
- data which is ordered somehow
- the intervals between each value are equal in measurement
- e.g. time, temperature
6
Q
acronym for the order of tests
A
simon - sign test
cowell - chi-squared
wants - wilcoxon
more - mann whitney
singers - spearman’s rho
receiving - related t-test
unanimous - unrelated t-test
praise - pearson’s r
7
Q
parametric and non-parametric tests
A
-parametric tests are preferred as they’re more powerful, but they require certain criteria:
- data should be interval
- data should be drawn from an underlying normal distribution
- should be homogeneity of variance (the variances in the two groups shouldn’t be significantly different)
- the 3 parametric tests required are the related t-test, the unrelated t-test and pearson’s r
