inferential statistics Flashcards
what are inferential statistical tests?
- Tests based on the mathematics of probability, they help researchers decide how likely a hypothesis is to be true, given the results from a study. They test the null hypothesis.
- Statisticians have produced tests and critical value tables which are used to claim a certain level of certainty that the results are not a chance finding (fluke)
- Psychologists have agreed to use the probability value of p<0.05, which means the probability the results are due to chance at 5% or less
what is nominal data?
data is in seperate categories
what is ordinal data?
data is ordered - differences between them may not necessarily be the same
what is interval data?
data is measured using units of equal intervals (e.g. height measured in centimetres)
name the inferential statistical tests
- chi squared test
- mann whitney
- wilcoxon
- spearman’s rank/rho
when do you use chi squared?
- testing to see if there is a difference between two or more groups (test of difference)
- gathered nominal data
- used an independent groups design (pps do only one condition each)
NICK IS COOL
when do you use mann whitney?
- Testing to see if there is a difference between two or more groups
- Gathered at least Ordinal or higher data (Interval/ratio data)
- Independent Measures design (pps do only ONE condition each)
Ollie Is Messy
when do you use wilcoxon?
- Testing to see if there is a difference between two or more groups
- Gathered at least Ordinal or higher data (Interval/ratio data)
- used a repeated measures design
Ollie Runs Wild
when do you use spearman’s rank/rho?
- testing to see if there is a relationship between two or more groups
- gathered at least Ordinal or higher data (Interval/ratio data)
- looking for a correlation
(WHEN RANKING, LOWEST GETS NUMBER 1)
how do you calculate mann whitney?
- Rank all scores together, ignoring which group they belong to. The lowest score gets a rank of “1”, the next lowest gets a rank of “2”, and so on. If two or more scores are identical, this is a “tie”. They get the average of the ranks that they would have obtained, had they been different from each other.
- Add up the ranks from group a (𝚺) and do the same for group b
- Complete the formula for each group to calculate Ua and Ub
- Compare the smaller of the two U values to the critical value to see if the results are significant - make sure you use the correct table based on whether the hypothesis is one or two tailed and your p value (0.05 or 0.01 etc.) The calculated U value, must be equal to or less than the critical value for the results to be significant.
what is a type 1 error?
accepting the alternative hypotheses when the results are actually due to chance (false positive)
what is a type 2 error?
accepting the null hypothesis when the results were not due to chance (false negative)