choosing a statistical test Flashcards
(18 cards)
what things do you need to know in order to choose a statistical test?
•experimental design
•level of measurement
•difference/ association
when would you use the mann-whitney U test?
•difference
•independent groups (unrelated design)
•ordinal data
when would you use wilcoxon T?
•difference
•repeated measures (related design)
•ordinal data (self- report questionnaire)
when would you use the unrelated T test?
•difference
•independent groups
•interval data
when would you use the related t-test?
•difference
•repeated measures design
•interval data
when would you use spearman’s rho?
•correlation
•ordinal or interval data
when would you use pearson’s r?
•correlation
•interval data
when would you use the chi-squared test?
•difference or association
•independent groups
•nominal data
choosing a statistical test diagram
in which type of hypothesis would you use a one-tailed test?
a directional hypothesis
in which hypothesis would you use a two tailed test?
a non-directional hypothesis
what you need to know in order to figure out whether a result is significant:
- P value (usually 0.05)
- N (number of participants)
- directional/ non-directional hypothesis
- one tailed/ two tailed test
- calculated value
- critical value
how do you calculate the calculated value in the sign test?
s= experimental condition result- control condition result
the S value is the value of the less frequent sign (plus or minus)
what does the P value being 0.05 mean?
that there is a 5% chance that your result is due to chance
what tests does the calculated value have to be smaller for to get a significant result?
•sign test
•wilcoxon
•mann-whitney
the poo tests
which tests does the calculated value need to be larger for a significant result?
all apart from sign, wilcoxon and mann-whitney
what do you do when the result is significant (in terms of hypothesise)?
accept the alternate hypothesis, reject the null
P= 0.05
N= 18
critical value= 5
observed value= 7
write an exam answer to this
