what is the multiplicative law?
probability of a set of independant events all happeing is the product of the individual probabilities
what does the combination rule do
work out the chances of picking ‘r’ things out of N things
what are the steps for wilcoxon rank sum?
- rank data with both groups in one (but note which comes from which group)
- calc sum of ranks of group with smaller sample (if equal sample, lower sum of ranks)
- significant if W is smaller or equal to value in table.
where is Wilcoxon ranks sum in the decision tree?
continuous data more than one participant comparing central tendency unrelated sample (two separate groups) NOT both normal and variance equal
what is Wilcoxon rank sum test basically trying to find out?
how much two groups overlap
what are the steps for wilcoxon matched pairs?
- calc difference between each pair (remove those w no diff)
- rank differences, ignoring sign!
- calculate sum of pos diffs and neg diffs
- t= smaller of two ranks youve calculated
- significant if T is smaller or equal to value in table
how do you do a normality check for related samples
use the differences (doesnt matter what way round you subtract)
what are the steps of the binomial test?
that formula that goes a little something like:
prob(x) = N! / x! (N-x)! * P^xQ ^(N-x)
significant if its smaller/ equal to
in the binomial test, what is base rate?
used to describe the percentage of a population that demonstrates some characteristic (p and q part of formula i think)
what are the steps for the n-by-1 chi square test?
- calculate expected value (mean of observed values) THINK THIS IS ONLY IF THERES ONE ROW. IF NOT, (row total x column total)/grand total
- O - E
- square these
- divide these by E
- sum of these. gives us x^2 value
- calculate degrees of freedom (N-1)
- significant if higher
what are the steps for chi-square contingency?
- calculate expected value (row total x column total)/grand total
then same as n-by-1 chi square!!!
- O - E
- square these
- divide these by E
- sum of these. gives us x^2 value
- HOWEVER, degree of freedom is (row-1) x (column-1)
what must you remember about chi square test and one/two tailed?
chi-square will always be two tailed
what must you remember about chi square test and expected values
the calculation of chi-square assumes all expected values are at least 5. if not, the significance may be invalid because it leads to a loss of power aka harder to find significant test.
what are the steps of the variance test?
- calculate variance for each group using formula
- homogeneity of variance- do groups differ significance in variance?? divide larger v by smaller v (use formula). gives F value.
- significant if F exceeds value in table
what are the steps of the z test?
- calculate z using formula which is:
observed value- population mean / population standard deviation
what is the z-test used for?
used to see if 1 person mean is same/diff to population mean
whats that confusing thing that you have to remember for the z-test?
be careful of the question.
if its something like “what proportion of population are scoring LOWER than person” or “HIGHER” then use table score.
but if its “is this person diff from population as a whole” then DOUBLE SCORE to consider people higher and lower.
what is exploratory data analysis?
idfk
how can the unrelated sample t-test be split
equal sample and unequal sample size
how do you know whether to do wilcoxon rank sum or unrelated t-test?
- normality check
- calc variance for each group
no sig. diff = t-test
sig diff = wilcoxon
what are the steps for the unrelated samples t-test with EQUAL sample sizes
- calculate differences between means
- calulate standard error using formula
- calculate t statistic using formula
- calculate d.f. (2N-2)
- significant if it exceeds
what are the steps for the unrelated samples t-test with UNEQUAL sample sizes
- calculate the pooled variance estimate using formula
- use pooled variance estimate to calculate standard error using formula
- calculate t as before using pooled variance estimate standard error
- calculate d.f. (N1 + N2 - 2)
which are the tests which measure amount of overlap and tell you likelihood this happened by chance
related + unrelated t-tests, and z-test
of the question says “significantly higher / lower” what is it measuring?
central tendency
