A2 RESEARCH METHODS- inferential stats- measures of central tendancy/measures of dispersion Flashcards
what does inferential stats allow?
psychologists to draw conclusions based on probability that a particular pattern of results could have arisen by chance
what if unlikely that results occurred by chance?
described as significant
in most research what do psychologists say?
if results are sig if probability occurred by chance is less than 0.05
how is p less than 0.05 written?
p,0.05
what is it meant in other words?
risk that results occurred by chances is less than 0.05 / 5% so we can be more than 95% sure that they were due to effect of iv and dv
if psychologists want to be more significant what do they do?
p
what do inferential tests determine?
likelihood that results due to chance.
what do they produce?
observed value
what must an observed value be compared to?
critical value which comes from table
in spearmans rho and chi square tests the observed value must be?
must be greater than critical to be sig
in mann whitney and wilcoxon tests the observed value must be?
less than critical value for result to be sig
what are the two types of errors?
type 1 & type 2
what are type 1 errors?
rejecting a null hypothesis that is true and accepting a experimental hyp that is not true. happens when level of signifcance too high e.g 10%
what are type 2 errors?
accepting null when not true and rejecting experimental that is true -happens when sig too low e.g 1%
what are the types of levels of measurements?
nominal, ordinal, interval
what is nominal data?
data in separate categories such as grouped into tall,medium,short. names of categories are simply labels
what is ordinal data?
data ordered/ranked but gap btw each item not the same e.g rating how happy u are on scale of 5
what is interval data?
data measured using specific units of measurement which are equally spaced apart such as measuring height in cm
what are the types of inferential tests?
spearmans rho,
chi square
mann whitney u test
wilcoxon test
when would you use spearmans rho?
testing for correlation, data is ordinal interval or ratio, there are no or few tied ranks
when would you use chi square?
data is nominal, no other test suitable!
when would you use mann whitney?
testing for difference, indepdant groups design, data ordinal interval or rational
when would you use wilcoxon?
testing for difference, there is repeated measures design, data is ordinal interval or ratio
what are the measures of central tendancy?
mean
median
mode
