correlation Flashcards
correlation
degree to which scores on two variables corelate
coefficient
(r) index of the extent to which 2 variables are linearly correlated
person’s product moment correlation coefficient
default
z value
(x-u)/SD
observed value - mean / SD
reporting the results of C
f(df)= p=
df in correlation
n-2
because always have two variables in correlation
assumptions of pearson
uses size of scores
2 variables must be measured on interval scales
2 variables must be linearly related
spearmans rank
(rs) used when scatterplot indicates the relationship is MONOTONIC but NON LINEAR (and when there are OUTLIERS)
also when have ORDINAL data
monotonic
relationship but non linear- like curved
parsons would …….. a monotonic relationshup
underestimate
because it’s looking for a straight line
type II error
non monotonic
v or inversed (u shaped)
neither ranked nor raw would work
split data in half
partial correaltion- when
when want to control for effect of third variable based on theoretical assumption
df for partial correlation
n-1
reporting partial correlation
partial c was conducted to test the strength of the relationship between x and y after controlling for z
pos neg sig r(df)= ,p=
pearsons would …… when there is a cluster on data and then 1 outlier
overestimate because straight line could be drawn.