Correlations

Question 1

why are correlations useful

Answer 1

if cant manipulate variables

absense of a correlation lets us discount theories

highlight possibilities for research

Question 2

name 3 measures of association

Answer 2

pearsons correlation coefficient (parametric)

spearmans coefficent of rank correlation (non-parametric)

covarience (parametric)

Question 3

whats parametric data

Answer 3

data with a normal distribution

Question 4

whats non parametric data

Answer 4

data without a normal distribution

Question 5

what is covarience

Answer 5

the mean of the product of the deviations from the mean

basically
add up all the deviations each score has from the mean
calculate the mean of this
thats covarience

when the two variables change a lot the covarience is large

when theres no relation the covariance is 0

Question 6

describe pearsons corrwlation coefficent

Answer 6

based on covarience

its independent of the overall variability as it divides the standard deviations of the variables

its only dependent on how related the variables are not how much they vary

values are between -1 and +1 (closer to these values the stronger the relationship)

Question 7

describe spearmans rank/rho

Answer 7

coefficient of correlation related to pearsons r value

non-parametric

used for ordinal data

first it ranks the data then works out pearsons r on the ranks

Question 8

how do you do correlations on SPSS

Answer 8

must have two columns of data - two variables

1 - plot data on scatter plot (graphs, legacydialogs, scatter)

2 - compute the coefficients - (analyze, correlate, bivariate).

3 add variables to right hand side box.

4 select pearsons or Spearman rank

Question 9

what do you get if you correlate a variable with itself

Answer 9

always +1

Question 10

what does a partial correlation do

Answer 10

measures part of the correlation of A and B that isnt to do with C

C is called partialled out or held constant

first order - 1 partialled out, second = 2 etc