Correlations in SPSS
In this example, we have correlated groups—in other words, “parent score” and “child
score” are a pair for the first person in each of the two rows.
To get a correlation coefficient, use the “correlate” command (again, in the “Analyze”
menu). Because a pair of variables will be correlated, this is a “bivariate” correlation. (If
you were to produce many correlations at once – a correlation matrix – you would still be
looking at each pair individually, so you would still use the “correlate/bivariate”
command).
This window appears:
Select these two variables and
then use the arrow to move them
both over to the “variables” area.
We’re assuming these variables are
I/R-level (i.e., “scales”), so the
Pearson correlation is the correct one.
If you had many variables and wanted to produce a correlation matrix showing how each
of them is related to each of the others, you would just move all of them to the right hand
column in this same dialog window.
Hit the “Options” key to go on:
Let’s get means and standard
deviations for these variables.
Hit “Continue,” and then “OK” to see the output:
Correlations
Parent's Score on Anxiety Measure
Pearson Correlation
Parent's Score on
Child's Score on
Anxiety Measure
Anxiety Measure
1
Sig. (2-tailed)
N
Child's Score on Anxiety Measure
.756*
.011
10
10
Pearson Correlation
.756*
1
Sig. (2-tailed)
.011
N
10
This is the correlation between
parent10and child scores …
*. Correlation is significant at the 0.05 level (2-tailed).
… and this is the same
correlation.
This p-value tells you whether
the correlation right above it is
statistically significant or not.
This output is called a correlation matrix. A correlation matrix has a list of all the
variables across the top, and the same list down the side. The diagonal is always all 1’s,
because that’s the correlation between each variable and itself.
A correlation matrix with more variables included would just have more rows and
columns. You can either read across (row name correlated with column name) or down
(column name correlated with row name) and get the same answer.
Paul F. Cook, University of Colorado Denver, Center for Nursing Research
Updated 1/10 with SPSS (PASW) version 18