# How to report a matrix correlation table?

1. When reporting correlations, is it necessary to report these in the matrix table format?

2. I ran a correlation between 7 variables, mainly to assess the influence of 6 independent variables on the dependent variable (DV). When I report these results is it necessary to make the matrix table 7$$\times$$7 as that would involve reporting the correlations between all the variables? Or is it appropriate to simply report the relationship between the 7 variables and the DV, making the table format 7$$\times$$1?

• Everyone can agree that (1) showing more results does no harm in the sense that readers can ignore what doesn't seem of interest or use to them (2) space is usually limited (3) showing irrelevant or unimportant detail is not a good idea. In practice there is a tension or trade-off between these simple principles (platitudes?). A reviewer or supervisor or examiner could criticise you either way, for not showing enough detail or for showing too much. Commented Oct 21, 2018 at 10:33

I recommend that you report the correlation matrix with means and standard deviations. Only the lower (or upper) half of the matrix need be reported. Here is an example taken from this link:

http://www.scielo.org.co/pdf/rlps/v47n1/v47n1a01.pdf

Unlike the example below, I normally place the M and SD values below the correlations rather than to the side since that approach allows for more correlations to be reported.

One reason to report the entire matrix is that it allows others to replicate your regression analysis without raw data. One can enter the correlation matrix, plus M and SD, and run various regression models without the raw data. One could also perform various structural equation models from the same correlation matrix.

As an example, using the correlation matrix above I entered the first four variables into SPSS using the following commands.

MATRIX DATA VARIABLES=age extro emotion agree
/FORMAT=LOWER DIAGONAL
/CONTENTS=MEAN SD N CORR.
BEGIN DATA.
22.75 3.26 2.77 3.56
4.81 0.42 0.57 0.34
228 228 228 228
1.00
0.03 1.00
0.04 -0.04 1.00
0.05 0.19 0.13 1.00
END DATA.
Regression
/MATRIX=IN(*)
/dep=agree
/enter=age extro emotion.


Below are the results of a regression with agreeableness as the outcome and age, extroversion, and emotional stability as the predictors.

• In terms of the example you have given, if i was only interested in reporting the correlations between variables 2-11 with age, in this format i would still be reporting all the correlations between all variables and not just the relationship i need. The correlation table will used to explain only one relationship between the variables of interest as support for my regression model which would be outlined following this
– Ali
Commented Apr 10, 2018 at 11:13