Re: MANCOVA: describing effect directions

On Sep 4, 12:51 pm, Niklaus Kuehnis <kuehnik_0...@xxxxxxxxxxxxxx>
wrote:
I computed MANCOVAS (SPSS GLM) with 8 simultaneous dependent metric
variables, one binary categorical IV and several metric covariates.
There was a main effect for the categorical IV, for several metric
covariates as well as some IV*covariate interactions.

My questions are:
a) Is it possible to know the *direction* of these effects by
looking at some output parameters? E.g., am I allowed to interpret
the direction of beta weights?

The rules for interpreting the betas are the same as in ordinary
regression: if your data are naturalistic observations then you need
to choose your words very carefully, to avoid even hinting at
causation.


Question (a) refers to the overall effects of the IVs/covariates as
well as to the univariate tests for every DV.

b) How to plot an effect of a metric covariate on a single metric DV?
Does it make sense to run a bivariate linear regression and make a
scatterplot with the regression line?

You need two lines, one for each group. If the covariate does not
interact with group then the lines will be parallel; if it does
interact then the lines will not be parallel.

c) How to plot an interaction between a binary IV and a covariate?
Run two bivariate regressions for each group and plot two linear
regressions?

Yes.


Generally speaking: Do linear plots make sense to describe MANCOVA
results?

Yes.


(xpost & followup 2 ssm)

I appreciate any comments.

--
Niklaus
.

Relevant Pages

  • Re: MANCOVA: describing effect directions
    ... How to plot an interaction between a binary IV and a covariate? ... many textbooks say there is no such thing as an ANCOVA ... an ANCOVA plot with regression lines with different slopes wouldn't ...
    (sci.stat.math)
  • Re: inclusion criteria for regression covariates
    ... > the list of potential variables to include in the multiple regression. ... > the treatment variable and the covariate. ... in a model does NOT in any way depend on its slope being similar to ...
    (sci.stat.consult)
  • Re: inclusion criteria for regression covariates
    ... >> the list of potential variables to include in the multiple regression. ... >> the treatment variable and the covariate. ... > in a model does NOT in any way depend on its slope being similar to ...
    (sci.stat.consult)
  • inclusion criteria for regression covariates
    ... the list of potential variables to include in the multiple regression. ... then the covariate should be included in the multiple ... Does anyone recognise this and know a reference? ...
    (sci.stat.consult)