ANCOVA - quadratic term as a interaction for comparing non-linear data? - Statistica General Discussion - Statistica - Dell Community

# ANCOVA - quadratic term as a interaction for comparing non-linear data?

#### ANCOVA - quadratic term as a interaction for comparing non-linear data?

Hello!

I have non-linear data (growth) and I would like to compare growth curves between two populations - does one population get bigger at a younger age?

I have read that I may be able to compare the elevations using an ANCOVA but including a quadratic term as a main effect and as an interaction such that:

x=age; y=size variable; 2 populations

size variable >= population; age; age^2; population*age population*age^2

However, I have been unable to figure out how to get an interaction in ANCOVA (using both GRM and GLZ). I'm only getting results for the main effects

Q1: is this an appropriate way of comparing 2 curves?

Q2: How can I get an ANCOVA to assess the interaction of my main effects?

I'm using v10

Thanks!

• Hi, Evan,

There are two ways you can achieve it.

Method 1:

After you get to the dialog with the "Between effects" button inactive, first click OK to get to the GLM result dialog,

In the GLM result dialog, then click on "Modify" button,

this will direct you back to the GLM Analysis of covariance dialog, where the "Between Effects" button becomes active to allow you specify interaction terms.

Method 2:

If Method 1 does not work for you, alternatively, you can choose "Analysis Wizard" as specification method after you initially open GLM module, Click OK.

A dialog will pop out to ask you choose your dependent, independent and covariates variables.

After you have chosen your variables, you will get to this GLM Analysis Wizard Between Design dialog,

Click on "Custom between design" tab, you will be able to specify interaction terms.

Hope it helps.

Thank you.

All Replies
• Hi, Evslite,

In Statistica, to perform ANCOVA in GLM module,

Choose "Analysis of covariance" as analysis type in GLM module.

Click on the "Variable" tab and choose the dependent and independent and cofactors for analysis.

Click OK. Click on "Between effects" to specify the main effect, co-factor effect and interaction terms

Click OK to obtain the ANCOVA analysis result dialog.

In your case, you can create another variable age^2 as continuous predictor, and specify the interactions terms population*age and population*age^2 in the analysis.

Hope it helps.

• Hi Jenny L,

Thank you for the reply and directions, but unfortunately I am unable to get to the Between Effects window. The button is just greyed out in GLM, GRM and GLZ. This is after I've selected all of the factors and codes. It looks like my ANCOVA function is just allowing main effects ANOVA action!

Can you think of anything I might be doing incorrectly?

Thanks

Evan

• Hi, Evan,

There are two ways you can achieve it.

Method 1:

After you get to the dialog with the "Between effects" button inactive, first click OK to get to the GLM result dialog,

In the GLM result dialog, then click on "Modify" button,

this will direct you back to the GLM Analysis of covariance dialog, where the "Between Effects" button becomes active to allow you specify interaction terms.

Method 2:

If Method 1 does not work for you, alternatively, you can choose "Analysis Wizard" as specification method after you initially open GLM module, Click OK.

A dialog will pop out to ask you choose your dependent, independent and covariates variables.

After you have chosen your variables, you will get to this GLM Analysis Wizard Between Design dialog,

Click on "Custom between design" tab, you will be able to specify interaction terms.

Hope it helps.

Thank you.

• Although the modify button didn't help, the Analysis Wizard worked. Thank you very much!

Is there a reason or setting that the Between Effects button is inactive?

• Hi, Evan,

Glad you can resolve using the Analysis Wizard.

I have reported this issue of inactive Between Effects tab to our development team with report ID 106049. They will check on it when they have time.

Thank you.