# ANCOVA results & covariate confusion

#### morepil

##### Member
This is a bit hard to explain so please bear with me till the end. Thanks for the effort.

I have 5 IV's and 1 DV.
I'm trying to find how much each IV influences the DV.

I run 5 ANCOVA's. In each of them I have one of the IV's as an IV, and the other 4 IV's as covariates b/c the 5 IV's are all closely related and research shows they can be/they are covariates..
So the IV's are not always IV's.
I understand that theoretically the model of the ANCOVA should be the same whether they are covariats or not.
Here are the results of two of the ANCOVA's. In the first one, for example, the actual IV, C_Fih_Total has a different F and Sig value then the second analysis, in which it is a covariate.
Researchers often cite doing an ANCOVA like this.. That amongst these 5 "IV's" researchers take 1 of the 5 variables as an IV while controlling for the other 4. They report 1 F and Sig value.
However, as can be seen from the chart below, these numbers change. So I was a bit worried that I was doing something wrong. But I guess it is OK that the F and sig value are different and that you only report the F and sig value for the actual IV, i.ie, the bottom variable in each analysis?
I really don't care about reporting the effect or F or Sig value of the covariates (B/c research doesn't). or should I be ?
So for example in citing the result of an ancova I can say that C_Fih_Total is related to PAR (F=2,67, p=,002) while controlling for cinsiyet, C_Dis_Total, C_Dih_Total, C_Fis_Total, and C_Cis_Total. Even though in the second analysis C_Fih_Total has a F and p value that is different?

#### Dason

What exactly is the difference between IV and covariate in your mind?

#### Dason

It seems SPSS will treat "covariates" as continuous no matter what. What does your data look like? Do you want these to be treated as continuous or as categorical?

#### morepil

##### Member
Hi. Thanks for the response.
So the difference between an IV and covariate is the fact that I am actually interested in the IV's At least in this case.

All but the DV and 1 of the covarites which is always a covariate in all the analyses are continious. DV and cinsiyet are categorical.
My data is positively skewed

So it doesn't look like the reason. Does it?

The DF's seem to change.. Somehow SPSS does not treat covariates and fixed factors (IVs) the same way, it seems.
Right?

#### morepil

##### Member
It seems SPSS will treat "covariates" as continuous no matter what. What does your data look like? Do you want these to be treated as continuous or as categorical?
They are all continuous in my mind Both the covarites and the fixed factor. i.e, there must be something else, no?

#### Dason

The models aren't treating your "IV" as continuous though. You can tell since they have more than one df for those variables

#### morepil

##### Member
I see. It does not treat the real IV as a continuous variable.. I just found ANCOVA is normally done with a categorical IV I'll try converting my IV to categorical and see what happens