Why making a CONTINUOUS covariate CATEGORICAL changes my results?

#1
Hello everyone,
I guess my question is stupid - I tried to figure it out myself, but couldn't, so please don't be too harsh :)

My study looks into how self- vs. others-view and engaging vs. disengaging emotions influence one's desire to announce one's helping behavior (continuous variable).

In a Univariate ANOVA, I have a significant interaction between view and emotion (p<.05), but the planned contrast of emotion is significant only for the self-view condition (p=.05) and not for the other-view (p>0.1).

One of the covariates that I tried is a continuous variable that measures one's predisposition to be helpful. No change. Interestingly however, if I make it a categorical variable of 2 levels, the contrast for the other-view condition becomes marginally significant (p=0.1) while the contrast for the self-view stays significant.

When I make a variable categorical, I decrease the variance in it, but how does this affect my significance level?

Thanks in advance!