I am running a logistic regression analysis to examine risk factors that predict the experience of intimate partner violence.

One of my risk factors is educational attainment (4 levels). The overall model is not significant (p=0.069) but when I look at the coefficients table, the coefficients for two of the educational attainment categories are significant. How can I interpret this? Would I include this predicor in a multivariate model?

Thanks in advance!