I am doing ANCOVA: main categorical variable for the comparison is "Street" and it contains 3 categories (Street1, Street2 and Street3). The outcome variable is social interaction time (mins) of individuals in these streets. We controlled for time of the day (because social interaction may be affected by the time of the day it was observed in). Reference category was street3. Because the initial model suffers from nonnormality and non-homogeneity of variance, we used the fast-wild bootstrapping technique, and here was the result:
As we see there is a p-value for each of the two dummy variables. May I somehow get an overall P-value for the whole categorical variable (street), just like SPSS and some other software do?
Is it possible to calculate Wald = sum of the squared t-values of the two dummy variables, and compare it with Chi-square (0.05, df = 2)?
