Assessing outliers in a proportion

#1
Hi all,

Just wondering how one would assess outliers in the following:

At the start of a two week period, people were asked to nominate a number of good deeds that they would perform in the next two weeks (N). At the end of the two weeks, they report how many they actually perform (C). There were two conditions, a control and an intervention to increase pro-social/helping behavior. The effect of condition on good deeds will be analyzed with logistic regression.

Note. each person could nominate however many deeds (N) they wished. In all cases, for each participant C <= N. Therefore the distribution of proportions was [0, 1].

Do outliers need to be assessed in this situation, and if so, how?

Thanks in advance everyone!

Al

In case it is relevant, this is the SPSS syntax for the regression:
GENLIN C OF N BY TREAT (ORDER=DESCENDING)
/MODEL TREAT INTERCEPT=YES
DISTRIBUTION=BINOMIAL LINK=LOGIT
/CRITERIA METHOD=FISHER(1) SCALE=PEARSON COVB=MODEL MAXITERATIONS=100 MAXSTEPHALVING=5
PCONVERGE=1E-006(ABSOLUTE) SINGULAR=1E-012 ANALYSISTYPE=3(WALD) CILEVEL=95 CITYPE=WALD
LIKELIHOOD=FULL
/MISSING CLASSMISSING=EXCLUDE
/PRINT CPS DESCRIPTIVES MODELINFO FIT SUMMARY SOLUTION.
 
#2
I would think you'd assess whether there are outliers in both C and N just like you would any other continuous variable by using the Explore procedure to generate a boxplot for your control and intervention groups. Unless I'm misunderstanding your question?