478 Chapter 12
were higher in convenience samples than
representative random samples. This find-
ing suggests that the people who volunteer to
be in studies of sexual harassment are more
likely to have been harassed. Thus to obtain a
good estimate of the frequency of sexual ha-
rassment, it is very important to have a rep-
resentative sample of participants and a high
response rate.
Good studies with representative
samples show that about half of women
and a substantial number of men experi-
ence unwanted sexual harassment at work
(Stockdale & Bhattacharya, 2009). The meta-
analytic review reached similar conclusions
but also noted that the incidence of harass-
ment depended on how it is assessed (Ilies
et al., 2003). When people are explicitly
asked in a survey if they have been sexually
harassed, fewer people report harassment
than when they are asked if any of a series
of behaviors such as those shown in Table
12.4 have occurred. Representative samples
using the first method revealed that 24%
of people reported harassment; the second
method yielded an estimate of 58%. Higher
percentages of harassment occur for milder
forms, such as sexual remarks (Stockdale &
Bhattacharya, 2009). Of all types of organiza-
tions, harassment rates were highest among
military samples (Ilies et al., 2003).
What is the incidence of sexual harass-
ment on college campuses? A 2005 nation-
ally representative survey of undergraduates
showed that sexual harassment is prevalent,
with two-thirds of students reporting some
kind of harassment (Hill & Silva, 2005),
which is consistent with another recent re-
port of 57% (Huerta et al., 2006). A major-
ity of harassment incidents consist of verbal
behaviors, such as lewd comments, jokes,
sexual innuendoes, and remarks about body
parts. Female and male students are equally
also associated with perceiving the person in
more favorable terms, which leads to a be-
havior not being interpreted as harassment
(Sheets & Braver, 1999). Thus if a high-status
person engages in an ambiguous behavior,
such as making sexual innuendoes, we will
be less likely to interpret the behavior as ha-
rassment if we know and like the person than
if we do not know and/or do not like the per-
son. This fits with other research that shows
we judge harassment by attractive men less
harshly than harassment by unattractive men
(Golden, Johnson, & Lopez, 2001). Thus, if
we like the person due to his or her status or
physical attractiveness, we will be less likely
to infer sexual harassment.
Incidence
In 2009, 12,696 charges of sexual harassment
were filed with the EEOC (U.S. Equal Em-
ployment Opportunity Commission, 2010).
A majority of charges were filed by women;
16% were filed by men.
How many people have experienced
sexual harassment during their working
history? The prevalence of sexual harass-
ment is typically measured with surveys.
One methodological difficulty with the sur-
vey method is that only a subset of people
complete them. The people who complete
the survey differ from the people who do
not complete the survey. The kind of person
who responds to a survey is likely to be in-
terested in the topic, and it makes sense that
the people who will be most interested in the
topic of sexual harassment are those who
have experienced it. There is evidence that
such a response bias exists.
A meta-analytic review of the litera-
ture showed that the sampling technique in-
fluenced reports of sexual harassment (Ilies
et al., 2003). Reports of sexual harassment
