.16, p  .01 (see Table 6). A simpl

.16, p  .01 (see Table 6). A simple slopes analysis (Cohen et al.,
2003; Figure 4) revealed that when trainers held low pretraining
expectations of the training, trainer posttraining assessment of the
training was positively related to the trainee’s performance, 

0.67, t(37)
4.03, p  .001. That is, within the trainers who
originally held low pretraining expectations, as posttraining evaluations
increased, trainee score increased. In contrast, trainers who
reported the highest pretraining expectations of trainees demonstrated
no relationship between trainer expectations and trainee
performance, 
–0.06, t(37)
–0.28, p
.78 (e.g., Goranson,
1976; Griffin & Ross, 1991; Jacoby & Kelley, 1987).
In addition, we tested the role of trainee weight in the relationship
between trainer evaluations and trainee performance. Consistent
with Hypothesis 4, in the obese condition, the relationship
between trainer evaluation of the training and trainee performance
was extremely high (r
.67, p
.001); that is, as trainer
posttraining evaluations increased, so did trainee performance.
However, within the average-weight condition (where neither positive
nor negative expectations were anticipated), there was no
relationship between trainer evaluation of the training and trainee
performance (r
–.03, p
.91). In sum, this exploratory analysis
supports Hypothesis 4 for inflexible trainers: Trainees instructed
by trainers who expected less and maintained these negative attitudes
performed worse on the task. Furthermore, performance
decrements were not elicited by trainers who adjusted their perceptions
of the trainee after the training interaction.
Discussion
In this study, trainers instructed female trainees in a computer
task after viewing a photograph depicting the trainee as obese or as
average weight. After viewing the purported photo of the trainee,
trainers provided their expectations of the training interaction and
the trainee. We hypothesized and found that, relative to the trainers
in the average-weight condition, trainers in the obese condition
expected less of the training (Hypothesis 1a). In addition, compared
with trainers in the average-weight condition, female trainers
expected less success and a lower work ethic from the obese
trainees (Hypothesis 1b). Furthermore, female trainers evaluated
the trainee and the training more negatively in the obese condition
(Hypothesis 2b). We further hypothesized and found that these
results extended to the obese trainees, who had no knowledge of
the picture manipulation and no reason to suspect differential
treatment. Relative to the trainees in the average-weight condition,
trainees in the obese condition with female trainers evaluated the
trainer and the training more negatively (Hypothesis 3b). Although
trainees in the obese condition did not ultimately perform more
poorly on the task (Hypothesis 4), exploratory analyses supported
the emergence of a self-fulfilling prophecy for inflexible trainers.
Thus, the present findings suggest that self-fulfilling prophecies
may occur when trainers infer negative stereotypes from trainee
characteristics. First, lower expectations held by female trainers
based on trainee weight manifested into lower quality training as
reported by the trainees. This is consistent with previous research
on the communication of stigma. Subtle negative behaviors tend to
“leak out” during interactions (e.g., Babad, Bernieri, & Rosenthal,
1989; Barr & Kleck, 1995; Fazio, Jackson, Dunton, & Williams,
1995) and, in general, are recognized by the other interactant
(Dovidio, Gaertner, Kawakami, & Hodson, 2002). Second, although
we did not find direct evidence that trainer expectations
impeded knowledge transfer, we did find, consistent with selffulfilling
prophecy theory (Jussim, 1986), that for inflexible trainers,
lower expectations because of trainee weight led to decrements
in trainee performance. Thus, this study contributes to both the
stigma and training literature, demonstrating that trainer bias can
influence the perceived and actual quality of a training interaction.
In addition, this study is one of the first organization expectancy
studies to find the strongest results among female trainers. Consistent
with organizational research that suggests that, compared
with men, women may be more critical of obesity (Decker, 1987;
Pingitore et al., 1994) and of other women (Graves & Powell,
1995; King et al., 2005; Mathison, 1986; Staines et al., 1974), the
current study demonstrates that female trainers in the obese condition
expected less of trainees, were evaluated worse by their
trainees, and evaluated the training to be worse. Researchers have
repeatedly suggested the importance of exploring the consistent
and mysterious finding that female leaders do not generally elicit
organizational self-fulfilling prophecies (Dvir et al., 1995; McNatt,