Popular Culture Review
86
or not I wore a tie. Then I began compounding the exogenous variation (that is,
my attire)—^partly to make the exam question more interesting, partly from
growing curiosity about the explanatory power of attire, and partly, frankly, on
larks. In later semesters, I alternated wearing a necktie with a solid-colored shirt,
not wearing a tie but with a similar shirt, and wearing a floral-print button-up.
On several occasions, I even wore the necktie with a floral-print shirt.
For each of the lectures where data was collected, I did my best—while
delivering a lecture, managing a classroom of students, and, particularly in the
earlier of those eleven courses, learning how to deliver a lecture and manage a
classroom—to record simple counts of a small number of student behaviors that
I considered at the time to be inappropriate but which I now recognize as
variable reactions of dissonance and avoidance. In particular, I counted how
many students each class period left the classroom (whether for the restroom or
for the duration of the period; I did not distinguish), how many read a newspaper
(particularly common at the first institution where I taught), and how many fell
asleep (more common at the later two institutions). Ultimately, I observed a total
1,501 attending students and recorded 42 instances of students leaving, 36 of
students reading, and 9 of students falling asleep.
Table 1: Instances Observed
Instance
1
2
3
4
5
6
7
8
9
10
11
Institution
A
A
A
A
B
C
D
D
D
D
D
# Students
12
12
24
24
19
46
27
25
25
24
24
# Days
8
8
8
8
4
8
4
4
4
4
4
Although I had whimsically proposed to students on those first two
examinations that the exogenous conditions of my attire might make a
difference, I was surprised at the extent to which they make a systematic
difference. I lumped the three counts (leaving, reading, and sleeping) as a
composite measure; calculated, for each class day measured, that lump sum as a
percentage of students present; and then averaged those percentages according
to the exogenous variation (such as whether or not I wore a tie). As Table 2
below shows, the percentages were more than twice as high (11.5% vs. 5.0%)
when I wore a necktie as when I did not. In other words, wearing a tie more than
doubled the incidence of behaviors I had hoped to avoid or at least reduce.
Wearing a floral-print shirt was associated with a small (but statistically
significant) reduction over wearing a solid-colored shirt without a tie (4.5% vs.
5.9%), and a larger reduction when the floral-print condition is compared with
the conditions of a solid-colored shirt both with and without a tie (4.5% vs.
8.7%). The highest incidence of problematic behavior occurred on the two dates
when I wore both a floral-print shirt and a tie (12.5%)—which, I soon realized.