6
Limiting the analysis here to just those who were part of the panel introduces some small amount of
bias into the findings, as it only includes those who made it through the class (and, to a lesser extent,
those who were willing to answer the survey questions). Of those who were not included in the panel,
the majority started the class and completed the first survey before dropping the course at some point.
A smaller number (~5) remained in the class all term but chose not to complete one of the surveys.
7
In his discussion of gender gaps in political knowledge, Mondak (1999) notes that one reason women
seem to know less than men is because women are less likely to take guesses when they don’t know the
answer. Men, who may occasionally guess correctly, are thus advantaged. To correct for that
possibility, we follow Mondak’s suggestion in using multiple choice questions and in encouraging
respondents to guess when they are not sure of the answer. If a respondent did leave a response item
blank, we randomly filled in a response (thus simulating a guess).
8
The largest gain among the textbook items, how the Chief Justice of the Supreme Court gets selected,
likely emerges because the confirmation of Chief Justice John Roberts occurred simultaneously with
the class and was a somewhat frequent topic of discussion.
9
These items are all highly correlated with each other. A factor analysis of these nine items showed
they all loaded very solidly on one factor. Even though the factor analysis could not tease out separate
dimensions of political skill, I continue to treat them as distinct elements here.
10
It is certainly debatable whether the item concerning skill in explaining views to others should be
placed in the category for managing information or working with people. I have placed it in the
former, as I believe the “I can persuade others to support political position item,” included under skills
of working with people, reflects the skill of being able to move someone to see things as you do.
“Explaining” one’s views to others, I believe, reflects more an ability to cull together useful
information to make a point – nothing about persuasion is implied in this item.
11
I obtained curious results for the item asking about skills in talking about social barriers such as race.
Men began the course feeling far more confident in their ability to do this, contrary to expectations that
women might be better at handling these sensitive matters. By the end of the course, women had
actually moved ahead of men in this area. These results clearly warrant further investigation down the
road.
12
Simulation grades were assigned on a 0-5 scale at the end of the simulation by the room facilitator.
Grades were checked by me; I had seen each simulation enough to have a sense if the grades seemed
realistic, which they almost always did. Grades were assigned long before any data analysis was done
to ensure that the grading was not affected by concerns about hypothesis testing.
13
A large racial difference was also noted in course grades, where white students averaged one full
letter grade higher than African-American students (3.1 versus 2.1, on a 4.0 scale; t=3.75, p<.000).
Convention