Jeffrey D. Holmes
278
scale construction. Each of the small groups should research one of three concise personality
traits. I use extraversion, openness to experience, and optimism because they are relatively
easy for students to grasp. Conversely I avoid assigning clinical symptom constructs, such as
anxiety and depression, to prevent student discomfort and the disclosure of information that
respondents might later regret. A lab assistant can offer guidance during students’ search of
Web-based or other literature to better understand and conceptualize their trait.
Following this background research and discussion of the literature, each student pro-
duces two or more original items to reflect the target trait. Instructors should provide
guidelines for effective item writing such as the importance of clear and concise language,
avoiding double negatives, and the advantages of using Likert scales to increase variability
(see Clark & Watson, 1995, for a brief review of test construction guidelines). My lab
groups initially produce 18–20 items each. The group members then debate the items
until they agree on the five that best represent the trait. All groups complete this process
independently, resulting in multiple measures of each trait. This procedure permits later
analyses of both intertrait and intratrait relationships.
Session Two: Data Collection
Prior to the second session, a student assistant combines each scale into a single instru-
ment for administration (Web-savvy instructors could put the items online to eliminate
the need for data entry). The assistant or instructor must maintain an item key to allow
for calculation of scale scores later. All students respond to the items written by all groups
as well as several online single-trait personality measures. For the latter, I typically use
scales measuring perfectionism, locus of control, sensation seeking, and Type A behavior.
An instructor might substitute other widely available scales (it would be impractical to list
Web sites here because Web addresses change frequently, but a brief search will yield many
options). The extra scales add complexity to the activity but could be omitted if necessary.
Because students can quickly complete the scales, they spend the remaining class time
researching their assigned traits (instructors might also assign this step as a homework
activity). Each student should retrieve empirical articles reporting correlations between
the assigned trait and other psychological constructs. The students use these sources to
generate hypotheses and later to refine their interpretation of the data.
Session Three: Analysis and Interpretation
After students have responded to the scales, an assistant compiles the data and the
students perform statistical analyses. The specifics will vary depending on the level
and nature of the course. In my introductory courses, I assign descriptive and corre-
lational analyses. Students examine gender differences in trait scores, as well as the
interrelations between similar and distinct trait scales. One of the most interesting
and pedagogically valuable lessons involves intratrait correlations. Typically these