2015 North American Society for Sport Management Conference (NASSM 2015)
Ottawa, ON June 2 – 6, 2015 Page 4
Applying the 3M Model of Personality and Motivation to Sport Event Emotional Responses
Minkyo Lee, Indiana University
Ju Young Lee, Indiana University
Paul M. Pedersen (Advisor), Indiana University
Communication Thursday, June 4, 2015 20-minute oral presentation
(including questions)
Abstract 2015-036 1:20 PM (Richelieu)
It is not a surprising fact that among the entertainment genres in televised programming sporting events often are
the favorite programs consumed by viewers. For these viewers, hedonic influences make up one of the main
motivators for this sportainment consumption (Hirschman & Holbrook, 1982; Raney, 2006). In line with the
hedonic paradigm, sport scholars have attempted to explore the relationship between various factors (e.g., biological,
psychological, cognitive, sociological) and the emotional responses of consumers (e.g., fans, viewers). For example, it
has been found that dispositional tendency (Zillmann, Bryant, & Sapolsky, 1989), suspense (Gan et al., 1997),
violence (Raney & Kinnally, 2009), gender (Sargent, Zillmann, & Weaver, 1998), game outcome (Bryant & Raney,
2000), and team identification (Wann, Royalty, & Rochelle, 2002) contribute to fans’ enjoyment of sporting contests.
Such theoretically based factors provide substantial explanations as to why people enjoy consuming televised sports.
However, one area needing scholarly examination involves how individual personality differences– which account
for plentiful variance in emotional feeling (Mowen, 2000) – interact with fans’ affective states.
Theoretical Perspective
The Meta-theoretic Model of Motivation (3M) provides a combined explanation of how personality traits interrelate
with situations and thus impact emotions, thoughts, and behavioral tendencies of individuals (Mowen, 2000). This
hierarchical model has a unique feature which contains both the broad and the specific level traits that influence each
other. The 3M conceptual framework provides the desirable and appropriate variables at each hierarchical level (i.e.,
elemental traits, compound traits, situational traits, and surface traits) to predict a specified criterion variable. This
model has been widely applied to various industries and areas ranging from advertisement and volunteerism to
traveling, gambling, and consumer behavior (e.g., Mowen, 2004; Mowen & Sujan, 2005; Mowen, Harris, & Bone,
2004).
The present study focuses on element and compound traits to predict emotional responses to a mediated sporting
event. Elemental traits (e.g., agreeableness, openness to experience, need for arousal, need for health, need for body
resource, introversion) are termed as the fundamental tendencies that result from genetics and a person’s early
learning (Mowen, 2000). Compound traits (e.g., competitiveness, impulsiveness, need for play) are classified as crosssituational predispositions to behave that develop from the interaction of elemental traits, culture, and the
individual’s prior learning. The research question assessed in the current research dealt with how personal traits (i.e.,
elemental and compound traits) influence fans’ emotional reactions to a sporting contest.
Method
A survey involving elemental and compound traits scales (Mowen, 2000) and a team identification scale (Robinson &
Trail, 2005) was distributed to 187 participants recruited from various sport management classes at a large
Midwestern university in the United States. The subjects were then randomly exposed to one of four different
manipulated versions of a video clip (an NBA contest) in a 2 (game outcome: win [lose] or lose [win] game) 2
(uncertain outcome: close or lopsided game) between-subjects design. After watching the video clip, emotional
responses to stimuli were measured with bipolar pairs of adjectives (e.g., enjoyable, entertaining, pleased).
Participants (n = 36) not affiliated with either team, univariate and multivariate outliers (n = 3), or with missing data
(n = 11) were excluded in the final data set. Also, no problematic multicollinearity was found. Thus, a total of 137
participants – greater than the 109 suggested by a priori power analysis (Faul, Erdfelder, Buchner, & Lang, 2009)
with a power of .80, an alpha of .05, and an effect size of .15 (a medium effect) – were included in the final data
analysis.