IS THE INVERTED-U HYPOTHESIS REALLY A "CATASTROPHE"?
DANIEL M. LANDERS AND MARC LOCHBAUM
ARIZONA STATE UNIVERSITY
Among the criticisms lodged against the inverted-U relationship between
arousal and performance has been that it "does not fit anecdotal evidence
from the field of sport psychology which suggests that this relationship
should not be symmetrical" (Hardy, 1990). Based on this "impression
of the literature," Hardy and Fazey (1987) created a complex multidimensional
"catastrophe" theory, which they believed better explained their belief
that when performers go over the top of the inverted-U curve, performance
"drops dramatically rather than gradually." The purpose of this study
was to demonstrate that Hardy and Fazey's (1987) initial belief was flawed
by attributing lack of asymmetry to theoretical shortcomings instead of
more obvious methodological limitations associated with tests of the inverted-U
hypothesis. Studies in this research literature have only manipulated
between 3-5 levels of arousal and most often these levels have not been
normalized to maximum arousal. To illustrate what would happen
to the shape of the inverted-U when seven levels of arousal are manipulated
(13, 25, 40, 50, 75, 89% of maximum arousal), a secondary analysis of 62
effect sizes (N = 310) derived from 7 cognitive performance studies was
undertaken. The results of this analysis showed that the performance
decrement following 50% of maximum arousal (i.e., the right side of the
inverted-U) resembled a more gradual decline instead of the "cliff-like"
drop that is predicted by catastrophe theory with subjects' cognitive anxiety
is dominant over self-confidence. It is concluded that when methodological
problems associated with tests of the inverted-U hypothesis are eliminated,
the overall results summed across anxious/nonanxious subjects shows no
support for the fundamental premise underlying predictions of catastrophe
theory. It is recommended that future research testing the inverted-U
hypothesis and catastrophe theory employ research designs which include
high and low congitively anxious subjects and also include at least seven
or more levels of arousal that are normalized to a well-recognized stardard
of maximum arousal or tension.
Journal of Sport & Exercise Psychology: 1998 NASPSPA Abstracts,
20(June Supplement), S16.