PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
Williams et al. / NEUROTICISM AND SYMPTOM REPORTS
Examination of the Neuroticism–Symptom Reporting
Relationship in Individuals With Type 2 Diabetes
Paula G. Williams
Washington State University
Craig R. Colder
State University of New York at Buffalo
James D. Lane
Cynthia C. McCaskill
Mark N. Feinglos
Richard S. Surwit
Duke University Medical Center
The current study utilized a within-subject, experience sampling
methodology (ESM) to examine the relationship between
neuroticism (N) and physical symptom reports. Individuals
with type 2 diabetes monitored diabetes-related symptoms, rated
negative and positive affect (NA and PA), estimated their blood
glucose (BG) levels, and tested their actual BG levels with a
glucometer four times per day for 7 days. Multilevel modeling
analyses indicated that N, NA, and PA were related to reported
symptom frequency. Neuroticism moderated the relation between
PA and symptom reports: Lower PA was more strongly related to
symptom reports among high-N individuals. In addition, there
was evidence that symptoms mediated the relationship between N
and state NA. Finally, N was related to overestimation of BG,
beyond that accounted for by state NA. Results are discussed with
respect to potential effects of N on the processing of negative self-
relevant information and on self-regulatory behavior in health
contexts.
In recent years, a major focus of research on personality
and health has been on the relationship between
neuroticism (N) and other anxiety-related dispositions
(e.g., trait anxiety, trait negative affectivity [NA]) and
both self-assessed and objective health. Initial reports on
these relationships (e.g., Costa & McCrae, 1987; Watson
& Pennebaker, 1989) noted that N may bear a stronger
relationship to perceptions of health status than to
actual health status. One consequence of this sugges-
tion, however, has been a shift toward conceptualizing N
as a “confound” in health research and controlling for it
statistically. Certainly, in some instances, controlling for
the effects of N is warranted. However, further research
is still needed to fully understand the mechanisms by
which N is related to both objective and subjective
health.
A second consequence of previous research on the
relation between N and symptom reports has been a ten-
dency to dismiss the utility of self-report measures of
health. Whereas the accuracy of self-report measures of
health status compared to “objective” (e.g., immune
functioning) measures warrants evaluation, one can
argue that self-report measures are important in their
own right, regardless of their veridicality. After all, when
we make important self-care decisions (e.g., when to take
medication, when to go to a physician, when to stay
home from work) we use our subjective impression of
our health and rarely do we check it against an objective
measure. Moreover, the manner in which we describe
our health to practitioners affects diagnosis and treat-
1015
Authors’ Note: Portions of these data were presented at the 21st annual
meeting of the Society of Behavioral Medicine, April 2000, Nashville,
Tennessee. This research was supported by a National Research Service
Award (Grant 5T32-MH-19109) and National Institute of Health
Grants 1 R01 DK 49066 and K05 MH00303. One-Touch II blood glu-
cose meters were provided by Lifescan, Inc., Milpitas, California.
Thanks to Greg Bailey, Irena Persky, Cori Scalzo, Michelle Wasserman,
and Venette Westhoven for assistance with data entry and checking.
Special thanks to Don Hedeker for invaluable advice on multilevel
modeling. Correspondence concerning this article should be ad-
dressed to Paula G. Williams, Department of Psychology, Washington
State University, Pullman, WA 99164-4820; e-mail: pwillms@wsu.edu.
PSPB, Vol. 28 No. 7, July 2002 1015-1025
© 2002 by the Society for Personality and Social Psychology, Inc.