You are on page 1of 3

Psychological Reports, 2008, 102,575-576.

O Psychological Reports 2008

COMPUTER ANALYSIS O F THE CONTENT O F SUICIDE


NOTES FROM MEN AND WOMEN '

DAVID LESTER
The Richard Stockton College of New Jersey

Summary.-A text analysis of 47 suicide notes identified n o reliable or psycholog


ically meaningful differences by sex, age, or method used for suicide.

There are strong and clear differences in suicidal behavior. Women attempt suicide more
than do men, whereas men complete suicide more than do women (Canetto & Lester, 1995).
Women use different methods for suicide than do men, preferring medications more and hang-
ing and firearms less (Lester, 1984). It is thought, perhaps erroneously, that the motives behind
the suicidal act differ, with women committing suicide over love crises and men for achieve-
ment-oriented crises (Canetto, 1992-1993).
Despite these differences, studies of the suicide notes left by women and men have identi-
fied few differences. In extensive study, Leenaars (1987, 1988a, 1988b) found no differences in
content, and Lester and Hummel (1980), Lester (19891, Lester and Heim (19921, Linn and
Lester (1996), and Lester and Linn (1997) also found no differences. Black (1989) found more
indications of depression, self-hostility, confusion, and despondency over the death of others in
the suicide notes of women, while Lester and Reeve (1982) found that notes from women had
fewer negative emotions but were more disorganized. All of these studies used judges to rate
the suicide notes.
In the present study a series of suicide notes was examined using a computer program
written to make exact counts of particular kinds of words, a technique first used by Ogilvie,
Stone, and Shneidman (1969). Pennebaker, Francis, and Booth (2001) have developed a text
analysis program (known as LIWC-Linguistic Inquiry and Word Count) that analyzes written
text on a word-by-word basis for over 70 language variables. The program can count pronouns
(for example, "I" and "me" versus "us" and "our"), emotional words (negative versus positive),
and particular words (such as "death"). This program has been used successfully to explore
differences in the suicide notes written by completed and attempted suicides, with the notes
from completed suicides having more future tense verbs and more positive emotions (Handels-
man & Lester, 20071, and also in the diary of a young woman who killed herself (Pennebaker
& Stone, 2004). The present study examined a series of 47 suicide notes for differences in con-
tent by sex using the text analysis program, LIWC.
Method.-A series of suicide notes from 47 completed suicides from a town in the north-
eastern USA was obtained from the medical examiners' files. There were 33 men and 14 wom-
en whose mean age was 43.2 yr. (SD= 17.7). The notes were scored by the LIWC program for
88 characteristics. The program presents the scores as percentages of the total number of
words.
Results and diJcussion.-Only three significant differences were found in the 88 compari-
sons at the two-tailed statistical significance of p 5 .05, no more than would be expected on the
basis of chance with so many variables. The 14 women used numbers more (point-biserial cor-
relation = .3 1, two-tailed p = .03), words related to hearing more ( r = .29, p = ,051, and fewer dic-
tionary words ( Y = -32, p= .03).

'Address enquiries to David Lester, Ph.D., Psychology Program, The k c h a r d Stockton College
of New Jersey, Jimmie Leeds Road, Pomona, NJ 08240-0195 or e-mail (david@stockton.edu).
D. LESTER

Incidentally, only three significant differences were found for age: long words (more than
six letters) (Pearson r = .45, p = .002), religious words (r = 3 3 , p = ,031, and apostrophes ( r =
-.32, p = ,041. Comparing those using pills ( n= 14) versus firearms ( n= 151, those using firearms
scored lower on the use of words such as would, should, and could (point biserial Y = -.45, p =
.01). Again, these few differences are no more than would be expected by chance with so many
statistical tests.
Overall, therefore, the present study showed no reliable differences in the content of sui-
cide notes by sex (and, in addition, by age and by method), and the few differences identified
had little or no psychological significance. However, the sample size was small and so power
was low.

REFERENCES
BLACK,S. T. (1989) Gender differences in the content of genuine and simulated suicide notes.
In D. Lester (Ed.), Suicide '89. Denver, CO: American Association of Suicidology. P. 8.
CANETTO, S. S. (1992-1993) She died for love and he for glory. Omega, 26, 1-17.
CANETTO, S. S., &LESTER, D. (1995) Women and suicidal behavior. New York: Springer.
HANDELSMAN, L. D., &LESTER,D. (2007) The content of suicide notes from attempters and
completers. Crisis, 28, 102-104.
LEENAARS, A. A. (1987) An empirical investigation of Shneidman's formulation regarding sui-
cide. Suicide & Life-Threatening Behavior, 17, 233-250.
LEENAARS, A. A. (1988a) Are women's suicides really different from men's? Women C flealth,
14, 17-33.
LEENAARS, A. A. (1988b) Suicide notes. New York: Human Sciences.
LESTER,D. (1984) Suicide. In C. S. Widom (Ed.), Sex roles and p~ychopathulogy.New York:
Plenum. Pp. 145-156.
LESTER,D. (1989) Sex differences in the motives expressed in suicide notes. Perceptual and
Motor Skills, 69, 642.
LESTER,D., & HEIM,N. (1992) Sex differences in suicide notes. Perceptual and Motor skill.^,
75, 582.
LESTER,D., &HUMMEL, H. (1980) Motives for suicide in elderly people. Psychological Reports,
47, 870.
LESTER,D., &LINN,M. (1997) Sex differences in suicide notes. Psychologzcal Reports, 80, 1302.
LESTER,D., &REEVE,C. (1982) The suicide notes of young and old people. Psychological Re-
ports, 50, 334.
LINN,M., &LESTER,D. (1996) Content differences in suicide notes by gender and age. Psycho-
logical Reports, 78, 370.
OGILVIE, N ,S. (1969) Some characteristics of genuine versus
D. M., STONE,P. J., & S H N E I D ~E.
simulated suicide notes. Bulletin of Suicidology, March, 27-32.
PENNEBAKER, J. W., FRANCIS, M. E., &BOOTH,R. J. (2001) Linguistic inquiry and word count.
Mahwah, NJ: Erlbaum.
PENNEBAKER, J. W., &STONE,L. D. (2004) What was she trying to say? In D. Lester (Ed.), Ka-
tie's diary. New York: Brunner-Routledge. Pp. 55-79.

Accepted March 24, 2008


576E

ERRATUM

LESTER, D. (2008) Computer analysis of the content of suicide notes from men and
women. Psychological Reports, 102, 575-576.

The above study had 48 participants. There was an error in the file type run
through the program, so the results require small corrections as follows. The Results and
Discussion section should read, “Only three significant differences were found in the 88
comparisons at the two-tailed significance of p =.05 or better, no more than would be
expected on the basis of chance with so many variables. The 14 women used numbers
more (point-biserial correlation =.31, two-tailed p = .03), words related to hearing more
(r = .29, p = .04), and fewer dictionary words (r = –.29, p = .03). Incidentally, only three
significant differences were found for age: long words (more than six letters) (Pearson r =
.45, p = .001), religious words (r = .33, p = .03), and negative emotions (r = .57, p =
.006).”

Erratum published in:


Psychological Reports, 2008, 102, 964. © Psychological Reports 2008