Testing Durkheim's Theory of Suicide
Testing Durkheim's Theory of Suicide
Testing Durkheim's Theory of Suicide
9 Springer-Verlag 1996
David Lester
D u r k h e i m (1897) p r o p o s e d that social forces w h i c h increase social integration should d e c r e a s e suicide, and research o v e r the p a s t century has s h o w n that m a r r i a g e and birth rates tend to be n e g a t i v e l y a s s o c i a t e d with suicide rates, both o v e r time and o v e r regions (Lester 1992). Recently, W i e d e n m a n n and W e y e r e r (1994) s h o w e d that in B a v a r i a f r o m 1865 to 1980 suicide rates were ass o c i a t e d with birth rates in the w a y that D u r k h e i m ' s theory predicts, but that m a r r i a g e rates were not s i g n i f i c a n t l y
D. Lester (N~) Center for the Study of Suicide, 5 Stonegate Court (RR 41), Blackwood, NY 08012, USA
a s s o c i a t e d with suicide rates. T h e present c o m m e n t explores this result using other nations. S u i c i d e rates were a v a i l a b l e f r o m the W o r l d H e a l t h Org a n i z a t i o n f r o m 1901 to 1988 for ten E u r o p e a n nations (see Table 1). M a r r i a g e and birth rates were a v a i l a b l e for this s a m e p e r i o d from M i t c h e l l (1992). T h e results o f a s i m p l e correlational analysis, a linear m u l t i p l e regression and a r e g r e s s i o n correcting for the serial autocorrelation in the data sets using the C o c h r a n e Orcutt technique (Doan 1990) are p r e s e n t e d in Table 1. It can be seen that eight o f the ten nations h a d negative correlation coefficients b e t w e e n suicide rates and birth rates, seven o f w h i c h were statistically significant. In contrast, six nations had n e g a t i v e associations b e t w e e n suicide rates and m a r r i a g e rates, only four o f w h i c h were statistically significant. Both o f these results are in the direction p r e d i c t e d b y D r u k h e i m ' s theory, but the associations for birth rates are stronger. In addition, the eight n e g a t i v e associations o f suicide rates with birth rates w e r e all larger than those with m a r r i a g e rates. The linear multiple regressions confirmed these results. However, it must be noted that correcting for the serial autocorrelation in the data sets eliminated the majority of the significant associations and produced little difference between the associations of suicide with birth and marriage rates.
Table 1 Correlational and regression results for time series analyses from 1900 to 1988 for ten nations England/Wales Finland Ireland Italy Netherlands Norway Scotland Spain" Sweden Switzerland
Correlation coefficients Marriage 0.21 b 0.04 0.01 0.05 _0.50 b _0.47 b -0.16 -0.26 b _0.35 b -0.02 Birth 0.05 -0.90 u -0.49 b 0.42 u -0.74 b _0.66 u -0.65 u -0.38 b _0.58 b -0.08
Regression coefficients (linear) Marriage 0.21 0.31 b 0.21 -0.08 -0.20 b _0.55 b -0.31 b -0.09 -0.25 b 0.04 Birth 0.02 -0.91 b -0.50 b 0. l0 b -0.19 b -0.37 b -0.28 b -0.06 b -0.31 b -0.06
Regression coefficients (corrected) Marriage 0.02 -0.31 0.05 -0.03 0.09 -0.30 b -0.03 -0.03 -0.25 -0.18 Birth 0.06 0.39 -0.16 0.08 -0.21 u -0.01 -0.03 -0.07 -0.03 0.34
"Spanish data were available only up to 1987 bStatistically significant at the 5% level or better