84 Views

Suicide and Self-Injury

Age, Race, and Sex-Stratified Demographic Variation of Lethal Means for Suicide in the United States Reduces to Two Dimensions

(PS2-D95) Age, Race, and Sex-stratified Demographic Variation of Lethal Means for Suicide in the United States Reduces to Two Dimensions

Friday, November 18, 2022

10:30 AM – 11:30 AM EST

Location: Westside Ballroom, 5th Floor

Author(s)

Joseph Sexton (he/him/his)

Undergraduate Student
Vanderbilt University
Nashville, Tennessee

Co-Author(s)

HW

Hao Wu, Ph.D.

Associate Professor of Psychology and Human Development
Vanderbilt University, Peabody College
Nashville, Tennessee

The mode of suicide may take several forms, not limited to drowning, hanging, self-immolation, and poisoning. Use of firearms for suicide carries the highest case-fatality rate, with roughly 90% of suicide attempts resulting in death. Poisoning, conversely, is utilized more frequently in suicide attempts but possesses an estimated case-fatality rate under 2%. Such disparities in lethality have frequently been used to explain sex differences in suicide rates, where males attempt less often but are still more likely to die by suicide. However, variation in suicide rate is also observed when comparing among racial or age groups. The current study aims to understand how varied demographics may adopt varied lethal means for suicide. Demographic data and cause of death for all deaths in the United States from 2009 to 2019 were compiled from the Centers for Disease Control and Prevention’s open-source National Vital Statistics System (N = 472,100). Correspondence analysis, an extension of chi-squared analysis and a means of dimension reduction for frequency data, was then used in an exploratory fashion to determine underlying variables in a number of contingency tables: 1) race vs. means; 2) intersectional race + sex vs. means; 3) age group vs. means; 4) intersectional age group + sex vs. means; 5) intersectional race + age group vs. means; and 6) intersectional race + age group + sex vs. means. According to qualitative assessment of associated scree plots, each contingency table could be reduced to one or two dimensions, and two dimensions explained over 90% of table variance in each case. These dimensions may be interpreted in terms of either demographic (row variables) or means (column variables). Dimension 1 seems to primarily distinguish between racial groups. Black and White demographics score higher in this dimension than American Indian / Alaskan Native (AIAN) and Asian demographics. Highly accessible methods (e.g., those available in the home, firearms, non-gas poisoning, stabbing) also score high on this dimension, indicating greater use in Black and White populations. Dimension 2 primarily distinguishes males and females, also explaining some variance in age groups. Males score higher on this dimension than females. Highly violent methods seem to generally score higher on this dimension as well, but this is not always true (e.g., gas poisoning ranks higher than jumping from a high place). In sum, correspondence analysis may be used to understand roughly how varied demographics utilize varied lethal means for suicide. This may have further implications in terms of targeted prevention efforts as well as the cognitive and cultural science undergirding suicidal decision-making. Future work may also rotate dimensions, in factor analytic fashion, to accomplish greater theoretical value.