Smart and suicidal? The social ecology of intelligence and suicide in Austria

The big g‐factor of national cognitive ability

The response deals with several controversial issues: theoretical concepts of cognitive abilities, their cultural relativity in definition or level, the meaning and validity of national cognitive ability, methodological questions like the ecological fallacy, the variance of intelligence at different levels of observation, multi‐level analysis, the correctness and importance of levels of analysis in cognitive‐ability research, the aggregation and adjustment process, and the similarities of different cognitive assessment approaches. Central to this research are questions of causality (the causes and consequences of national cognitive‐ability homogeneity and level), of malleability of these levels, and of ethical and political consequences of intelligence research. Copyright © 2007 John Wiley & Sons, Ltd.

The association between premorbid cognitive ability and social functioning and suicide among young men: A historical-prospective cohort study

Previous studies have found associations between low cognitive ability and later completed suicide. The aim of this study was to examine the association between cognitive ability and social functioning in adolescence, and later completed suicide in a large population-based longitudinal study. Data from the Israeli Draft Board Register for 634,655 Israeli male adolescents aged 16 and 17 was linked to a causes-of-death data registry, with a mean follow-up of 10.6 years for completed suicide. Our results show that in males without a psychiatric diagnosis, both low (adjusted HR=1.51, 95% CI: 1.19-1.92) and high (adjusted HR=1.36, 95% CI: 1.04-1.77) cognitive ability, and very poor (adjusted HR=2.30, 95% CI: 1.34-3.95) and poor (adjusted HR=1.64, 95% CI: 1.34-2.07) social functioning were associated with increased risk for later completed suicide; however positive predictive values were low (PPVs=0.09% and 0.10%, for low cognitive ability and very poor or poor social functioning, respectively). No association between cognitive ability or social functioning and risk for suicide was found in males with a psychiatric diagnosis. These data do not support the clinical utility of screening for such potential predictors.

Regional Intelligence and Suicide Rate: New Data for Australia and a Synthesis of Research

Previous research has shown for the most part positive correlations between intelligence and suicide prevalence on the national level. However, this study found proxies for regional intelligence in Australia (international average domain scores from the PISA 2000 study) to be significantly negatively correlated with the total, male, and female suicide rates of the different administrative divisions of Australia, and this finding was independent of regional wealth. A research synthesis of the current results and those from similar studies of other countries (positive correlations for Austria, Belarus, The British Isles, Denmark, and The Netherlands; inconclusive findings for France, Germany, and the USA) was conducted. This synthesis of research findings showed that positive ecological correlations of intelligence with suicide rate were more likely observed for nations with higher suicide rates and poorer general living conditions, whereas there was no relation with national IQ.

State IQ and Suicide Rates in the United States

Previous findings pertaining to the direction (positive or negative) of the ecological correlation between intelligence and suicide rates in the USA have been conflicting. Using novel state IQ estimates, derived from the Scholastic Assessment Test, the American College Test, these tests combined, or the National Assessment of Educational Progress state scores, these estimates were not consistently associated to state suicide rates. Whereas ACT-derived state IQ was significantly positively correlated with suicide rates, the correlation with composite ACT-SAT-derived state IQ was significantly negative and with both SAT-derived and NAEP-derived state IQ also negative but not significant. Validity checks pointed to possible methodological problems with the state IQ estimates. Currently available estimates of state IQ, therefore, seem not appropriate to resolve the question of the direction of the ecological correlation of intelligence and suicide mortality across the USA.

Evidence for a Positive Ecological Correlation of Regional Intelligence and Suicide Mortality in the United States during the Early 20th Century

Several contemporary cross-national and intranational geographic studies have reported positive ecological (group-level) associations of intelligence and suicide mortality. These findings are consistent with facts from suicide research and with an evolutionary view of suicidal behavior. The present research extended these accounts cross-temporally. Analysis of E. L. Thorndike's state-level personal quality scores and standardized birth rates of eminent persons, taken as proxy variables for regional intelligence, along with historical state suicide rates (1913–1924 and 1928–1932) showed that intelligence and suicide mortality across the USA were already clearly positively related during the early 20th century, suggesting time stability of the effect. Within the USA, the effect is possibly due to state differences in the ethnic composition, which correspond to both suicide rates and intelligence proxies. It is argued that the most parsimonious interpretation of these ecological findings remains that they indeed reflect individual-level effects, that a positive link between intelligence and suicide is entirely compatible with positive overall links between intelligence and health and longevity, and that the ultimate explanative background for the positive link between intelligence and suicide may be provided through the framework of Rushton's differential K theory.

Regional Intelligence and Suicide Rate in Germany

Regional intelligence proxies (intranational average domain scores from the PISA 2000 and 2003 studies) were not reliably associated with the suicide rate across the states of Germany, thereby not replicating findings of a positive ecological correlation of regional intelligence and suicide prevalence from other within-nation studies (Austria, Belarus, the British Isles, Denmark, The Netherlands, and the USA).

National Differences in Intelligence and Population Rates of Suicidal Ideation, Suicide Plans, and Attempted Suicide

Population rates of suicidal ideation, suicide plans or attempted suicide from three independent multinational epidemiological surveys (the WHO/EURO and WHO SUPRE-MISS studies and a third one) were not significantly associated with national IQ figures. This result conflicts with previous evidence from cross-national studies (by Lester and by Voracek) of a positive ecological correlation between level of national intelligence and rates of completed suicide. Across nations, the indicators for the prevalence of suicidal behavior also lacked close correspondence with suicide rates, which may be due to the higher unreliability of assessing covert suicidal behavior (suicide ideas, thoughts, plans, and, to a lesser extent, also suicide attempts) as compared to suicidal behavior that is indisputably overt (completed suicide). Several alternative explanations for the current nil findings are also discussed.

Regional Intelligence and Suicide Rate in Denmark

Consistent with evidence from several recent geographical (cross-national and within-nation) studies (by Lester and by Voracek), a positive ecological (aggregate-level) correlation of regional intelligence and suicide rate was found across the seven major geographical regions of Denmark.

Educational Attainment, Intelligence, Interstate Migration, and Suicide Rates in the United States: Rejoinder to Abel and Kruger (2005)

Abel and Kruger (2005) reported a negative association of educational attainment and suicide rate across the United States. Given strong links between intelligence and educational attainment, this appears to be inconsistent with positive associations of intelligence and suicide rate reported in several other geographical studies. However, the apparent inconsistencies may reflect interstate migration within the U.S. adult population. Testing this hypothesis showed that interstate migration was more strongly related to state suicide rates than educational attainment. Thus, the different results obtained by Abel and Kruger could reflect migration effects, which for the U.S. might render adult-age indicators of educational attainment inappropriate to reflect regional variation in intelligence by place of birth.

National Intelligence and Suicide Rate across Europe: An Alternative Test Using Educational Attainment Data

National mean scores on a historical knowledge test (taken as a proxy for intelligence), stemming from representative samples of male (and female) 9th-grade school students from 26 European countries in a 1999 report by Wilberg and Lynn, were significantly positively associated with the national male (and female) suicide rates, independent of the general quality-of-living conditions in these countries. This finding replicates previous evidence from cross-national studies (by Lester and by Voracek), of a correspondence of higher national IQ to higher suicide rates, with an alternative measure of national IQ that is independent of the national IQ estimates recently published by Lynn and Vanhanen which have been used in prior studies.

The Nexus of Suicide Prevalence, Helping Behavior, Pace of Life, Affluence, and Intelligence: Contrary Results from Comparisons across Nations and within the United States

This study examined interrelations among suicide prevalence, affluence, intelligence, pace of life, and helping behavior on the level of aggregate analysis, both cross-nationally and within the USA, using four extant data sets (8, 15, and 29 countries; and 17 U.S. states) for the latter two variables. Pertaining to the association of suicide rates with pace of life, the results replicated discrepant previous findings of Lester. Across nations, suicide is more prevalent in populations living in faster environments, whereas within the USA suicide rates are higher in slower environments. Further, seen from an international perspective, there appears to be a nexus of variables, such that more people tend to commit suicide in the more affluent and faster environments which are populated by more intelligent, but less friendly, individuals. Within the USA, the direction of the correlations of suicide prevalence with these four variables is exactly reversed. Possible reasons for these opposed findings and implications for research are discussed.

Regional Intelligence and Suicide Rate in Germany, Revisited

Across the 10 westerly and 6 easterly states of Germany, better-than-average school-end grades, taken as a proxy for intelligence, corresponded to higher regional suicide rates. This finding adds to similar evidence from other countries (Austria, Belarus, British Isles, Denmark, Netherlands, and USA), suggestive of positive ecologic (group-level) associations between indicators for cognitive ability and suicide prevalence, whilst offsetting prior inconclusive evidence for Germany from Voracek (2006) on the same question.

The relationship between elderly suicide rates and different components of education: a cross-national study

Background:

Two recent studies reported a curvilinear (U-shaped) between elderly suicide rates and educational attainment measured by the United Nation’s Education Index. A study examining the curvilinear (U-shaped) relationship between elderly suicide rates and the individual components of the Education Index (adult literacy rate, percentage of children of relevant age group enrolled in primary schools and percentage of children of relevant age group enrolled for secondary schools) and one other measure of educational attainment (youth literacy rate) was undertaken to partial out the effects of the individual components of Education Index on elderly suicides.

Methods:

A cross-national study examining the relationship between elderly suicide rates (Y-axis) and different measures of educational attainment (X-axis) was undertaken using data from the World Health Organization and the United Nations data banks using Curve estimation regression models.

Results:

The relationship between elderly suicide rates with the adult literacy rate, the percentage of children of relevant age group enrolled for secondary schools and the youth literacy rate was curvilinear (U-shaped curve). This relationship was absent with the percentage of children of relevant age group enrolled in primary schools.

Conclusion:

Given the cross-sectional study design, a causal relationship between elderly suicide rates and measures of educational attainment, including the adult literacy rate, the percentage of children of relevant age group enrolled for secondary schools and the youth literacy rate, cannot be assumed. However, the findings suggest that future studies of elderly suicide rates and educational attainment should focus on the adult literacy rate, the percentage of children of relevant age group enrolled for secondary schools and the youth literacy rate as measures of educational attainment.

Predictors of National Suicide Rates: A Reply to Voracek (2004, 2006, 2009)

Voracek (2004, 2006a, 2006b, 2009) reported that cognitive ability predicts national suicide rates, even after plausible controls. Yet, national IQs were not a significant predictor of suicide rates when regressed with indices of national religiousness and perceptions of personal health.

National Intelligence, Suicide Rate, and Subjective Well-Being

Across 73 countries around the world, national intelligence estimates predicted significant increments in the variance of male and female suicide rates over and above the predictive contributions of the cognitive and hedonic facets of subjective well-being (life satisfaction, happiness, and unhappiness). This new result adds to current evidence for cognitive ability as an ecologic (aggregate-level) correlate of suicide prevalence.

Regional Intelligence and Suicide Rate in Italy

Replicating the majority finding from prior related studies of European countries, regional IQ estimates and suicide rates in Italy were positively correlated; the association was retained with application of controls for regional affluence. Directions for further inquiry into aggregate-level associations of cognitive-ability indicators and suicide risk within nations are discussed.

Suicide Rates, National Intelligence Estimates, and Differential K Theory

In a nation sample of 75 countries around the world, higher suicide rates of the total, male, and female population corresponded to higher levels on the superordinate K factor from differential K theory, thought to reflect a set of mutually interrelated life history and reproductive strategy traits. Countries ranking high on suicide rates concurrently ranked high on national intelligence estimates, longevity, and affluence, whilst low on rates of births, infant mortality, HIV/AIDS, and crimes (rape, serious assault, and homicide). These findings integrate previously reported positive population-level associations between suicide rates and cognitive ability variables into the conceptual space of differential K theory. The propensity toward suicidal behavior is a positive correlate of the K superfactor.

The relationship between general population suicide rates and educational attainment: a cross-national study

Suicides are associated with both high and low levels of intelligence and educational attainment in both individual-level and aggregate-level studies. A cross-national study examining the relationship between general population suicide rates (y) and educational attainment (x) was undertaken with the a priori hypothesis that the relationship would be curvilinear and follow a U-shaped curve with the quadratic equation Y = A + BX + CX(2), where A, B, and C are constants. Data on suicide rates and the Education Index (a proxy measure of educational attainment) were ascertained from the World Health Organization and the United Nations Web sites, respectively. The main finding was the predicted curvilinear relationship between general population suicide rates, in both sexes, and the Education Index fit the above quadratic equation. Given the cross-sectional study design, a causal relationship cannot be assumed. The impact of educational attainment on general population suicide rates may occur through interaction with other factors, mediation of the effects of other factors, or by its effects being mediated by other factors, and require further study.

Is there a relationship between elderly suicide rates and educational attainment? A cross-national study

BACKGROUND

Suicides are associated with both high and low levels of intelligence and educational attainment in both individual-level and aggregate-level studies, but this has been rarely studied in the elderly.

METHODS

A cross-national study examining the relationship between elderly suicide rates (y-axis) and educational attainment (x-axis) was undertaken with the 'a priori' hypothesis that the relationship would be curvilinear and follow a U-shaped curve with the quadratic equation Y = A + BX + CX(2), where A, B and C are constants. Data on suicide rates for both sexes in the age-bands 65-74 years and 75+ years, and the Education Index (a proxy measure of educational attainment) were ascertained from the World Health Organisation and the United Nations websites, respectively.

RESULTS

The main finding was the predicted curvilinear relationship between suicide rates, in both sexes in both the elderly age-bands, and the Education Index fitting the quadratic equation Y = A + BX + CX(2).

CONCLUSIONS

Given the cross-sectional study design, a causal relationship cannot be assumed. The impact of educational attainment on elderly suicide rates may occur through interaction with other factors, mediation of the effects of other factors, or by its effects being mediated by other factors, and require further study.

Cross-National Social Ecology of Intelligence and Suicide Prevalence: Integration, Refinement, and Update of Studies

This study integrates, refines, and updates previous findings pertaining to positive ecologic (population-level) associations between intelligence and suicide prevalence across nations by using corrected and revised national IQ estimates and, further, a quality-of-human-conditions index, both recently published by Lynn and Vanhanen. Across a global 85-nation sample of sex-specific total suicide rates and a Eurasian 48-nation sample of sex-specific elderly suicide rates, these were positively associated with updated national IQ estimates. The associations were stronger for the general population than for elderly persons, independent of the quality of human conditions, and notedly stronger in exponential fitting of suicide rates with national IQ than in linear fitting, thereby indicating that shifts or differences in national IQ correspond to proportional, not absolute, changes in suicide rates. Implications of these findings and the question of generalizability of such associations to the individual level are discussed.

Genetics of suicide: a systematic review of twin studies

OBJECTIVES

Convergent evidence from a multitude of research designs (adoption, family, genomescan, geographical, immigrant, molecular genetic, surname, and twin studies of suicide) suggests genetic contributions to suicide risk. The present account provides a comprehensive and up-to-date review of the twin studies on this topic.

METHODS

A total of 32 studies (19 case reports, 5 twin register-based studies, 4 population-based epidemiological studies, 4 studies of surviving co-twins) located through extensive literature search strategies are summarized and discussed here. This literature corpus was published between 1812 and 2006 in six languages and reports data from 13 countries.

RESULTS

A meta-analysis of all register-based studies and all case reports aggregated shows that concordance for completed suicide is significantly more frequent among monozygotic than dizygotic twin pairs. The results of co-twin studies rule out exclusively psychosocially based explanations of this pattern. Population-based epidemiological studies demonstrate a significant contribution of additive genetic factors (heritability estimates: 30-55%) to the broader phenotype of suicidal behavior (suicide thoughts, plans and attempts) that largely overlaps for different types of suicidal behavior and is largely independent of the inheritance of psychiatric disorders. Nonshared environmental effects (i.e. personal experiences) also contribute substantially to the risk of suicidal behavior, whereas effects of shared (family) environment do not.

CONCLUSIONS

The totality of evidence from twin studies of suicide strongly suggests genetic contributions to liability for suicidal behavior. To further research progress in this area, an extensive discussion of design limitations, shortcomings of the literature and further points is provided, including sources of bias, gaps in the literature, errors in previous reviews, age and sex effects and twin-singleton differences in suicide risk, and notes from a history-of-science view.

Psychometric properties of the Revised Facts on Suicide Quiz in Austrian medical and psychology undergraduates

Psychometric properties and demographic correlates of a German form of R. W. Hubbard and J. L. McIntosh's (1992) Revised Facts on Suicide Quiz (RFOS), an inventory for assessing overall knowledge about suicide, were investigated in a sample of 1,093 Austrian medical and psychology students. Internal consistency of the RFOS was weak, as were many of the item-total correlations, indicating a fair amount of item-content heterogeneity. Demographic variables (sex, age, field of study, and study level) were significant predictors of RFOS scores and item responses. On the whole, women, advanced, older, and medical students outperformed men, junior, younger, and psychology students. Various directions for further inquiry are discussed

Surname study of suicide in Austria: differences in regional suicide rates correspond to the genetic structure of the population

OBJECTIVES

There is convergent evidence from adoption, family, geographical, immigrant, molecular genetic, twin and, most recently, surname studies of suicide for genetic contributions to suicide risk. Surnames carry information about genetic relatedness or distance and, in patrilineal surname systems, are a close substitute for Y-chromosome markers and haplotypes, since surname transmission is similar to the transmission of the nonrecombining part of the Y chromosome. This study investigated whether differences in regional suicide rates correspond to the genetic structure of the Austrian population.

METHODS

Differences in district-level standardized suicide rates 1988-94 between the five major surname regions identified for Austria were analyzed. The surname regions used in the analysis reflect the contemporary population structure and closely follow the natural borders found in the topography of Austria, less so its administrative division into nine states.

RESULTS

Surname region accounted for a significant (P < 0.001) and substantial (38%) portion of the variance in district-level suicide rates. Adjusting the suicide rates for a set of five social and economic indicators that are established ecological correlates of suicide prevalence (income, and rates of the divorced, unemployed, elderly and Roman Catholics) left the results essentially unchanged.

CONCLUSIONS

Regional differences in suicide rates within Austria correspond to the genetic structure of the population. The present evidence adds to related findings from geographical and surname studies of suicide that suggest a role for genetic risk factors for suicidal behavior. Genetic differences between subpopulations may partially account for the geography of suicide. Study limitations and directions for future research are discussed.

Facts and myths about seasonal variation in suicide

The prevalence of suicide presents a universal seasonal pattern. In the Northern hemisphere, suicides peak during spring and early summer and the trough occurs during winter. This peculiar pattern might be counterintuitive for everyday reasoning. Data from 1,093 medical and psychology undergraduates from Austria (382 men and 711 women; M age 25.0 yr., SD=6.6) indicated an almost perfectly reversed pattern of beliefs about suicide seasonality compared with the actual seasonal distribution. The vast majority of respondents believed the peak to be located in late autumn and early winter and the trough occurring in late spring and the summer months. Implications for education and practice are discussed.

Social ecology of intelligence and suicide in the United States

Two proxy variables for regional intelligence, state-level mean scores of 1994-2004 on the American College Test and of 1982, 1993, 1994, and 1998-2004 on the Scholastic Assessment Test, were consistently positively correlated with the averaged regional U.S. suicide rates of 1990-1994. The effects observed in these ecological (group-level) analyses were of small size, with correlation coefficients ranging from .05 to .25, thus mostly falling below the nominal statistical significance level. The present evidence nevertheless adds to several recent findings from international geographical studies by Lester and by Voracek of a positive association of intelligence and suicide mortality by indicating that this pattern may also exist on the within-nation level.