Back to school blues: Seasonality of youth suicide and the academic calendar

https://doi.org/10.1016/j.econedurev.2011.04.012Get rights and content

Abstract

Previous research has found evidence of academic benefits to longer school years. This paper investigates one of the many potential costs of increased school year length, documenting a dramatic decrease in youth suicide in months when school is not in session. A detailed analysis does not find that other potential explanations such as economic conditions, weather or seasonal affective disorder patterns can explain the decrease. This evidence suggests that youth may face increased stress and decreased mental health when school is in session.

Highlights

► We document a large decrease in youth suicide in during summer. ► Adults from a slightly older age ranges exhibit no summer decrease in suicide. ► The summer decline in youth suicide is not explained by weather, unemployment, or SAD. ► The increase rate of youth suicide during non-summer months aligns with school calendar. ► That increase may be indicative of broader stress experienced by youth in school.

Introduction

Suicide is the third leading cause of death among youth and the suicide rate of 15–19 year olds tripled between 1950 and 1990. Over the same time period, suicide rates for older individuals decreased. A significant literature exists exploring the determinants of youth suicide, with recent studies showing that risky behaviors such as alcohol consumption and sexual activity are associated with youth suicide (Carpenter, 2004, Sabia, 2008). Despite this literature, no previous work (to our knowledge) has directly explored the relationship between school and youth suicide, even though teens are in school for over three-quarters of the calendar year. This paper examines the seasonality of youth suicide with a specific focus on how youth suicide may be related to the typical academic calendar.

Using a panel of state suicide rates between 1980 and 2004, our data show during months that students tend to be on break from school (June, July, August and December), youth suicide is significantly lower than the rest of the year. This pattern does not hold for adults. One possible explanation for the summer decline in youth suicide is the prevalence of seasonal affective disorder (SAD), where individuals become more depressed during the winter months. SAD is most prominent in northern states where the sunlight is limited during the winter, but relatively abundant during the summer. Results show that suicides also decrease during the month of December, when students are typically on winter break, and further analysis shows that the suicide decrease in the summer months is not driven by northern states. Furthermore, SAD affects youth females significantly more than youth males, but males are driving the results below. In addition to investigating the role that SAD plays, economic conditions, weather patterns and divorce rates are also considered and when controlling for these potential factors, the seasonal youth suicide pattern remains.

Recent work shows the benefits of increased schooling,1 but rarely is the negative impact of schooling considered. Our results suggest that the increased stress that students face throughout the school year may exacerbate mental health issues and increase youth suicide. This potential cost of education should be taken into consideration as school districts debate increasing the length of the academic school year.

Section snippets

Background

The seminal work on suicide in economics by Hamermesh and Soss (1974) develops a rational-suicide theory that argues that suicide rates should increase as individuals get older. While their data from the early half of the 1900s support their predictions that older age groups should have higher suicide rates, recent trends in suicide across age groups tend to deviate from their findings. Most of the research on youth suicide looks at the patterns of youth suicide across ages, time periods,

Data

The data used in the paper comes from a variety of sources. The mortality data is from the Multiple Cause-of-Death Public Use Files, which are published annually by the National Center for Health Statistics. Between 1977 and 1999, the International Classification of Diseases, 9th Edition (ICD-9) was used to code mortality and currently the ICD-10 is used. Suicides are defined in the ICD-9 using code 350 in the “34 Recode” classification and code 040 in ICD-10 using the “39 Recode”

Results

This section estimates the size of the decrease in suicides for youth in the summer and explores whether the drop can be explained by observable characteristics. The seasonal pattern in youth suicides is compared with adult suicide patterns and the stability of the summer decrease is analyzed. In order to determine the role that SAD plays, the summer decrease is analyzed by region and by gender. The summer pattern of other common causes-of-death for the youth are analyzed in order to rule out

Conclusion

Recent high profile criminal cases in Massachusetts have anecdotally demonstrated the increased stress and decreased mental health that students can face as a result of being in school. Jacob and Lefgren (2003) begin to address some of these issues, finding a greater prevalence of violence when school is in session, which they attribute to increased negative social interactions. If negative social interactions are more likely when school is in session, then summer break could lead to a time

References (31)

  • Evans, W. N., & Moore, T. J. Liquidity, activity, mortality. Review of Economics and Statistics, in...
  • D.G. Freeman

    Determinants of youth suicide: The Easterlin-Holinger cohort hypothesis re-examined

    American Journal of Economics and Sociology

    (1998)
  • L. Friedberg

    Did unilateral divorce raise divorce rates? Evidence from panel data

    American Economics Review

    (1998)
  • D. Hamermesh et al.

    An economic theory of suicide

    Journal of Political Economy

    (1974)
  • B. Hansen

    School year length and student performance: Quasi-experimental evidence

    (2008)
  • Cited by (0)

    We are grateful to helpful comments from Dave Marcotte, Lars Lefgren, Ben Cowan, Marianne Page, Hilary Hoynes, Scott Carroll, Daniel Rees, Jason Lindo and Glen Waddell. We also thank participants at the 2010 WEAI and 2011 SOLE annual meetings and the UC Davis public economics and University of Oregon micro-economics seminars. Lastly, we are grateful to comments from two referees and the editor which improved the quality of the paper.

    View full text