EconC175_Modeling+Effects+of+Economic+Burden+on+Suicide+Rates

EconC175_Modeling+Effects+of+Economic+Burden+on+Suicide+Rates

Info iconThis preview shows pages 1–23. Sign up to view the full content.

View Full Document Right Arrow Icon
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Background image of page 2
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Background image of page 4
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Background image of page 6
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Background image of page 8
Background image of page 9

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Background image of page 10
Background image of page 11

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Background image of page 12
Background image of page 13

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Background image of page 14
Background image of page 15

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Background image of page 16
Background image of page 17

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Background image of page 18
Background image of page 19

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Background image of page 20
Background image of page 21

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full Document Right Arrow Icon
Background image of page 22
Background image of page 23
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Ame WWI-3v iii-TH}: mi); how “mm” Q0 W W L. "1‘ FIRM WWW" DwsPT A“ Modeling the Effects of Economic Burden on Suicide Rates: A Multi-Variable Regression Ned Kaida-Yip Prof. Ronald Lee I. Introduction The old-age dependency ratio has long been considered a vital statistic to observe in order to predict the burden that the retired population is likely to place upon the working population. As overall population ages, greater burdens are placed upon the working class via pensions, social security, and more direct care-taking efforts. Policy makers have rightly considered rising old-age dependency to be issues of economic import. However, the impact on the mental well-being of a population is also worth considering. The most extreme cases of mental duress manifest as suicide, the taking of ones own life. Understanding whether or not economic burden, represented by the old~age dependency ratio, contributes to the suicide rate coule be an important step towards suicide prevention. The hypothesis is that there exists_ a. positive correlation between the old-age dependency u) ex, ratio and the suicide rate. This wiggng im 1 that the burdens of an agingpo‘puulamtion L7 Hm—m deteriorate mental health and economic welljaeing, The last available set of data on suicide by state was compiled by the Center for Disease Control in 2008, the data from 2005. The data for old-age dependency was taken from the United States Census Bureau’s American Community 2005. A multiple regression analysis indicates that a 1% increase in the old-age dependency ratio produces a 093% increase in the suicide rate. In this multiple regression /' n. analysis a number of other variables were included: Child dependencyfiratig, sex ratio, employment percentage, high school completion, and White. These other variables were included to attemfltgcpntrol omitted yariable bias whichwouldrendertheicoeflficients i @ inaccurate. II) The Hypothesis and Data -‘ Old-age dependency ratios across states are used to approximate the economic burden upon the population. These ratios were taken from the United States Census Bureau’s American Community Survey 2005. This continuing survey is distributed via form to a random sampling of addresses that also received the 2005 Census form. This provides 51 observations across all states and the District of Columbia. The old-age dependency ratio is generated by dividing the population aged 65+ by the population aged 18-64 and then multiplying by 100. Table 1 states the mean old-age dependency ratio is 19.45 with a standard deviation of 2.73. The other variables were also derived from the same American Community Survey. Child dependency is calculated the same as old-age dependency only with minors aged under 18. The sex ratio is the number of males to every 100 females. Gender is included because according to the 2005 data, men tend to commit suicide with greater frequency than females. Employment percentage is the number of individuals employed between the ages 18-64. High school completion is the percentage of individuals over 25 who have either graduated high school or achieved an equivalent degree. These two variables are included in order to control for differences in education and employment. They approximate education level and employment across states. Finally, there is a variable for the percentage of whites in the state because according to the 2005 Suicide Data white males were the most prevalent committers of suicide. a The suicide rates are taken from the American Association of Suicidology’s final 2005 1 official data report. The rate is calculated as the number of suicides in a state divided by the population of the state and then multiplied by 100,000; the number of suicides per 100,000 a, J citizens. The first regression run is simply that the suicide rate is a function of the old-age dependency ratio. The other variables each could affect the suicide rate, so they are added in to help control for omitted variable bias. The last regression run includes all the variables. The Q?) regressions run will chart a linear relationship between the suicide rate and the included explanatory variables. Old-age dependency increases as the population ages. This places a greater burden upon the population via social services, pensions, and personal caretaking. Thus the hypothesis is that the additional economic burden would take a toll on the mental health of the population and thus a positive correlation would exist between the old-age dependency ratio and the suicide rate. The sex ratio should have a positive correlation as males commit suicide more often and increase in males per 100 females should increase suicide. White should also be positively correlated as white males have the highest percentage of suicides. However, employment and high school completion should be negatively correlated. As education and employment increase suicide should decrease because a more educated and employed populace probably has a higher level of mental health. III) The Multiple-Regression Equation The multiple regression equation appears as: m. Yn = B1 ann + B2ann + B3nXsn + B4nXen + B5nth + B6ann Where Y is the suicide rate and 11 takes a value between 1 and 51. The multiple regression model will use the ordinary least squares method to approximate the best fit line. This model assumes a linear relationship between the Y, suicide rate, and the explanatory variables. As variables are added each one can be held constant in order to better isolate the effect of old-age dependency upon suicide rate. IV) The Regressions In the appendix, there are the results of the regressions as each additional variable was added. Along with the t-statistic and the p-value, I have also provided a hypothesis test. The null hypothesis in each case is that the coefficient of old-age dependency is 0, or that the ratio has no effect on suicide rates and is thus irrelevant. The hypothesis is tested at a 95% confidence interval and the t-stat and p-value are given for a 5% significance level. The initial regression found in Regression 1 is only suicide rate regressed for old age. Initially, the coefficient is -. 197 which goes against the hypothesis that the coefficient for old-age dependency would be positive. However, the result is not statistically significant at a 5% level and so it is likely that old-age coefficient is simply 0. Regression 2 adds the explanatory variable child dependency. The coefficient remains negative but it is slightly closer to zero at -. l 87. However, it remains statistically insignificant at the 5% level. With the addition of the gender ratio in Regression 3, the coefficient takes a positive value of .124. However, the result still remains statistically insignificant. However, as hypothesized, gender ratio has a positive correlation of 1.01 and is statistically significant. So as of regression 3, gender ratio is significant to suicide rate. Therefore through the first 3 regressions the results have yet to be significant at any meaningful level. The likelihood that old-age dependency’s coefficient is zero is high. Regression 4 adds employment as a variable. As expected the coefficient is negative, indicating that as employment increases by 1% suicide rates fall by .314%. This coefficient is significant at a 5% level. Old-age dependency remains statistically insignificant with a small positive coefficient. Gender remains significant and positive. Regression 5 adds percentage of high school completed citizens and it shows a small negative coefficient. It is statistically insignificant. In Regression 5 only gender remains significant. In Regression 6 we add the final variable: White. At this point we see that gender and employment are significant but none of the other variables carry any weight at the 5% level. Old-age dependency has a coefficient of 0.093 and remains statistically insignificant. V) Conclusion For all regressions, save 1 and 2, old-age dependency carried a small, positive coefficient. However, it was never statistically significant at a 5% level for any of the regressions. Meaning in all regressions there is a 95% probability that old-age dependency had no impact on suicide rate and thus, coefficient zero. Thus the proposed hypothesis was not proven. An aging population and increasing old-age dependency ratio also indicates falling mortality which acts against suicide rates. Furthermore, people may be living longer due to greater satisfaction and mental health. Finally, the significant burdens of old-age dependency are often taken from the working age in taxes and not as a direct burden that may break them down mentally. These factors all help explain why old-age dependency may not have been the best possible explanatory variable. In the process of testing, however, a few superior explanatory variables were revealed. Namely, gender and employment. These variables were almost always significant once they were introduced into the model. As males increase, so does the rate of suicide. Conversely, increases in employment leads to decreases in the rate of suicide. Unemployment rates may have been a better choice of variable to represent economic burden upon workers 18-64 than old-age dependency ratios. Finally, there is a_b_it of a missin link in this examination in that on] 1 ear 2005) was tested. Suicide rates vary greatly over the years, so drawing additional samples from the 51 states W over time and doing a time sample would help to better reveal which variables have significant _' act upon suicide rates. While the lack of significance in any regression is compelling evidence that old-age dependency ratios have no effect on suicide rates, a study with observations over many years may provide different results. Appendix Table 1: Variable Obs Mean Std. Dev. Min Max Child ratio 51 39.36667 3.326299 32.7 49.8 Old ratio 51 19.45294 2.731033 10.2 27.6 Sex ratio 51 96.31372 2.65413 89 103.1 Employment 51 70.92941 3.823417 61.7 78.2 High school 51 85.64902 3.610394 78.5 91.3 Suicide rate 51 12.432 3.593443 6 22 White 51 73.28431 16.40613 23 96 Regression l: Suicide rate| Coef. Std. Err. t P>|t| [95% Conf. Interval] ........... --+-------------------------------------..-------..-_--.._.._---------- Oldratiol -.1965_3_97 .1863333 -1.05 0.297 -.5711879 .1781085 _cons| 16.26059 3.665094 4.44 0.000 8.891428 23.62976 H0: Old ratio coefficient = 0 F-Stat: 1.11 P-Value:0.297 Regression 2: Suicide ratel Coef. Std. Err. t P>|t| [95% Conf. Interval] ........... -..+--------..---_-_-_------------------------------------....--..---..-- Oldratiol -.1872106 .2178107 -0.86 0.394 -.6253892 .2509679 Childratiol .0152632 .1791368 0.09 0.932 -.3451135 .37564 _cons| 15.47865 9.896388 1.56 0.125 -4.430312 35.38762 H0: Old ratio coefficient =0 F-Stat = 0.74 P-Value = 0.394 Regression 3: Suicide ratel Coef. Std. Err. t P>|t| [95% Conf. Interval] ........... --+__-------------------_-----_--___---_-__-_-____-__--_----------- Old ratio] .1240758 .1695459 0.73 0.468 -.2l72025 .465354 Child ratiol -.l34092 .1354 -0.99 0.327 -.4066381 .1384542 Sexratiol 1.014983 .162689 6.24 0.000 .6875071 1.342459 _cons| -82.53197 17.34945 -4.76 0.000 417.4546 -47.60933 H0: Old Ratio Coefficient = 0 F-stat = 0.54 P-Value = 0.468 Regression 4: Suicide ratel Coef. Std. Err. t P>|t| [95% Conf. Interval] ........... --+-----------------_------...--..-....-----------....------------------- Oldratiol .1461271 .1_603_742 0.91 0.367 -.1770858 .46934 Childratiol -2521331 .1378453 -l.83 0.074 -.529942 .0256758 Sexratiol 1.222332 .1781232 6.86 0.000 .8633481 1.581316 Employment | -.2981318 .1560177 -1.91 0.063 -.6125648 .0163011 Highschooll -.0263395 .1776728 .015 0.883 -.3844154 .3317365 _cons| -74.85802 16.71975 —4.48 0.000 -108.5545 -41.16158 H0: Old Ratio Coefficient = 0 F-Stat = 0.83 P-Value = 0.367 Regression 5 Suicide rate | Coef. Std. Err. t P>|t| [95% Conf. Interval] ........... .-+..--------_-_-----------..-..-..-.....-------------------------------.. Oldratiol .093301 .1591544 0.59 0.561 ~.2276645 .4142665 Child ratiol 4719—32389 TI3W7—18 -1.40 0.170 -.4722922 .0858144 Sexratiol 1.140948 .1795185 6.36 0.000 .7789142 1.502981 Employmentl -.3396205 .1539134 -2.21 0.033 -.6500166 -.0292245 Highschooll -.0533829 .1739526 -0.31 0.760 -.4041917 .2974259 White] .0468161 .0259833 1.80 0.079 -.0055841 .0992164 _cons| -66.47563 16.95921 -3.92 0.000 -100.6771 -32.27413 nun-"uuun—-———_-—- ------------------------------------------------------- n- Works Cited Ashenfelter, Orley, Phillip B. Levine, and David J. Zimmerman. Statistics and Econometrics: Methods and Applications. New York: J. Wiley, 2003. Print. H.-S. Kung, D. L. Hoyert, J. Xu, & S. L. Murphy. (2008, January). Deaths: Final Data for 2005. National Vital Statistics Reports, 56(10). http2//www.cdc.gov/nchs/data/nvsr/nver6/nvsr56_10.pdf (Retrieved April 9, 2010). United States -- States; and Puerto Rico GCT0105. Age Dependency Ratio of the Total Population: 2005. United States Census Bureau. American Community Survey 2005. http://factfinder.census.gov/servlet/GCTTable? bm=y&-geo id=01000US&- ds name=ACS 2005 EST G00 &- lan =en&-redoLo =false&-format=US-9&- mt name=ACS 2005 EST G00 ocrmos US9&-CONTEXT=gct (Retrieved April 9, 2010) United States -- States; and Puerto Rico GCT0106. Child Dependency Ratio of the Total Population: 2005. United States Census Bureau. American Community Survey 2005. http://factfinder.census.gov/servlet/GCTTable? bm=y&-geo id=01000US&- ds name=ACS 2005 EST GOO &- 1ang:—en&-redoLog=false&-format=US-9&- mt name=ACS 2005 EST G00 GCT0106 US9&-CONTEXT=gct (Retrieved April 9, 2010) United States -- States; and Puerto Rico GC T23 03. Employment/Population Ratio for the Population 16 to 64 Years Old: 2005. United States Census Bureau. American Community Survey 2005. United States Census Bureau. American Community Survey 2005. http://factfinder.census.gov/servlet/GCTTable? bm=y&-geo id=01000US&- ds name=ACS 2005 EST GOO &- lan =en&-redoLo =false&-format=US-9&- mt name=ACS 2005 EST GOO GCT2303 US9&-CONTEXT=gct (Retrieved April 9, 2010) United States -- States; and Puerto Rico GCT1401. Percent of People 25 Years and Over Who Have Completed High School (Includes Equivalency): 2005. United States Census Bureau. American Community Survey 2005. http://factfinder.census.gov/servlet/GCTTable? bm=y&-geo id=01000US&- ds name=ACS 2005 EST GOO &- lang:—en&-redoLog=false&-format=US—9&- mt name=ACS 2005 EST GOO GCT1401 US9&-CONTEXT=gct (Retrieved April 9, 2010) United States -- States; and Puerto Rico GCT0201. Percent of the Total Population Who Are White Alone: 2005 United States Census Bureau. American Community Survey 2005. http://factfinder.census.gov/servlet/GCTTable? bm=y&-geo id=01000US&- ds name=ACS 2005 EST G00 &- lang:—en&-redoLog=false&—format=US-9&- mt name=ACS 2005 EST G00 GCT0201 US9&-CONTEXT=gct (Retrieved April 9, 2010) United States -- States; and Puerto Rico GCT0102. Sex Ratio of the Total Population: 2005. United States Census Bureau. American Community Survey 2005. http://factfinder.census.gov/scrvlet/GCTTable? bm—:y&-geo id=OlOOOUS&- ds name=ACS 2005 EST G00 &- 1ang=en&-redoLog=false&-format=US-9&- mt name=ACS 2005 EST G00 GCT0102 US9&-CONTEXT=gct (Retrieved April 9, 2010) 10 Works Cited Ashenfelter, Orley, Phillip B. Levine, and David J. Zimmerman. Statistics and Econometrics: Methods and Applications. New York: J. Wiley, 2003. Print. H.-S. Kung, D. L. Hoyert, J. Xu, & S. L. Murphy. (2008, January). Deaths: Final Data for 2005. National Vital Statistics Reports, 56(10). http://www.cdc.gov/nchs/data/nvsr/nver6/nvsr56_10.pdf (Retrieved April 9, 2010). United States -- States; and Puerto Rico GCT0105. Age Dependency Ratio of the Total Population: 2005. United States Census Bureau. American Community Survey 2005. hfipM/factfindencensus.gov/servlet/GCTTable? bm=y&-geo id=01000US&- ds name=ACS 2005 EST GOO &- lan =en&-redoLo =false&-fonnat=US-9&- mt name=ACS 2005 EST GOO GCTOIOS US9&-CONTEXT=gct (Retrieved April 9, 2010) United States -- States; and Puerto Rico GCT0106. Child Dependency Ratio of the Total Population: 2005. United States Census Bureau. American Community Survey 2005. hflpd/factfindencensus.gov/servlet/GCTTable? bm=y&-geo id=01000US&- ds name=ACS 2005 EST G00 &- Iang=en&-redoLog=false&-format=USo9&- mt name=ACS 2005 EST GOO GCT0106 US9&-CONTEXT=gct (Retrieved April 9, 2010) United States -- States; and Puerto Rico GCT23 03. Employment/Population Ratio for the Population 16 to 64 Years Old: 2005. United States Census Bureau. American Community Survey 2005. United States Census Bureau. American Community Survey 2005. http://factfinder.census.gov/servlet/GCTTable? bm=y&-geo id=01000US&- ds name=ACS 2005 EST G00 &- lang=en&-redoLog=false&-format=US-9&- mt name=ACS 2005 EST G00 GCT2303 US9&-CONTEXT=gct (Retrieved April 9, 2010) United States ~- States; and Puerto Rico GCTI 401. Percent of People 25 Years and Over Who Have Completed High School (Includes Equivalency): 2005. United States Census Bureau. American Community Survey 2005. http://factfinder.census.gov/servlet/GCTTable? bm=y&-geo id=01000US&- ds name=ACS 2005 EST GOO &- [ang=en&-redoLog=false&-format=US-9&- mt name=ACS 2005 EST GOO GCTI401 US9&-CONTEXT=gct (Retrieved April 9, 2010) United States -- States; and Puerto Rico GCT0201. Percent of the Total Population Who Are White Alone: 2005 United States Census Bureau. American Community Survey 2005. http://factfinder.census. gov/servlet/GCTTable? bm=y&-geo id=01 000US&- ds name=ACS 2005 EST G00 &- lang=en&-redoLog=false&-fonnat=US-9&- mt name=ACS 2005 EST GOO GCT0201 US9&-CONTEXT=gct (Retrieved April 9, 2010) 11 United States -- States; and Puerto Rico GCT0102. Sex Ratio of the Total Population: 2005. United States Census Bureau. American Community Survey 2005. http://factfinder.census.gov/servlet/GCTTable? bm—fi&-geo id=01000US&- ds name=ACS 2005 EST G00 &- lan =en&-redoLo =false&-fonnat=US-9&- mt name=ACS 2005 EST G00 GCT0102 US9&—CONTEXT=gct (Retrieved April 9, 2010) Y 7 Ned «668399 A/ gm P— ‘— “% Q‘Hjflmv \lexa wen Wbpqv) ma;qu mama)» «mam»: (bo’r 995 WW We ’YO\OO\« ax- «he weerde 69(398 ‘ @®1+h<m?m @113 my \Ico deed Om‘PM \IW b53611 @uem gem W985“ OD dfififiemfi 95 appswqa‘g gm “GENE” *0 W ‘%%‘W 9%an a swan 039d Q‘Wvddkd. @VQ‘V W am (waved dPfimsst 952 3M9 emwfiw, @Vm good!) I am wmgsséad amcm we“ \IOU 096%83 *‘(Q \‘PmQerQmeJ‘ (345 I Hemalmpmvemms @5596 "We 99> a poefl‘A/e Cowaéfibfl mum we dd age dependency (($90 and fine mfifide (ade‘: 01%ij 96 five «Qmm \IOU anéNC‘bpfi. HMVQr,® fig“ Sher \loo ©®Qd 2 “Wes (mud 9mm! max me burden?) 096m @9919 pOpUkBY‘RSM Worse mema\ beam and QCOflomPc we» befixg‘; You camo’v 6&9 causafiw even *muagn you have We ngecr comrdgo w F9f€>*,\1008396 There 9% carexabipan and mg) were ‘aflenfioo ware «(5me am» causamvl @wfl’r 6mm Pmfigwds or (QQQW ©93c???c «ends? Ex 903m, 9* cooxcs be -+ha+ ammo 059% mg, a mgh dd ~deperrjawcxf @Wo ,. have mm Mg W W06? @692 may do my Mwacv m @091 Qqu QBXWCQPGMQ)‘ N63, (13st *Vxemefi wen (X2098 9mm Mae 666m \eave Mme Pom W?» 92> \eQ’s 964m \6398 Qmmake. 1H6 my mmame W 62059 t W ENS“ 66‘9‘9 Q5 “3‘ \‘00 Cfinndr esxabWsh @3689“ 306* Rem mam fixaQO @994» @6193 new) M98ng mex and \mxgx wee».uh%% we cmgamrm a a 1M): eggfic‘v W) *0 $3?er 5051- Am 6001858 Ya @EQDW \KJo name Wfle @9869; @5th km?» ,wbo‘ ‘ MW} (399999 Jr‘fiafim \geU‘D" 63%??? «if (3% (30*; @KP @‘QQK‘JVD 0%; 90$ (3*) , . FLNAI/ DRASPT ’r ' . \100 more??? %ij% b43635 we *0 5% an \[OQ needed .qu om W (@965; Modeling the Effects of Economic Burden on Suicide Rates: , A Multi-Variable Regression Ned Kaida-Yin Prof._Ronald Lee-s I. Introduction The old-age dependency ratio has long been considered a vital statistic to observe in order to predict the burden that the retired population is likely to place upon the working population. As overall population ages, greater burdens are placed upon the working class via pensions, social security, and more direct care-taking efforts. Policy makers have rightly considered rising old-age dependency to be issues of economic import. However, the impact on the mental well-being of a population is also worth considering. The most extreme cases of mental duress manifest as suicide, the taking of ones own life. Understanding whether or not economic burden, represented by the old-age dependency ratio, contributes to the suicide rate could be an important step towards suicide prevention. The hypothesis is that there exists a positive correlation between the old-age dependency ratio and the suicide rate. The last available set of data on suicide by state was compiled by the Center for Disease Control in 2008, the data from 2005. The data for old-age dependency was taken from the United States Census Bureau’s American Community 2005. A multiple regression analysis indicates that a 1% increase in the old-age dependency ratio produces a 093% increase in the suicide rate. In this multiple regression analysis a number of other variables were included: Child dependency ratio, sex ratio, employment percentage, high school completion, and White. These other variables were included to attempt to control omitted variable bias which would render the coefficients inaccurate. II) The Hypothesis and Data Old-age dependency ratios across states are used to approximate the economic burden upon the population. These ratios were taken from the United States Census Bureau’s American Community Survey 2005. This continuing survey is distributed via form to a random sampling of addresses that also received the 2005 Census form. This provides 51 observations across all states and the District of Columbia. The old-age dependency ratio is generated by dividing the population aged 65+ by the population aged 18-64 and then multiplying by 100. Table 1 states the mean old-age dependency ratio is 19.45 with a standard deviation of 2.73. The other variables were also derived from the same American Community Survey. Child dependency is calculated the same as old-age dependency only with minors aged under 18. The sex ratio is the number of males to every 100 females. Gender is included because according to the 2005 data, men tend to commit suicide with greater frequency than females. Employment percentage is the number of individuals employed between the ages 18-64. High school completion is the percentage of individuals over 25 who have either graduated high school or achieved an equivalent degree. These two variables are included in order to control for differences in education and employment. They approximate education level and employment across states. Finally, there is a variable for the percentage of whites in the state because . according to the 2005 Suicide Data white males were the most prevalent committers of suicide. The suicide rates are taken from the American Association of Suicidology’s final 2005 official data report. The rate is calculated as the number of suicides in a state divided by the population of the state and then multiplied by 100,000; the number of suicides per 100,000 citizens. The first regression run is simply that the suicide rate is a fiinction of the old-age dependency ratio. The other variables each could affect the suicide rate, so they are added in to help control for omitted variable bias. The last regression run includes all the variables. The regressions run will chart a linear relationship between the suicide rate and the included explanatory variables. With the control variables in place to help correct for omitted variable bias, the regression becomes more precise. However, there are other factors that are much more difficult to correct for. Firstly, there are region s e ' . tes simply have a higher population than other states, which intuitively means that these states are likely to have a higher number of M suicides. However the problem of state population is further compounded because higher J population states tend to have higher wages as well. These two issues are related to the old—age dependency ratio in that states with a lower ratio may be younger because they attract young labor force members. Intuitively, it makes sense that states with higher wages and population tend to attract younger workers and this increases the population and decreases the old-age dependency ratio. However, as population rises the number of suicides may increase due to the raw number of people in the state. Also, workers leaving for higher wage states leave may leave a less capable pool of workers in a lower wage state, as well as increasing the old-age dependency ratio. These region specific trends may well influence suicide rates and I was unable to find a way to correct for these issues. A way help correct for these region specific trends is to M. examine them over time. As time trends over, it would be possible to examine a state suicide rate, W old-age dependency ratio 0 ulation and state er—capita wage over time. This time trend could help correct for the region-specific issues that prevent causality from being established. M*‘ wva an So although we can look for correlation between old-age dependency ratios and suicide rates it is not possible to establish causality via a state-to-state comparison. Old-age dependency increases as the population ages. This places a greater burden upon the population via social services, pensions, and personal caretaking. Thus the hypothesis is that the additional economic burden would take a toll on the mental health of the population and thus a positive correlation would exist between the old-age dependency ratio and the suicide rate. The sex ratio should have a positive correlation as males commit suicide more often and increase in males per 100 females should increase suicide. White should also be positively correlated as white males have the highest percentage of suicides. However, employment and high school completion should be negatively correlated. As education and employment increase suicide should decrease because a more educated and employed populace probably has a higher level of mental health. 111) The Multiple-Regression Equation The multiple regression equation appears as: Yn = B] nXon + BZann + B3nXsn + B4nXen + B5nth + B6ann Where Y is the suicide rate and n takes a value between 1 and 51. The multiple regression model will use the ordinary least squares method to approximate the best fit line. This model assumes a linear relationship between the Y, suicide rate, and the explanatory variables. As variables are added each one can be held constant in order to better isolate the effect of old-age dependency upon suicide rate. The variables represent the following: Xon = old age ratio, ch = child ratio, Xsn = sex ratio, Xen = employment ratio, th = high school graduation rate for 18-25 year olds, Xwn = White ratio of the population. IV) The Regressions In the appendix, there are the results of the regressions as each additional variable was added. Along with the t-statistic and the p-value, I have also provided a hypothesis test. The null hypothesis in each case is that the coefficient of old-age dependency is 0, or that the ratio has no effect on suicide rates and is thus irrelevant. The hypothesis is tested at a 95% confidence interval and the t-stat and p-value are given for a 5% significance level. The initial regression found in Regression l is only suicide rate regressed for old age. Initially, the coefficient is -.197 which goes against the hypothesis that the coefficient for old-age dependency would be positive. However, the result is not statistically significant at a 5% level and so it is likely that old-age coefficient is simply 0. Regression 2 adds the explanatory variable child dependency. The coefficient remains negative but it is slightly closer to zero at -.187. However, it remains statistically insignificant at the 5% level. With the addition of the gender ratio in Regression 3, the coefficient takes a positive value of .124. However, the result still remains statistically insignificant. However, as hypothesized, gender ratio has a positive correlation of 1.01 and is statistically significant. So as of regression 3, gender ratio is significant to suicide rate. Therefore through the first 3 regressions the results have yet to be significant at any meaningful level. The likelihood that old-age dependency’s coefficient is zero is high. Regression 4 adds employment as a variable. As expected the coefficient is negative, indicating that as employment increases by 1% suicide rates fall by .314%. This coefficient is significant at a 5% level. Old-age dependency remains statistically insignificant with a small positive coefficient. Gender remains significant and positive. Regression 5 adds percentage of high school completed citizens and it shows a small negative coefficient. It is statistically insignificant. In Regression 5 only gender remains significant. In Regression 6 we add the final variable: White. At this point we see that gender and employment are significant but none of the other variables carry any weight at the 5% level. Old-age dependency has a coefficient of 0.093 and remains statistically insignificant. V) Conclusion For all regressions, save 1 and 2, old-age dependency carried a small, positive coefficient. However, it was never statistically significant at a 5% level for any of the regressions. Meaning in all regressions there is a 95% probability that old-age dependency had no impact on suicide rate and thus, coefficient zero. Thus the proposed hypothesis was not proven. An aging population and increasing old-age dependency ratio also indicates falling mortality which acts against suicide rates. Furthermore, people may be living longer due to greater satisfaction and mental health. Finally, the significant burdens of old-age dependency are often taken from the working age in taxes and not as a direct burden that may break them down mentally. These factors all help explain why old-age dependency may not have been the best possible explanatory variable. In the process of testing, however, a few superior explanatory variables were revealed. Namely, gender and employment. These variables were almost always significant once they were introduced into the model. As males increase, so does the rate of suicide. Conversely, increases in employment leads to decreases in the rate of suicide. Unemployment rates may have been a better choice of variable to represent economic burden upon workers 18-64 than old-age dependency ratios. Finally, there is a bit of a missing link in this examination in that only 1 year (2005) was tested. Suicide rates vary greatly over the years, so drawing additional samples from the 51 states over time and doing a time sample would help to better reveal which variables have significant impact upon suicide rates. While the lack of significance in any regression is compelling evidence that old-age dependency ratios have no effect on suicide rates, a study with observations over many years may provide different results. Appendix Table 1: Variable Obs Mean Std. Dev. Min Max Child ratio 51 39.36667 3.326299 32.7 49.8 Old ratio 51 19.45294 2.731033 10.2 27.6 Sex ratio 51 96.31372 2.65413 89 103.1 Employment 51 70.92941 3.823417 61.7 78.2 High school 51 85.64902 3.610394 78.5 91.3 Suicide rate 51 12.432 3.593443 6 22 White 51 73.28431 16.40613 23 96 Regression 1: Suicide rate| Coef. Std. Err. t P>|t| [95% Conf. Interval] ........... --+------------------------------------------_--------------_--_--- Old ratiol -.1965397 .1863333 -1.05 0.297 -.5711879 .1781085 _cons| 16.26059 3.665094 4.44 0.000 8.891428 23.62976 H0: Old ratio coefficient = 0 F-Stat: 1.11 P-Value:0.297 Regression 2: Suicide rate| Coef. Std. Err. t P>|t| [95% Conf. Interval] ........... --+---------------------------------------------------------------- Old ratio | -.1872106 .2178107 -0.86 0.394 -.6253 892 .2509679 Child ratio 1 .0152632 .1791368 0.09 0.932 -.3451135 .37564 _cons| 15.47865 9.896388 1.56 0.125 -4.430312 35.38762 HO: Old ratio coefficient =0 F-Stat = 0.74 P-Value = 0.394 Regression 3: ‘ Suicide rate | Coef. Std. Err. t P>|t| [95% Conf. Interval] ........... --+------....---....--------------..--------..------..-------------------- Oldratiol .1240758 .1695459 0.73 0.468 -.2172025 .465354 Child ratiol -.l34092 .1354 -0.99 0.327 -.4066381 .1384542 Sexratiol 1.014983 .162689 6.24 0.000 .6875071 1.342459 _cons| -82.53197 17.34945 -4.76 0.000 -117.4546 -47.60933 H0: Old Ratio Coefficient = 0 F-stat = 0.54 P-Value = 0.468 Regression 4: Suicide rate| Coef. Std. Err. t P>|t| [95% Conf. Interval] ........... -_+_----_-----_-..---_-_-_----..__------_-------__-..---_.......--..-....-.._- Old ratiol .1461271 .1603742 0.91 0.367 -.l770858 .46934 Childratiol -.2521331 .1378453 -1.83 0.074 -.529942 .0256758 Sexratiol 1.222332 .1781232 6.86 0.000 .8633481 1.581316 Bmploymentl -.2981318 .1560177 -l.9l 0.063 -.6125648 .0163011 High schooll -.0263395 .1776728 -0.15 0.883 -.3844154 .3317365 _cons| -74.85802 16.71975 -4.48 0.000 -108.5545 -41.16158 H0: Old Ratio Coefficient = O F-Stat = 0.83 P-Value = 0.367 Regression 5 Suicide rate] Coef. Std. Err. t P>|t| [95% Conf. Interval] ........... -.+..-.._-----..----------------------------------.------------------- Old ratiol .093301 .1591544 0.59 0.561 -.2276645 .4142665 Child ratiol -.1932389 .1383718 -1.40 0.170 -.4722922 .0858144 Sexratiol 1.140948 .1795185 6.36 0.000 .7789142 1.502981 Employmentl -.3396205 .1539134 -2.21 0.033 -.6500166 -.0292245 High schooll -.0533829 .1739526 -0.31 0.760 -.4041917 .2974259 Whitel .0468161 .0259833 1.80 0.079 -.0055841 .0992164 _cons| -66.47563 16.95921 -3.92 0.000 -100.6771 -32.27413 ...
View Full Document

{[ snackBarMessage ]}

Page1 / 23

EconC175_Modeling+Effects+of+Economic+Burden+on+Suicide+Rates

This preview shows document pages 1 - 23. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online