Do the ages of death differ for different Ethnicities?

Introduction

Age at death is assuming that death was not caused by an accident can be a difficult thing to model given the many variables involved in determining an individual’s health. Several studies have been put forward at attempting to explain the potential population differences in the year of death at different times. The explanations for disparities in life expectancies have been widely explaining on the basis of socio-economic differences among members of the population. In the U.S, for instance, while there has been a general growth in the life expectancy due to generally improving life standards, researchers have observed a widening gap in the age at death of Americans (Olshansky, et al., 2012). The gap varies with total years of schooling, total household income and differences in certain races. Informed individuals generally have a high tendency to see factors such as hygiene, attend to proper medical care when seek and so on (Olshansky, et al., 2012). The medical implication on age at death can also be attributed to better methods in dealing with diseases through break throughs in medicine and medical technology effectively improving the body’s ability to fight infections. The impact of socio-economic differences can also be viewed on behavioural risks that increase likelihood of death such as obesity, stress and declined immunity due to the social and economic pressures among individuals. Although most studies indicate relatively economically marginalized groups dominated by the black community in the US, other studies have focused on behavioral aspects that puts in danger human life and reduce life expectancy.  The study by McGehee comparing the life expectancy of blacks and whites in Arkansas showed that the tendency to fall into high risk groups involved in crime, drug use and who do not emphasize on proper nutrition as a possible reason for the short longevity of black people. McGhee’s research shows that risky behaviors in the white community include unhealthy eating behaviors that expose them to lung disease, and also emphasizes that white people have a higher suicide rate in this research area. A contradictory finding was found by Kochanek, Arias, & Anderson, (2013) who observed that prior to 2010, black population in the U.S had an average a life longevity 3.8 years shorter compared to the white people due to the heart diseases, cancer, diabetes. While it is not clear why black individua have higher rates of these diseases, one can conclusively say that there is possibly little impact of socio-economy as these diseases point to life style most people lead. In fact, an isolation of the different factors that have been attributed to lower longevity across races, social-economic status was estimated to account for a maximum of 67%, while lifestyle diseases were estimated to account for about 17% of the reasons for short longevity (Mariotto, et al., 2018). In this regard, some studies have endeavoured to understand that while efforts are being made to reduce the gap in longevity attributed to racial differences, the efforts should also be focussed on the prevention of long-term illnesses especially among the most affected persons (Wong, Ettner, Boscardin, & Shapiro, 2009). This report aims at contributing to the body of knowledge of the racial factors that contribute to differences in longevity using statistical inference methods.

Data collection methods of analysis

The data were secondary data downloaded from the Centres for Disease Control and prevention (CDC) as yearly average age at death of the different races of people. The use of this information was for research purposes, and the data came in a format that did not specify the names of the owners of this information so privacy was maintained. The data was based on census estimates. Both descriptive and inference statistics using methods such summary statistics, graphics and simple one-way ANOVA were used. Where inference is used for comparison, the level of alpha was 0.05.

Statement of the task

The background of the study has provided that despite the many different factors that can be used to predict age at death, the most impactful are the socio-economic and demographic factors. Despite the improving life expectancy in the U.S and other developed countries, observed there are observed differences in the socio-demographic differences in the U.S, with educational attainment and overall income playing a vital role. The present study aims at isolating the demographic factors of race and assess possible differences in age at death of the sample members. While the age of death could also be genetically determined, it would appear that access to a quality life impacts the life expectancy of people as opposed to other different factors. Additionally, differences in the quality of life due to race are also gradually closing making the general American population “more homogenous” in terms of quality of life. However, equal quality of life and access to quality medical care have not been fully achieved so the study predicts that the race or ethnic group with more people in lower quality of life would have a lower age at death. I will start by calculating the mean, standard deviation and median.

Reason for research topic

My interest in this topic is drawn from an informal observation that people from relatively low socio-economic class who are in most cases people of the African American and non-white people tend to die at relatively young ages. A quick look on the internet is my suspicion that in countries where economic performances is poor, the average life expectancy of men and women is typically less than 60 years. However, some research suggests the importance of race when controlling for the economic condition. To get a definitive answer, I set to collect data and assess the age at death across different races. The statistical analysis used includes both inferential and descriptive statistics.

Data and Analysis

The raw data contains the average pf white people, the age of the African American and Hispanic race, and all genders are further divided according to gender. Each race however had here columns for the male and female. The data were for the years 1900 to 2017.

See table 1 below.

Table 1: Raw data

Descriptive Analysis and Mathematical processes

Across the three races, the females lived longer on average than males. Overall age at death for the white race was 75.75 years while for the African American the age at death was 69.81 years. The highest variation was observed in the African American females at a standard deviation of 7.11 years. Below is the table 2 with the summary statistics.

Table 2: Summary Statistics

The scatterplot of age at death for the three races shows that the Hispanic and White races had an almost similar trend over the years, of average age at death. However, while the data shows an initial fall back in the races age at death, at the beginning of 2015 the black race appears to have caught up with the other two races in terms of age at death. See graph below.

Figure 1: Scatterplot of Age at death among different races.

Inference

This section of the analysis estimates if the differences in age at death are statistically significant. The analysis carried out compared the means using a one-way ANOVA under the hypotheses;

(Age at death across the three race groups does not significantly differ)

At least one of the race groups significantly differs in terms of age

At least one of the groups significantly varied from the others in terms of age at death as seen in the lower value of p-value against the set level of 0.05, p=0.000

However, since the ANOVA table only shows there is a difference, it cannot be concluded safely which of the groups were different. Therefore, a comparison was made. The table below shows the absolute differences in means for the groups.

Using the Honest Significant Difference (HSD) statistic defined as

HSD=  if any of the above computed absolute differences are higher than the HSD then the groups being compared are significantly different. Q is a list of HSDs of 3 and 120 degrees of freedom, and MSE is the mean error in the ANOVA table. Therefore,

HSD=

Analysis of results

From the table, the Hispanic age at death was significantly higher than the black age at death 5.37>2.982 which was also the case with the White age at death and the black age at death 5.95>.2982. Hispanic and white age at death were however not significantly different since their absolute differences did not exceed the threshold value of 2.982 at 0.05 level of significance.

Validity and Conclusions

The interest of the present study was to understand whether age at death of the different races significantly varies. The chosen method of analysis compared the means across the races (race being a categorical variable and age being the dependent and continuous variable). The method of choice was ANOVA method, which was the appropriate method. While people identify with certain races, possible intermarriages in the family tree with people from other races may have hindered proper identification of race, hence creating a possible confounder. Future studies ought to do analysis based on observational data as opposed to census data to help understand at the grassroot level the differences in age at death of different races. In conclusion, our study results resonate with those of (Mariotto, et al., 2018; Kochanek, Arias, & Anderson, 2013 and Wong, Ettner, Boscardin, & Shapiro, 2009) that people from the African American race have significantly lower age at death.

References

Kochanek, K. D., Arias, E., & Anderson, R. N. (2013). How did cause of death contribute to racial differences in life expectancy in the United States in 2010? . US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics.

Mariotto, A. B., Zou, Z., Johnson, C. J., Scoppa, S., Weir, H. K., & Huang, B. (2018). eographical, racial and socio-economic variation in life expectancy in the US and their impact on cancer relative survival. PloS one, 13(7).

Olshansky, S. J., Antonucci, T., Berkman, L., Binstock, R. H., Boersch-Supan, A., Cacioppo, J. T., & Jackson, J. (2012). Differences in life expectancy due to race and educational differences are widening, and many may not catch up. Health affairs, 31(8), 1803-1813.

Wong, M. D., Ettner, S. L., Boscardin, W. J., & Shapiro, M. F. (2009). he contribution of cancer incidence, stage at diagnosis and survival to racial differences in years of life expectancy. Journal of general internal medicine, 24(4), 475-481

Sciencing. 2020. What is R2 Linear Regression? | Sciencing. [ONLINE] Available at:

https://sciencing.com/r2-linear-regression-8712606.html

http://pages.stat.wisc.edu/~st571-1/Fall2005/lec18-21.2.pdf

http://www.nr.usu.edu/ArcWebpage/old_wp/2001%20presentations/NASA/ARC%20nstudy.pdf

Appendix

Table 1: Raw data

Hispanic and white age at death were however not significantly different since their absolute differences did not exceed the threshold value of 2.982 at 0.05 level of significance.