May 11, 2022 | News | Interview
“No Single Metric Is a Panacea”
Jiaxin Shi, PhD Student in the Research Group: Lifespan Inequalities and colleagues suggest to use the stratification index to analyze mortality inequalities between groups. In the interview he explains why this metric offers a new perspective and what he found when applying it to Finnish register data.
Mr. Shi, traditionally, researchers who studied mortality differences between groups focused on metrics that describe average levels such as life expectancy and standardized mortality rates. You suggest another metric. Why?
Comparing averages is what social scientists do most of the time. Life expectancy is the average of a lifespan distribution. But this does not tell the whole story. A good example is lifespan variability. It shows the dispersion of a lifespan distribution. There are other dimensions, such as median, mode, skewness, etc. They give complementary perspectives.
What exactly does this metric, the stratification index, look at?
Visually, the stratification index looks at the proportion of non-overlapping areas in two lifespan distributions. A larger proportion indicates that the two distributions are more distinct. In statistics, the stratification index and many other related measures quantify the statistical distance or dissimilarity of distributions. Measures like life expectancy and standardized mortality rates can also help compare mortality differences between groups, but they focus on the average, not distributional differences.
What advantages does the stratification index have for you?
One advantage is that it enriches our understanding of how mortality is different from (sub)populations beyond the conventional comparisons of averages. Hence, it offers an alternative perspective of mortality differences. As we show in the paper, lifespan stratification can increase even when differences in life expectancy become smaller, and it is the same for the other way around. So, it uncovers new dimensions of inequalities that we have paid less attention to previously.
Another advantage is that it reflects the sociological concept of “social stratification”. Like the stratification of rock, social stratification refers to the phenomenon when a society is hierarchically layered based on social characteristics such as class, gender, and ethnicity. Our index can quantify this sociological concept.
… and the disadvantages?
One is that it is a relative measure, so it does not tell how big the actual gap is, which might be more relevant for health policies.
You applied the metric to Finnish register data. What did you find out?
The most important finding is that, overall, lifespan stratification between low- and high-income individuals has been increasing between 1996 and 2017. For men, there was a slight decline in recent years. This is in contrast to the steadily narrowing gap in life expectancy in the recent decade.
What does this mean for the society?
Rising lifespan stratification by income means that poor and rich people are experiencing increasingly different mortality regimes. This applies not only to individuals themselves but also to people surrounding them—their friends, relatives, etc., because people tend to be around people with similar social status. This puts an extra burden on the social network which is important for health. Like any other kind of social stratification—no matter what the causes are—a large lifespan stratification can lead to individual emotional stress and even societal unrest. Thus, rising stratification calls for policy actions on reducing health inequalities.
What kind of data does one need to apply the stratification index?
The metric can be applied to contexts other than lifespans— for example, healthy lifespan, or to be used to compare between education, sexes, nations, etc. You will just need the distributions for each group. Individual-level data are not necessary.
What do you conclude from your findings?
No single metric is a panacea. The best solution for mortality analysts is to incorporate a range of metrics to monitor how mortality inequalities evolve over time, thus providing comprehensive and timely information.
Shi, J., Aburto, J. M., Martikainen, P., Tarkiainen, L., van Raalte, A. A.: A distributional approach to measuring lifespan stratification. Population Studies (2022). DOI: 10.1080/00324728.2022.2057576