Monday, April 17, 2017

Stress and Elections

Can you predict a vote based on the level of stress groups of people are feeling? Maybe.

Recently, WalletHub published the list of states that have the most stress (Alabama) tops the list. (Minnesota has the least stress.) When this is compared to the margin that Democrats or Republicans had during the presidential election (by state), there is a high correlation between two groups of stress factors. But not with any one set of factors alone. The closest is that Health and Safety Related stress factors explain 26% of the Democratic-to-Republican presidential vote margins.
These two sets of factors shown on the chart above TOGETHER explain almost all of the election results but not alone. Reminder: Correlation is not the same as causation. Just because there is a correlation (like the level of Bangladeshi butter imports to election results) doesn't mean that there's a cause-and-effect relationship...unless you can describe the mechanism for it. If it's simply wealth and  health, there have been other studies that show rich people live longer because of access to high quality healthcare.

For the statistics nerds, when you run the data through multiple regression ANOVA analysis, a p-value result of 0.0000...(nearly impossible that none of the factors explain the results) with two groups being significant and two groups not being significant.

The Money and Health factors are shown, as described by WalletHub. Money stress factors:

  • Median income
  • Debt per median income
  • Personal bankruptcy rate
  • Share of population below poverty line
  • Housing affordability
Health and Safety stress factors:

  • Share of adults who felt stress yesterday
  • Share of adults in poor or fair health
  • Share of adults diagnosed with depression
  • Mental health
  • Suicide rate
  • Unaffordability of doctor visits (which would be related to money)
  • Share of population with health insurance coverage
  • Psychologists per capita (hopefully high if you have a higher share of depression)
  • Prevalence of binge drinking
  • Physical activity rate
  • Share of adults getting adequate sleep
  • Rate of bullying incidents
  • Crime rate per capita
  • Quality of infrastructure
By the way, the two groups of stress factors that don't explain much of the presidential results are Work Related and Family Related stress. Of course, we probably have all experienced stress in one area that carries over into stress in other areas. So there you go.

If you (and here it would be the political party leaders of the Democrats and Republicans) believe there is causation, then the data might be helpful in setting strategy for the next election.

No comments:

Post a Comment