Here is the premise of the study: A surge in suicide rates has been clearly observed in unemployed individuals, particularly men. Cardiovascular diseases peak in face of financial stress and preventive ontological care declines, thus contributing to excess mortality. A ph.d is not needed to see there could be upticks in suicides and a reduction in slef-care when people lose their jobs and their health care. if you look closer at the paper you will see it is a time series autoregression analysis taking mortality rates and unemployment statistics. We rely on a Bayesian method to construct posterior estimates of model unknowns including the VAR coefficients and the smoothed growth rates of the life expectancy and mortality rates. Equipped with posterior estimates for the VAR parameters, we use an impulse response analysis based on a Cholesky decomposition method to assess the effects of an increase in unemployment on the life expectancy and mortality rates. The main message arising from our exercise is that the typical unemployment shock results in a statistically significant decline in life expectancy and increase in mortality rates for the overall population. It is 3 variable auto regression... If you ever did econometrics you will see these kinid of time series have wide standards of deviation and present all types of errors. To say increased unemployment leads to an uptick in death rates was already well proven (people lose jobs, they lose health care and stop preventive measures). To try and add race as a 4 variable creates what we call in statistics garbage in-garbage out.
Also the study found the percent change in age adjustmed death rate impact is highest on White Males v. AA Males. But that is not a headline worth printing even though based on faulty estimations. The study failed to account for the variable of economic class. This is a common tactic in studies to predetermine what variables you choose because you are prebiased in the outcome. A real study should include all variables of race, M/F, and economic class to see the effects across the population. Let me save you the Ph.D..... unemployment shockes lead to higher death rates, those rates are higher in poorer groups who suffer unemployment at higher rates than middle and upper class groups. Most academic studies are just Ph.ds trying to make themselves relevant since you are mandated to publish regularly. I just summarized the findings without lifting a finger or using made up statistics.
Just like with global warming studies.... and studies about tax cuts... and many others... If a study uses a model and its published it almost always reflects the bias of the researchers and / or those paying for it. (because they can select the inputs and the assumptions) The readers job is to discern whether the model is useful.
its funny cause the algo has nailed the presidential elections 3 times in a row. Tony the moron will post my personal joking around from the romney loss ... but the algo tells us to use the previous election turnouts as the template for unskewing the polls allowing for about 1-3 points of slosh. It picked Trump's election perfectly And this year it stated that Biden was winning inside the margin for error. I also stated it all depends on the turnout this year and I could not predict the turnout.
You weren't joking idiot.Just as you weren't joking when your dumb ass said the birthers would win their birther lawsuits and Obama wouldn't be on multiple state ballots in 2012 you stupid motherfucker.