Two things the COVID-19 Pandemic has taught me {Partially due to listening in on a UPenn Social Work Research Methods graduate class being taught from my living room via Zoom} :
- Numbers Matter – From transmission rates to job loss percentages in specific demographics
- I Am Not a Statistician AND I Should Not Play One on Facebook
I am quarantined with a professional, published
statistician whose work is cited around the world. She has reviewed and written on diverse studies from incarceration data for the National Institute of Justice to drowning data for NOAA National Weather Service to mental health programs for the University of Pennsylvania Center for High Impact Philanthropy. Few of us will be publishing papers {Viral FB posts do not count} on the COVID epidemic, but the scientific process and concepts that Michelle touches on may help you filter the FAKE from the REAL and the SUPFERFLOUS from the CRITICAL. We are all going to be faced with personal decisions that will impact the health and well-being of ourselves, our families and our communities. Here’s a little framework to help you through the process.
From Dr Michelle Evans-Chase:
I am a stats geek. I love stats and the scientific method.
I also embrace how this unnatural love guides my understanding of the world. Because of this unnatural love for the very things that my students dread most (research methods and statistics) I found myself curious about the Covid-19 numbers presented in Governor Murphy’s press conference as measures of where we are as a state in reducing the spread and public health consequences of the virus. What I was curious about was what those numbers could tell us if we understood them in a more explicit context.
In statistics, context is everything. Without context numbers become “floating statistics” and of themselves, can become meaningless. When presented with numbers that represent anything having to do with people I always ask: 1) compared to what/who? If “4 out of 5 dentists prefer trident chewing gum” – what do they prefer it to? Chewing sugar cubes? Tobacco? And 2) what would we expect? If 20 people drop out of a race – is that average? More than usual? Fewer than usual?
So, for no other reason than my own curiosity I decided to put the numbers of new hospitalizations reported across NJ on May 11 into the context of population proportionality (not even a real term, but you know what I mean). I wanted to know: compared to what? What would we expect? To do this I found some old(ish) population numbers for NJ broken down by region: North, Central, and South, similar to the way the state was broken down in the press conference numbers. I found population numbers from 2013 here https://www.nj.com/news/2015/04/mapping_njs_unofficial_north_central_and_south_jer.html.
Not ideal but again, this was for my own interest, not something I planned on submitting for peer review. With these numbers I could now place new hospitalizations in the context of regional populations and see if they were roughly equivalent given the population of each region. If the regions were equivalent then we would expect the number of cases to be roughly proportional to the percent of the NJ population of that region. This is what I found:
| Population | % NJ population | New Hospitalizations | % New Hospitalizations | |
| North | 3866532 | 0.44 | 179 | 0.50 |
| Central | 3042937 | 0.34 | 63 | 0.18 |
| South | 1922937 | 0.22 | 118 | 0.33 |
| Total | 8832406 | 360 |
South Jersey makes up about 22% of the overall population of NJ. If South Jersey’s proportion of new hospitalizations reflected this, then we would expect it to have roughly 22% of the new hospitalizations in the state. However, South Jersey had 33% of the state’s new hospitalizations on May 11 and therefore had a disproportionately high number of the new Covid-19 hospitalizations.
Comparing regions directly, South Jersey had roughly ½ of the number of people in its region compared to North Jersey. On May 11 North Jersey had 179 new hospitalizations and thus, all else being equal, we would expect South Jersey to have roughly half as many new hospitalizations. But rather than the 88 new hospitalizations we would expect (179/2=88), South Jersey had 118 new hospitalizations. Again, South Jersey’s numbers were disproportionately higher than we would expect if the difference in the number of cases between South and North Jersey where simply a reflection of the difference in size of each region’s population.
From Bruckner:
So whether you are in South Jersey or the South Pacific, whether your news source is NPR, the BBC or Trevor Noah the key take-away is to recognize that numbers {And people} need context. Whether you are staying at home, going to work, protesting or posting take the time to think critically and find the context that aligns with your situation and life. We wanted to share the building blocks for applying science and research based critical thinking . We encourage you to take the time to apply the same to your decisions to stay safe, healthy and sane.
BLUE JOURNEYS 