We’ve been talking with journalists around the country during this crisis, and have seen some recurring opportunities to enhance COVID stories that feature a data component.
Some of these points are well-known and oft-used; some are not. But they are all important.
At the Policy Data Institute, we work with journalists to get them the data they need to tell interesting and accurate stories. We offer our own compiled data as well as access to our team of analysts – all for the cause of better journalism.
Feel free to contact me to discuss how we can provide you with data, insight and analysis – even on short notice.
Ted Ballantine
Cell: 847.371.1964
Email: EBallantine@CivicMeter.com
____________
1. Look for peer comparisons.
Show comparisons to similar cities or counties – by population density, income, climate and other measures. Compare with locales and states with similar business opening policies.
2. Localize.
Do not stop at state or even county-level data. Use city, zip code and other granular geographic data.
3. Contextualize case counts with testing data.
Focus on meaningful testing data – number of subjects versus number of samples; serum anti-body testing versus direct swab virus testing. Look for rationales for testing benchmarks as promulgated by local and state health officials.
4. Look for impact on sub-groups.
In addition to income and race, look for age, local, neighborhood density, among other factors.
5. Data triage: which COVID-19 numbers to emphasize.
Use per capita, number of deaths, number of hospitalizations. Daily cases are a lot of noise – look at changes in the seven-day moving average.
6. Follow up with reacts from experts and decision makers.
For any data observation, experts and high-level decision makers can offer context and relevance. Data + insight = better stories. One extra data feed and one extra phone call can generate improved journalism.
7. Use simple graphics.
First, don’t forget the graph. Second, keep it simple. Really simple. You may lack the NYT animation flair, but reporters will be better able to connect the data to purpose of the story. Simple bar and line graphs are often better than graphs with bubbles of different sizes, floating in different quadrants, often in different colors.
8. Avoid lower-caliber medical data.
Aside from the fraud and controversies, reporters often treat all medical information as being equal. It is not. Be sure to weigh the value of the information by not just the messenger (e.g. CDC), but also the source (quality of the journal). Understand study design issues when evaluating the study results (cohort, case studies, etc.).
9. Communicate probabilities and likelihoods along with the black and white decisions.
Readers are impatient – but not dumb. Yes, they want answers. However, many will appreciate nuance and reality along with drama and certainty.
10. Look for reinforcing factors.
As with the sub-groups, look for data that could explain a trend. Climate; housing density; income; race; focus of cases in jails and nursing homes. These factors explain trends as well as highlight conditions among sub-groups.
