I recently flew on United Airlines, and one of the passengers did not wear a mask for most of the 8-hour flight.
This action infuriated me, so naturally, I vented about it to a friend a few hours after landing. I lamented my fellow passenger’s callous disregard for human life, hardcore judging his lack of effort to protect the strangers around him. Apparently my friend did not share the same sentiments, and our conversation devolved into a long argument about the effectiveness of wearing a mask to reduce the transmission of Covid-19.
In short, my friend believes that mask mandates are not an effective means of preventing the spread of the coronavirus pandemic.
“CDC research says masks don’t work!” — anonymous friend
To be clear, the CDC does not say that. However, I had no idea that my friend held such beliefs about masks. Neither one of us are ideological zealots. Quite the opposite, we are both fact-driven, science-minded, educated young professionals, so I was curious what evidence he had to support this view. While a number of different points were raised, I would like to focus on one piece of his supposed evidence from the CDC in the hopes that doing so might save other people from misinterpreting this data.
If you consider the table in isolation as it was presented to me, it seems to suggest that masks are not effective in preventing Covid-19. The section in red shows that there is basically no difference in mask wearing behavior between the case-patients with Covid-19 and the control participants. Furthermore, the p-value highlighted in yellow suggests that we can’t statistically distinguish between mask and non-mask wearers who got sick.
Case closed! We can all take off our masks now.
Not so fast…
In order to properly understand this table, we need to define a few terms:
Control group — the standard to which comparisons are made in an experiment. The control group and the experimental group are used to measure the effect of the treatment.
P-value or “Statistical Significance”— the probability of obtaining results at least as extreme as the observed results of a…