Danger, Danger

In my recent online time-wasting, I’ve been seeing colorful charts that categorize activities (Tennis! Casinos!) by COVID-19 risk.  There are several versions, but they all sport a color scheme ranging from calming green to ominous red, use qualitative terms such as “medium-low risk”, and are labeled with official-seeming numbers from one to nine.  People seem to really like them.  I get why: we all want clear guidelines to help us navigate the medical and sociopolitical plagues of 2020.  But these over-simplified charts give my inner logician hives.

My first gripe is that the selected activities are peculiarly specific, and seem to have been designed by people with a lot of leisure time and no job experience as baristas.   Is playing tennis equivalent to ineptly throwing a Frisbee around in a park?  Are casinos like community theaters, or are they more akin to the shared computers at public libraries?  

My second gripe is that the activities are frustratingly vague.  Yes, it’s possible to be too specific and too vague at the same time.  At the casino, is everyone wearing a mask?  Is it a sparse Tuesday morning? Am I visiting once for an hour, or gambling from dawn to dusk every weekend?   

My third gripe… gets nerdy. 

When I teach Natural Resources Measurement and Inventory at the University of Alaska, Fairbanks, I start the first lecture with a discussion of measurement scales.  This is one of those things we all think we understand, but don’t.  (Later in the class we get into statistics, which is one of the things many people already know they absolutely don’t understand.)  Many types of scales are useful, but if you mix them up, you’re going to cause problems.

Ratio scales start at zero, and each increment is the same size as every other increment.  Six quarts is three times as much as two quarts.  We’re a bit less familiar with logarithmic scales.  Each increment is some multiple (usually ten) of the previous one.  An earthquake that registers at 7.0 has ten times the measured amplitude of a 6.0.  With interval scales, the increments are all equal, but there is no true zero.  If you wear size twelve shoes, your feet are not twice as long as those of someone who wears size six.  Some scales are merely ordinal – gold, silver, bronze.  The winner of a race might cross the line a hairsbreadth ahead of the second-place finisher, and then there might then be a laughable lag before third place arrives.

In the case of the colorful COVID charts, the scale seems like it might be a ratio scale.  In fact, it begs to be read as one – otherwise, why have numbers at all?  But it definitely isn’t. Three “level-three” masked trips to buy corn flakes and lettuce are not as likely to make you sick as one “level-nine” wild night in the mosh pit.  Going for seven different “level-one” jogs is not equivalent to attending “level-seven” junior high for a year.  To me, the one-through-nine scale seems closer to being logarithmic – maybe in binary, or base three? – but no way was it designed or calibrated that way.  Thus, all we can really say about the charts is that they are ordinal: two is worse than one, and eight is worse than seven.  But how useful is that in governing our choices, not to mention the apportionment of our limited angst supply? 

I’d like to create a model that uses something closer to real risk values to assign meaningful numbers to a wide range of activities, so that we can compare casinos to tennis in a way that warms my scientist heart.  But I have too many hobbies already, including sewing masks and endlessly scrolling online, so it might take a while.  In the meantime, it’s back to the basics, otherwise known as “knowing something about how COVID-19 spreads”.

Distance from other humans is crucial.

Length of exposure matters.

Air flow makes a big difference.

Masks help.  A lot.

Pay attention to the medical professionals.

Ignore the politicians who aren’t paying attention to the medical professionals.

Close the damn bars.

As noted, when I teach, I start with a lecture about measurement scales.  I end with one about the various ways in which graphs and charts can be misleading.  In this vein, I made a new chart.  It’s not any more useful than any of the others. 

You’re welcome.