WHY IS IT SO HARD TO KNOW WHAT TO DO FOR THANKSGIVING THIS YEAR AND WHY DON'T WE ALL AGREE? A data analyst's perspective. Thread.

MANAGING CORONAVIRUS RISK IS HARD BECAUSE IT'S NOT JUST ABOUT THE SCIENCE. Don't get me wrong. It's not completely unrelated to the science either. It's in between. Why? Because it requires us to answer questions that are unanswerable by science alone. Questions like...

1. WHAT'S AN ACCEPTABLE RISK? SCIENCE can't decide this for us. Logic can give us a guide for defining our risk preferences so that they're consistent. For instance, if I'm afraid of LESS risky stuff than coronavirus, it's illogical for me not to fear coronavirus as well.

Additionally, if I already have a more risky lifestyle than I want, I should rationally want to avoid the additional risk of coronavirus. Unfortunately, logic can only go so far. It can't tell me how much risk is worth the benefits of freely living my life. That's up to me.

2. WHAT ARE THE COSTS AND BENEFITS TO YOU? The costs and benefits of our decisions are specific to us and our situation. A mathematical modeler can't know what your life is like, but if they know your perception of the costs and benefits, modelers can put that into their models.

3. HOW MUCH DO YOU CARE ABOUT COSTS AND BENEFITS TO OTHERS? How much do you care about other people? Would you be willing to give up $100 of benefit if another person got $101 of benefit. What if it was only $99. What's the right level of trade-off? The answer is personal to you.

4. ARE YOU CONCERNED ABOUT COLLECTIVISM VS AUTONOMY? Would you prefer a plan that tries to get the best outcome for the group or just for you personally? Do you favor working with others to achieve common goals or does that sound like COMMUNISM to you? This is a personal choice.

5. WHOM DO YOU TRUST? This can determine what EVIDENCE you consider to be relevant evidence for making your decision. Most of us don't want to use data from untrustworthy sources. It's biased to exclude evidence AFTER you hear what the data says...

...but it's not necessarily biased to have prior criteria for what data you consider valid. Scientific knowledge can help with defining logical exclusion criteria but these criteria will always be a personal, subjective choice.

THESE DECISIONS HAVE TO BE MADE before we can objectively analyze available data and make the best decision. In order for experts to advise the public, they must make assumptions about the right answers to these questions. These assumptions affect their advice.

Sometimes the answers are based on their personal knowledge of the preferences of the end users of the advice. Sometimes the answers come from their personality or their politics. In my opinion, the general public isn't dumb at all to be skeptical of expertise.

The public is discerning enough to tell that there's a gap between their decision-making preferences and the preferences of the average expert. Unfortunately, as a society, we rarely do the work to translate the science into a form that respects individual-level preferences.

This might be because it's a vastly harder problem to re-engineer a system of simple top-level rules like "wash your hands" and "stand six feet apart" into a personalized rule set for each of the 7 billion humans globally.

I think we might need an app for that.