It is no surprise that I enjoy talking about data science and using data to make better decisions. Unfortunately, everyone is not as excited about data as I am. I occasionally face some resistance. One of the most common arguments goes something like this:
Why would I put in the extra work to find data that will tell me what I already know?
Because you might not really know what you think you know.
People are very comfortable making decisions based off gut feelings, assumptions, and personal biases. They obviously think the decisions are the correct ones. Many times the gut feeling is correct. However, the problem is those times when the gut feeling is not the right decision. In those cases, I would prefer to have some good data behind my choice. Here are a couple scenarios:
A Website Scenario
We should choose the white background because it produces 75% more signups than the blue background.
or
We should choose white because I got a good feeling people will like it better.
A Healthcare Scenario
You should try these pills because numerous tests have shown good results for people your age and weight.
or
I got a hunch these pills will make you better.
Both of these decisions are made everyday. Surely the doctor cannot knowingly prescribe a harmful drug, but many times the prescription is just for the standard drug matching the symptoms. Are all patients standard? Personally, I would feel more confident about the first decision in each scenario. But hey, I like data.
Why are people so skeptical to use data for better decision making? I do not know. Maybe some people are afraid the data might not match their assumptions. Maybe they just don’t want to put in the extra effort. I probably need some better data to properly answer this question.
Have you faced similar resistance? How do you convince people to use data for decision making?
Absolutely – especially with more “black box” (e.g. neural nets). One thing I like to do is always show a decision tree – it’s not always the best technique but it’s easy for non-technical types to understand. Some of the branches should generally relate directly to their experiences which helps “sell” the other more complicated stuff. Of course depending on the person this can also act as proof that they don’t really need data mining! Still, I’ve found a decision tree “on the side” can help to get buy in.
A decision tree is a good idea. Thanks for sharing.
Ryan
[…] Resistance to Making Decisions based upon Data (datascience101.wordpress.com) […]