The Data Detective: Introduction, Rules One + Two

As you read through the Introduction and Rules One + Two feel free to use this as a place to share your thoughts, questions, ideas, and reflections!

Rule 1: Search your feelings

This chapter reminds me of the Feynman quote: “The first principle is that you must not fool yourself and you are the easiest person to fool.” You can easily treat your theories like your children – it is hard to be objective about them.

There are so many cases where feelings cause people to be less thorough than they should be. For example, there is a famous story where CBS and Dan Rather were duped by six documents containing false claims about George W. Bush. Within hours of releasing the documents, experts were able to show them to be fake. Dan wanted to believe his story so badly that he did not look at some basic things like the use of a proportional font on a typewritten document.

Rule 2: Ponder Your Personal Experience

I used to work with a very good Procurement manager who used something she called the “Snicker Test” – does the claim agree with what she has seen to be true in the past? She would carefully explain what she had seen in the past and then ask “What is different now?” I was amazed at how often she found issues with vendor claims by calling on her previous experience and asking what is different this time.

1 Like

these are such great points! and the idea of a “snicker test” is fantastic - it really highlights the need to slow down a bit and reflect, and not always jump in based on what we’re thinking or feeling in the moment!

I loved how the author presented both extremes: looking at statistics from the viewpoint of a tool that can be used to manipulate people and ideas, and also being indispensable to public health research.

one thing that I’d love to hear from everyone is times that you’ve (successfully!) used data to convince stakeholders about something - it doesn’t have to be as monumental as the correlation between smoking and lung cancer!

I think a lot about my early days in data science, when I would just assume that everyone had a baseline level of knowledge in data work. I’d pull together these big, beautiful reports and send them out thinking “ah yes, NOW we’ll make some changes!” and they never had the effect that I expected. it wasn’t until a coworker pulled me aside to tell me that the reports were pretty, but no one understood how to read the graphs or interpret the statistics :sweat_smile:

related, I’d also love to hear if anyone has any favorite strategies around teaching//communicating statistics and data sciency-results to non-stats folks!