Suggestions

:speech_balloon:

Insights: correlation is not causation

Currently insights are worded as “This factor makes you Feel…” but what it’s actually showing is correlation, not causation.

For example, I’m not a regular caffeine drinker, but I often use small doses of caffeine to function on days when I’m suffering from a lot of fatigue. Thus, my insights note a strong correlation between exhaustion and caffeine use. However, the way it’s displayed in the app suggests that caffeine is making me feel exhausted, which is silly. Of course I can ignore that wording, but the correlation/causation fallacy is worryingly common in our society; many users may get the wrong idea from the app’s wording.

I suggest changing the wording to something more neutral regarding causation, and adding a note in the walkthrough/help explaining that the correlations generated by insights can help users discover relationships but they shouldn’t necessarily be interpreted as causal without further investigation.

For the insights screen, I suggest rewording “This Factor makes you Feel…” to say “This Factor relates to when you Feel…”

57 votes

Tagged as Suggestion

Suggested 20 July 2021 by user Bex

  • Sign in to comment and vote. Sign in by email
  • 20 July 2021 Bex suggested this task

  • 21 July 2021 James @Bearable approved this task

  • avatar

    Though just a word, this is actually bothering me too every time I read it. It’s as if I first have to overwrite that “wrong” information in my brain. Would be great to see that fixed.

    02 September 2021
  • avatar

    Bex, I wonder if this could be accompanied by a time-based correlation system that treats the earlier factor as the “lead” and the later factor as the “follower” when it comes to the correlations? Then you could see your exhaustion as the”lead” and the caffeine as the “follower”.

    04 February 2023
  • avatar

    That’s a great idea. I’ve been frustrated that it’s difficult to track whether a factor is actually helpful or not, this would help

    07 October 2023
  • avatar

    I have this problem a lot too. I also find that many of my factor affect how I feel the next day. When I am feeling at my worst, my husband will give me a massage which helps me sleep. Good sleep leads to better days the following day. But in the correlation overview, massage is always correlated with my worst days. I know that I can use the “deeper” insights to see next day or following two day effects. But it would be nice if there was a way for Bearable to recognize particularly strong next-day or future correlations and bring them to our attention.

    26 November 2023