On Instagram @AndreaHardyRD, I talk all about how important it is to consider statistical significance.

  • Image of @andreahardyrd Instagram post about statistical significance


Statistical significance, but make it fashion 💰🤩 #scicommunitythursday⁠⠀
⁠⠀
One time I won a LOT of money at craps rolling 6’s in Las Vegas. Was it due to chance? Or did I have a rolling technique to actually roll 6’s consistently? ⁠⠀
⁠⠀
We could find out the likelihood of that through statistics! ⁠⠀
⁠⠀
Many times, when we read papers, we may look at P-value. This is the probably looking at whether:⁠⠀
⁉️the result due to random chance⁠⠀
⁉️or due to a specific effect (skill, in this case?)⁠⠀
⁠⠀
The lower the P-value, the more likely the results were due to the intervention.⁠⠀
⁠⠀
While you may get a statistically significant result based off the number of rolls I made (many times statistical significance is set at P<0.05 meaning we can say with 95% confidence the result was due to the intervention and not chance), we also need to recognize the limitations of that. ⁠⠀
⁠⠀
I can tell you – my rolling had no technique. (Unless vodka was the technique 🍹)⁠⠀
⁠⠀
Yet, it may have been statistically significant – there’s STILL the possibility that it’s due to chance (it was.) Which is why multiple studies, studies need to be powered adequately (the right amount of statistics), and the right type of statistical methods are used to draw conclusions about things! ⁠⠀
⁠⠀
I am not a statistics expert, and this post reminded me why that is – so I’m thankful for all the statisticians that make interpreting research possible! ⁠⠀