The fact is that food-frequency questionnaires are flawed BUT The flaws are EQUALLY DISTRIBUTED This means that everyone answers them wrong, but in the same ways (usually underestimating calories) which makes them extremely useful for researchhttps://twitter.com/dnunan79/status/1038326067298217985 …
-
-
The best part about issues like these is that they can actually make a study stronger! This sounds counterintuitive, but it is often true
Show this thread -
This is because issues that affect all groups in a study tend to bias the results towards the null hypothesis - i.e. they tend to make any results you DO find more significant
Show this thread -
For example: if you measured calorie intake across 5 groups, and the 'true' result is 1. 1800 2. 2000 3. 2200 4. 2400 5. 2600
Show this thread -
Now, generally speaking people underestimate their calorie intake. Let's say using this food questionnaire it is by 20% So the values you OBSERVE are: 1. 1440 2. 1600 3. 1760 4. 1920 5. 2080
Show this thread -
In our first example - the TRUE value - the difference between each group is 200 In our second example - the OBSERVED value - the difference is only 160
Show this thread -
So what's happened here is that the issues with our study have changed the results, thus biasing towards the null hypothesis
Show this thread -
If we still manage to find a difference, it means that the difference is even more compelling An issue with our measurement is now actually making our results even more impressive!
Show this thread -
This is one of many reasons why epidemiology can be pretty confusing End thread
Show this thread
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
+ RT
Lolz