@JonathanRepple we actually simulated that for standard sample sizes in neuroimaging: https://t.co/X9t3ATqY0D
2,287 followers
2,287 followers
@GordonFeld Very nice visualisiation and explanation! We looked at this in 2D with @johalgermissen. way lese fancy compated to yours, but maybe still of interest: https://t.co/KPyIQ0BLGw
2,287 followers
@johancarlin @cl445 @UK_Muenster @INM7_ISN @TheFlorey @KordingLab @talyarkoni @danilobzdok Sorry for the late reply, just saw this. Yes, this is what we mean with bias - small N papers are overrepresented in the literature. Here a short demonstration with
2,287 followers
In our mini-review on power in neuroscience w @johalgermissen we have simulated the inflation of ES for a commonly used N (25) and a 80% powered N (90), assuming a true effect of d=0.3 (alpha 0.05) under publication bias and found a similar difference http