You're supposed to publish results that can be reproduced as you describe. And it wasn't an outlier. The simplest comparison I can make to illustrate the poin

You're supposed to publish results that can be reproduced as you describe. And it wasn't an outlier. The simplest comparison I can make to illustrate the point without boring you is measuring the power of a car engine. Imagine that you get a very high torque number, but horsepower doesn't match. Issue is: torque and horsepower are linearly dependent torque = HP * RMP / 5252 Measurements that linerarly correlated with the voltage published were all consistent with a number 10/15% lower.