Question

Do you eliminate gross error before computing Dixon Q test?

Answer 1

what do you mean by gross error, specifically?

Answer 2

When running one of my trials, in the weight room someone nudged my arm, so a little of my sample splashed out of my dish... Isn't that considered a gross error?

Answer 3

and so your value was far off from the expected?

just do a q test for that value you got, you have to show that it's an outlier, statistically. Clearly, you can eliminate it without performing the test, but, you should show the work

so the answer should be no.

but once you eliminate it (by showing that it's an outlier) you can exclude it from the range

Answer 4

The data for that one trial is right in the middle of my range... So I wouldn't believe it would be an outlier... But my professor said that all gross errors should be rejected.

Answer 5

that's weird, it would be right in the middle of the range if you spilled some of the contents. But i guess if that's what your professor said, then you should do so. So, yes i would eliminate it from the range before doing the q test for other potential outliers

Answer 6

I found it to be weird that the value is in the middle of my range too. And she did say that we should eliminate all gross errors we KNOW about. I marked all my samples differently so I would know which one is which... That's why I know which particular trial spilled. But I wasn't sure...

And if I am to eliminate it... instead of N=4 for the 4 trials I completed it would be 3?

Answer 7

yeah, you would eliminate the data point (including it's trial)

Answer 8

Okay. thank you!

Answer 9

no problem.