Removing outliers in paired samples

[Svetlana Bondar]

Hello.

I have two paired samples - patients' blood pressure at 2 visits to the doctor. I see outliers at visits 1 and 2 (3 outliers each) and when comparing the difference between visits (4 outliers). What outliers should I remove from the analysis? Only on visits? I tried to do this, but I got other outliers when calculating the difference.

[Carlo Lazzaro]

Svetlana:
unless the so called outliers are blatant examples of mistaken data entry (or implausible bp measurements), you should not omit those observations.

[Svetlana Bondar]

Svetlana:
unless the so called outliers are blatant examples of mistaken data entry (or implausible bp measurements), you should not omit those observations.

Thanks, Carlo. I will use them.

[Svetlana Bondar]

Thanks, Carlo. I will use them.