I can't find your later message, but the fact that the data were collected at different times does not invalidate the independent samples test. They are still two independent samples, but you cannot determine whether any difference found is due to the change in procedures or the difference in times or both. Those two possible factors are completely confounded with these data.
------------------------------
Jon Peck
------------------------------
Original Message:
Sent: Tue April 20, 2021 04:53 PM
From: Jon Peck
Subject: What procedure is best for this odd design?
Why is this not independent samples?
------------------------------
Jon Peck
Original Message:
Sent: Tue April 20, 2021 10:48 AM
From: Serdar Atav
Subject: What procedure is best for this odd design?
Good Afternoon Folks:
I think I have a simple question:
I have a data set for outcomes of individual patients PRIOR to a change in procedures; and then outcomes of DIFFERENT patients AFTER the change in procedures. Basically the design looks something like this:
O1
X O2
It is not paired samples but it is not quite independent samples. Would it be appropriate to use independent samples t-test to analyze this data or is there a better way to go at it.
Thanks in advance for your ideas.