Network Analyzer Normalization(Callibrated vs Uncalibrated measurement)

Hello,

I have a question regarding Network analyzer Normalization( Calibrated vs Uncalibrated measurement)

My counterpart is measuring a Insertion Loss drift transmission line and from 10MHz to 67Ghz. Currently Im trying to duplicate the "drift failure" that he sees in the transmission line.

The setup
IF BW = 1Khz
Start Freq = 10Mhz
Stop Freq = 67Ghz
Power = -15dBm
Normalize with Uncalibrated Network analyzer
DUT is connected with High freq cables, 1.85mm connectors.

The response that he is getting something like this over a certain time period

[/url]

Question;

Is the method used to measure drift is valid while the Network Analyzer is not calibrated?

I understand that in a uncalibrated system, the power received by the DUT might be lower than what is set on the Network Analyzer(-15dBm), and the insertion loss (due to skin effect and long transmission line) at high frequencies would be worst(stability wise)

Appreciate if someone could enlighten me on this issue.

Thanx in advance.