In a recent QST article, (Gable 2021) gave the following advice:
SelfSWR is commonly known as Insertion VSWR. The article contains several errors in definition of SWR, Return Loss and Insertion Loss… but suffice to say that he uses \(InsertionLossdB=s21dB\).
In the example given above, VSWR=1.2 means ReturnLoss=20.8dB (and s11=20.8dB).
InsertionLoss in this case is due to input MismatchLoss (failure to capture the maximum power available from the source) and Loss (conversion of some energy to heat).
Let’s check the last statement in the quote by calculating the scenario using Calculate Loss from s11 and s21.
InsertionLoss of 0.5dB in the presence of input VSWR=1.2 (s11=20.8dB) has Loss=0.4637dB. We can calculate the power lost in the switch itself
as heat as \(P=50 (110^{\frac{0.4637}{10}})=5.06 \: W\), somewhat less than Gable’s 5.8W. (Since the calculator gives loss as a numeric value of 1.113, a simpler calculation is \(P=50 (1\frac1{1.113})=5.07 \: W\).)
Note that the inferences of these measurements with a VNA apply to sources that are well represented by a Thevenin equivalent circuit with Zth=50+j0Ω. Most ham transmitters do not comply with that requirement, and the power output under mismatch is best assessed with a directional wattmeter.
It is difficult to understand why QST would publish an article that purports to drill down on these effects when it is not based on sound concepts and theory.
References

 Calculate Loss from s11 and s21 – convenient online calculator
 Gable, R. Feb 2021. Identifying common feed line miscreants. In QST Feb 2021.