With the usual caveats, start with the relevant frozen Design Standard and ACs.
Background reading in how to do it might usefully start with
(a)
FAR 25
(b)
AC 25-7C.
(Noting that your aircraft of particular interest may have used now-superseded data in which case you need to dig around for that older data rather than use the current data above).
Generally, any OEM is out to get the absolute best results feasible during the test program. Dud runs will be discarded, in turn, by the FT crew, and then the performance aerodynamicists back at the factory. A bit more argy bargy will ensue with the particular Certification folk to end up with an agreed basis for the AFM.
(a) idle thrust selection will be appropriate to the aircraft's handling characteristics and one should refer to the flight procedure in the AFM to get an idea of what might have transpired.
(b) the 50ft point is worked out post flight test by analysing the FT data records
(c) actual test speeds will be as close as the FT crew can get them to the nominated Vref. Final data tweaking post test will refine the data for any minor errors. Generally, tests are done in optimum weather and the TP generally nails the numbers pretty accurately.
(d) test analysis will be to get basic raw data referenced back to the desired numbers
(e) AFM landing data will be according to the regulatory requirements. The AFM will disclose what the specific data basis is for the Type and Model.
(f) operational landing data will use factored data - normally 1.67 for normal operations. This is the pilot's basic protection and please accept that the line pilot out in the real world isn't going to be able to reproduce the test data. The raw data should be viewed as the absolute minimum brick wall ..
I suggest that it's best not to mix up certification and operational data and requirements - two different animals.