The method of calculation, the model used and the assumptions that went into the model will also have an effect. Just because you have some calculated values, it doesn't mean they are correct, it only means that they are correct for a certain fixed set of conditions.
Also, test methods and conditions at the time of test are important. Getting the voltage probes in the right places can make a big difference! And "finger trouble" is all too frequent - wrong connections, misread scales, transcription errors. Also make reasonable allowance for instrument error, accuracy, precision, repeatability etc. YOu won't get meaningful 4 decimal place results if one of your measurements is only accurate to 1 significant digit!
I always try to get our commisssioning people to measure line impedances at commissioning time, and the results are always different from the model. When they are a lot different, I assume human error and try to rectify the problem.
Bung