Reproducibility depends on many factors such as instrument, operator and etc. However, from my own experience, interface connection needs special attention. For instance, if you apply silver paste and seal edges of your working electrode, significant lower resistance will be obtained, and importantly more more reproducible results! Don't forget the interface! They are crucially important! 

This depends very much on the tested material and solution. A very passive metal without any pitting corrosion will give very reproducible results in potentiodynamic polarization (PDP). Also a metal that has a very good, non-defective, coating will give reproducible results in EIS. Similarly, a very active metal surface will give reproducible results (early breakdowns, low pitting potential in PDP, low resistance in EIS). But everything in between can have a very high uncertainty and needs therefore many repetitions. If you are for example testing pitting corrosion by PDP for a stainless steel with a relatively low number of inclusions, different polished surfaces of this material will result in different values of pitting potential, simply due to a low/medium probability of having a very pitting-susceptible inclusion present on the surface.

From my experience, electrochemical tests can be reproduced but depends on some variables as my colleagues have mentioned. Even as far as external factors such as atmospheric temperature variation. But, yes, if all things been equal, results can be reproduced.

Reproducible data is crucial for research. On my research, I always folllow the same procedure of sample preparation and experimental setup. EIS is a techniques track the interface changes. When you have no reproducibility EIS will tell help in this matter. Please keep in mind that the experimental and human error. When this is minimum, you can compare the data. Make sure you search for reproducibility and not for the same values. 

After experiences with measurement of corrosion properties of metallic powder immobilized on PIGE working electrode, after several hours of training, I must say that by potentiodynamic polarization You can achieve nicely reproducible results. But in this case, operator factor play a huge role ... 

I agree with all of them. I can only add that EIS experiments are very sensitive. You should do it away from any possible noise source, such as, any strong electromagnetic field. I suggest you to repeat the experiments with the same environmental conditions.

Corrosion tests are specially difficult to reproduce but, as all experiments, this goal can be achieved if all variables are accurately controlled in the lab and the human factor is also reliable...