Random Error or Experimental Error?

[#1]

Please, could you help me by clearing this doubt?
When analyzing the validity of the results using the residual, which of the errors is referred to: experimental error or random error?
Thanks in advance.

[#2]

I'll try my best to answer...but a specific example of your analysis would be helpful. Terminology can be tricky. Experimental error, from an analysis point of view, is the quantity expressed by the mean square for error resulting from the specific model you have fit to a given set of data. Random error is generally the mean square for error as a result of replicate runs in a given set of data. Random error will be a component of experimental error. Most commonly encountered with some manner of lack of fit test. Both can be visualized by examining residual plots in one form or another. 

[#3]

Yes, the terminology can be challenging.  Use of the terms often requires context.  

It is often thought that error is either random (unpredictable, chance) or systematic (assignable).  

When running an experiment, you want to compare the effects of the factors in your experiment (e.g., your model terms) to an estimate of the true error in the system (which is likely a combination of both random and systematic errors).  This estimate is often referred to as the experimental error.  How that error is estimated (randomized replicates, et. al.) and how representative of the true error in the system are very important considerations for the analysis and interpretation of your experimental results.  If, for example, the errors associated with your experiment are much less than typical error in the system (e.g., future conditions), you likely will commit an alpha error.  If the errors in your experiment are much greater than typical, then you commit the beta error.

Back to your question, what errors are represented in your residuals are a function of how the experiment was conducted (e.g., inference space) and the model you are analyzing (e.g., what terms are in your model).