Conducting an experiment is related to providing certain experimental results, mostly numerical data. The chapter discusses the possibilities for adequate processing of such data, as well as the appropriate interpretation of results and assessments from the analysis conducted. In its main part, the processing of experimental data is based on conducted statistical analysis and graphical interpretation of the estimates. This defines the three main parts included in the chapter. In the first part, an approach is proposed for primary processing of experimental data (determination of definitional areas, individual criteria for evaluation, logical completeness of the evaluation and the need for preliminary filtering of the sample). The second part is dedicated to the basic processing of experimental data, where statistical approaches for this are discussed (basic statistics, correlation and regression analysis, logistic regression, variance analysis). In the last part, the basic methods are tabular and graphical interpretation of estimates from the analysis of experimental data.
Top1. Introduction
When conducting experiments using different approaches (modeling, monitoring, benchmark, etc.), information (registrations) is provided for the set of observable variables forming a vector of the studied observable parameters 
= <X1, X2, …, Xn>. Each parameter is considered as a random variable, and the recorded values – as its specific manifestation during the experiment. Conclusions are drawn based on the analysis of the accumulated information, which requires that the modeling data be highly informative. This is also related to the effectiveness of the experiments, which is determined by a properly defined plan. Guidelines for planning experiments are presented in the editorial (Curtis et al., 2022), the purpose of which is to specify the current requirements and good practices in this direction. It is a follow-up to two previous publications on the topic, with the article referenced here addressing new issues and emphasizing three aspects of design: randomization, blinded analysis, and balance of group sizes.
The statistical interpretation of experimental results is an important part of conducting research in any field, it is stated in (Lúcio & Sari, 2017). Of great importance is the correct determination of the main factors that may be subject to variability when conducting the experiments. To avoid this drawback and to reduce the experimental error, it is necessary to pay important attention to the planning of the experiment. The application of statistical tests for the analysis of experimental data should be based on their good knowledge and statistical validity for the specific field of study. The article suggests possible alternatives for reducing factorial variability due to uncontrolled effects in a given experiment with a recommendation for appropriate statistical analyses. The aim is to provide information for the proper planning and conduct of experiments and in the analysis and interpretation of experimental results.
An example of conducting statistical analysis of experimental data is presented in (Romansky, 2023). In this case, the main goal is to analyze the delay factors in accessing remote information resources in active learning processes. It is known that in the latter network communications and access to distributed resources in the network space are of primary importance. It is assumed that clients access different types and content of resources in the digital space, maintained and offered by different servers – general purpose and specialized. To investigate parameters (variable factors) of these communications between client and server, monitoring was conducted to collect empirical data used to form initial samples of values (registrations). These experimental data are subjected to statistical analysis to determine estimates of significant service parameters. Different servers supporting purely text objects and resources containing text and images were analyzed, and statistical estimates were determined for total time for dynamic resources and influence of transfer parameters at different types of resources. A summary of the obtained results is made in graphic form of formed functional dependences of the studied time parameters on the volume of available resources.
Modeling also provides data that can be further processed based on developed plan for statistical experimentation. Regarding the final phase for analyzing the model results, the plan should include the following main steps:
- •
Determining the meaningful meaning of the vector of model parameters.
- •
Planning of primary processing of model information.
- •
Planning the main processing of the model information.
- •
Analysis and interpretation of the obtained estimates.
While the first step is the object of decision already at the initial stage of organization of experiments (including modeling), the remaining three are related to the two main parts of the “Analysis and interpretation of model results” phase presented below (see Figure 10 in Chapter 3).