Experimental data from: "A Study on the Suitability of Visual Languages for Non-Expert Robot Programmers"
Once the three courses for the three learning scenarios - C# OOP programming, Sphero Edu visual programming and VEDILS authoring tool - were taught, the three student groups were asked to indicate using a scale between one and four - to avoid the selection of neutral options - their perception of the clarity and the interest of the exposition (CL and IT indicators), as well as the time spent studying the course contents (ST indicator).
All the students also undertook the same three programming exercises, which were marked using the logical sense of the algorithms and the implementation degree of the functionalities required by the statements as evaluation criteria, and a scale between 0 and 2 (0 - Poor: The student did not resolve the exercise at all and the code has no logical sense, 1 - Average: The student only solved part of the exercise and part of the code has logical sense, 2 - Good: The student solved the whole exercise and all the code has logical sense) in order to reduce the subjectivity in the marking process.
The maximum time available for completing the three exercises was ninety minutes. This time could be freely distributed among the exercises.
Finally, the student groups were asked to indicate using a scale between one and four their perception of the difficulty level (DF1, DF2 and DF3 indicators) of the programming exercises, as well as the perceived enjoyment level in relation to the use of the development environment (EJ indicator).
This dataset can be used as input files for the IBM SPSS statistical software in order to obtain the results showed in the article entitled Experimental data from: "A Study on the Suitability of Visual Languages for Non-Expert Robot Programmers” (Tables I to V).
The meaning of the sample data is explained in the text of the article and in the abstract of this dataset. From each text file, the SPSS software can be used to calculate the mean value of the indicators and their standard deviations showed in Tables I and II.
Note that before the program can determine the existence of significant differences between the data from the experimental groups (see Tables III, IV and V), the contents of the three text files must be combined in pairs, so that three new SPSS files must be created, each one containing the data from two groups. Each new file will contain data from two different learning scenarios.
Finally, a binary integer variable must be added to each new file in order to identify the observations from each of the groups.