Task 1. Selection of valid players’ responses for subsequent analysis.
6573 successions of responses(patterns) to GWBS questionnaire items were obtained. To evaluate the level of players’ emotional well-being in connection with other variables it is necessary at first to determine whether it is possible to use the set of questionnaire items as a psychodiagnostics method. For this, the players’ responses should be subjected to psychometric analysis, and it should be determined whether it is possible to build a measuring scale for evaluation of emotional well-being level on the basis of the obtained responses. When building a scale, we should be sure that the players’ responses to the questionnaire items really reflected those emotional peculiarities that corresponded to the contents of the items, i.e. the responses to the items should be valid from the point of view of a player’s behavior in relation to the items. This parameter is called “procedure validity”. In other words, only procedure valid responses should be subject to psychometrical processing for the purpose of building of an emotional well-being scale. Procedure valid answers are the answers not determined by psychological peculiarities of a player corresponding to the contents of the items. So, the first task was to select procedure valid answers. It can be done manually by comparing the responses to 22 items of each player out of 6573 and detecting the inconsistent responses in the response patterns. This task is too time and resource consuming. That’s why the process of selection of valid patterns was automated on the basis of the index of quality of players’ responses. For this Rasch metric system was applied. All the items were evaluated as one general scale of emotional well-being. In МСР the index of the quality of answers was calculated foreach player. It is formed based on the difference between the predicted and the actual probability of responses to the questionnaire items, i. e. on so called remainders. The smaller is the remainder, the more qualitative and valid is the pattern of the players’ responses. The range of values of the quality index varies from 0 to 4. The less is the index value, the more predictable and homogeneous is the response pattern, and the less is the dispersion in the players’ answers. From the point of view of meaningful validity of the questionnaire items the scale of emotional well-being is rather homogenous. That’s why the procedure valid responses should have low index values and be rather homogenous. The upper noise boundary is 1.3. Too predictable answers are also invalid. During the selection process 100 highly predictable patterns were removed, including null, identical and low-dispersive answers as well as 1739 responses with the high noise level, while the part of relatively “noisy” patterns were left in the sample. 4735 patterns of responses were remained for subsequent analysis, i.e. 70% of the data from the total sample.
Task 2. Building of a scale of emotional well-being measurement and psychometrical parameters evaluation.
The selected relatively valid players’ responses further were subject to psychometrical processing with building of emotional well-being measurement scale. During the process of scale building the validity of each questionnaire item was evaluated. All the items showed acceptable validity. The two items (No.2 and 15) had the worst validity indicators.The item No. 2 refers to concern about serious physical disease, the item No.15describes the state of cheerfulness and vivacity. The scale of emotional well-being has excellent reliability (0,95). The structural validity is characterized by moderate heterogeneity. Two components can be distinguished in it. The first component was formed by the items 1, 4, 6, 9, 10, 12, 15, 16, 19 and 21. These items meaningfully describe the state of well-being from the positive side and characterize the state of enthusiasm, activity and vigor. The second component includes the items 3, 5, 8, 11, 13, 14, 17, 18 and 22.These items describe the symptoms of tension, anxiety, depressiveness, stress and nervousness.
Scale calibration was made. The scale evaluates 12 levels of expressiveness of emotional well-being. This is a good indicator of discriminative characteristics of the method. The table below shows the total scores and interim measures in logits with standard errors. The levels of expressiveness were calculated with the help of logits and errors. The levels interpretation in also presented in the table. Now interpretation can be presented to the players on the basis of the new criteria of interpretation.
To learn more about the study results, please check the full article - click here.