INTRODUCTION
Technologies (computers, cell phones, tablets, among others. (CCPT&O)) are in continuous and rapid evolution and interact increasingly with our lives.^{1} Improvement is so rapid that very few of us are prepared to absorb or follow this progression. Abusive use is further intensified by the use of the internet, social networks, applications and everything else offered to us.
New technologies CCPT&O have promoted changes^{2} in habits, behaviors, personal and social relationships. Now and in the foreseeable future it seems impossible to avoid all the ensuing effects (benefits and losses) resulting from this interactivity.^{2} It is thus of paramount importance to investigate and understand such changes in all possible contexts.
The “normal” use of technologies is one that allows us to take advantage of personal growth, work, social relationships, among others.^{3} Daily prolonged use does not constitute, per se, a pathological dependence. Pathological dependence is characterized by a sufficient level of inadequate use and must be accompanied by a history of symptoms in order to be diagnosed.^{3}
The purpose of the construction of validated scales in the area of digital dependence^{4} is to provide researchers with appropriate instruments for carrying out specific studies. We intend to perfect and train health professionals to deal with this new demand for help, namely the diagnosis and treatment of such “technology-dependent” individuals who increasingly seek this kind of support.
The objective of this study is to validate a scale to evaluate the abusive use of technologies in general through a TAUS in the daily life of individuals.
MATERIALS AND METHODS
At the time of writing, there are no specific scales for the quantitative evaluation of certain forms of abuse of digital technologies. To produce and validate any such scale, it is necessary to develop its content rigorously aligned with the subject and the objectives of the end-product. This task must be undertaken by experts in the field: then it must be tested on volunteers and the results statistically analyzed for validity. There is no consensus to define the number of specialists who should participate in the validation of a scale; this is left to the discretion and accessibility of the researcher: the greater the number of specialists, the greater the disagreement; conversely, the smaller the number (less than 3) the greater the risk of the agreement being 100%.
Accordingly, the production, validation and testing of this scale to evaluate abuse of digital technologies was carried out in 5 phases.
Construction of an initial questionnaire scale; six specialists trained in the area of digital dependence^{4} were selected. Based on published studies,^{5}^{,}^{6}^{,}^{7} they constructed a scale with 20 questions.
Evaluation of the questions by a second group of six similarly trained specialists, who analyzed the content regarding presentation, clarity, relevance and understanding. Thus, a preliminary validation was provided.
Application of the scale to 200 volunteers, divided into two groups: MAIN Group, including 100 participants with presumed abuse of technologies (CCPT&O); CONTROL Group including 100 participants with no presumed abuse of technologies (CCPT&O). For inclusion in the Main or Control groups, volunteers were previously submitted to the Internet Addiction Test (IAT) scale^{8} The Main Group included volunteers with IAT scores ≥ 50, whereas the Control Group included IAT scores < 50.
Statistical analysis and evaluation of the results.
Preparation of the validated final version.
The 200 volunteers participating in the research were asked to insert values opposite each question, as follows: Never/Rarely (0 points); Often (1 point), Always (2 points). The final sum of the results obtained ranked responders as follows. no dependence - 0 to 10 points; mild dependence - 11 to 20 points; moderate dependence - 21 to 30 points; severe dependence - 31 to 40 points. Orientations referring to each range of points was offered.
Sample. Volunteers included in the TAUS were (i) patients seen at our facility with complaints of abuse symptoms and prolonged daily use of technologies (CCPT&O). (ii) accompanying persons (iii) students, employees, any persons who agreed to participate. All were randomly recruited through posters at the institution, verbal communication from person to person and on social networks.
Factor analysis was used for the orthogonal model. The method used was Principal Components based on Spearman’s correlation matrix. For data analysis we used the R statistical program, version 3.4.2. ^{9}and packages “dplyr” .^{10} “psy” .^{11} “paran” ^{12} into R.
Inclusion Criteria. Participants should be between the ages of 17 and 65 and have a cell phone, tablet, computer, etc. with or without internet access.
Exclusion Criteria. illiterate candidates and persons with some kind of mental impairment that would prevent them from using technologies.
We discarded 5 Main and 10 Control group participants. Discarded volunteers presented incomplete questionnaires, discontinued participation or lack of accompanying persons when minors. The included results were entered into a database for statistical analysis.
RESULTS
Table 1 shows the results of the demographic statistics (age group, gender, degree of education) of the sample. For each characteristic, the absolute number of elements with the characteristic and the proportion within its group are displayed. Demographic data were collected for statistical purposes and not considered in the statistical evaluation.
Gender | ||||||
---|---|---|---|---|---|---|
Male | Female | |||||
Control | 28 (31.1%) | 62 (68.9%) | ||||
Main | 34 (36.2%) | 60 (63.8%) | ||||
Age range | ||||||
15-25 | 26-36 | 37-47 | 48-58 | 59-69 | ||
Control | 29 (32.2%) | 23 (25.6%) | 11 (12.2%) | 11 (12.2%) | 16 (17.8%) | |
Main | 44 (46.8%) | 23 (24.5%) | 20 (21.3%) | 5 (5.3%) | 2 (2.1%) | |
Education | ||||||
Middle | College | Graduate | Master | Doctoral | NI | |
Control | 21 (23.3%) | 26 (28.9%) | 37 (41.1%) | 2 (2.2%) | 3 (3.3%) | 1 (1.1%) |
Main | 53 (56.4%) | 26 (27.7%) | 9 (9.6%) | 5 (5.3%) | 0 (0%) | 1 (1.1%) |
NI = Not informed
Scores for the 20 original question scale. The mean ± standard deviation score for the Control group was 12.71 ± 8.42, while the corresponding value for the Main group was 19.47 ± 7.27 The t-test of means between the two groups produced a p-value < 0.001 (t-statistic = 5.820); this indicates a significantly higher level of damage in the Main group vs. the Control group. This difference ratifies, prima facie, the characteristics of the groups, mainly dependence in the main group and little or no dependence in the control group.
Factor analysis. The first test performed was the Bartlett sphericity test to verify if the variables are correlated with each other. In this test, the null hypothesis is that the correlation matrix, based on Spearman’s correlation, is equal to the identity matrix. For the data set, a statistic equal to 1806.758 and a p-value <0.001 was found, implying that the covariance matrix is not equal to the identity.
The next criterion used to verify the adequacy of the factor analysis was the Kaiser-Meyer-Olkin (KMO) criterion. Its value found was equal to 0.877; values above 0.8 are considered good.^{13} Table 2 DISPLAYS the Measurement of Sampling Adequacy (MSA) indices for each of the variables.
TAUS.1 | TAUS.2 | TAUS.3 | TAUS.4 | TAUS.5 |
0.819 | 0.811 | 0.922 | 0.888 | 0.925 |
TAUS.6 | TAUS.7 | TAUS.8 | TAUS.9 | TAUS.10 |
0.913 | 0.892 | 0.864 | 0.910 | 0.827 |
TAUS.11 | TAUS.12 | TAUS.13 | TAUS.14 | TAUS.15 |
0.756 | 0.851 | 0.897 | 0.893 | 0.870 |
TAUS.16 | TAUS.17 | TAUS.18 | TAUS.19 | TAUS.20 |
0.875 | 0.914 | 0.874 | 0.919 | 0.893 |
Due to the results for the Bartlett test and the KMO criterion, we considered it appropriate to carry out the factorial analysis for the scale.
The next step was to check the factor loads to determine the number of relevant factors. We used 3 criteria: Factorial Load, Screeplot and Parallel Analysis. Table 3 shows the factorial loads, using the Principal Components as method:
PC1 | PC2 | PC3 | PC4 | PC5 | |
Standard deviation | 2.762 | 1.328 | 1.205 | 1.114 | 1.059 |
Variance proportion | 0.381 | 0.088 | 0.073 | 0.062 | 0.056 |
Cumulative proportion | 0.381 | 0.470 | 0.542 | 0.604 | 0.660 |
PC6 | PC7 | PC8 | PC9 | PC10 | |
Standard deviation | 0.942 | 0.904 | 0.874 | 0.775 | 0.732 |
Variance proportion | 0.044 | 0.041 | 0.038 | 0.030 | 0.027 |
Cumulative proportion | 0.705 | 0.746 | 0.784 | 0.814 | 0.841 |
PC11 | PC12 | PC13 | PC14 | PC15 | |
Standard deviation | 0.684 | 0.668 | 0.620 | 0.600 | 0.573 |
Variance proportion | 0.023 | 0.022 | 0.019 | 0.018 | 0.016 |
Cumulative proportion | 0.864 | 0.886 | 0.906 | 0.924 | 0.940 |
PC16 | PC17 | PC18 | PC19 | PC20 | |
Standard deviation | 0.545 | 0.514 | 0.511 | 0.442 | 0.428 |
Variance proportion | 0.015 | 0.013 | 0.013 | 0.010 | 0.009 |
Cumulative proportion | 0.955 | 0.968 | 0.981 | 0.991 | 1.000 |
PC = Principal Components
Factor loads with cumulative proportions above 0.9 are considered satisfactory.^{13} For this data set, we would have to use 13 factors, which in practice would not solve the problem of data reduction.
The Screeplot criterion of the correlation matrix was tested: in this test we eliminate the factors related to Eigenvalues > 1. Figure 1 presents this criterion:
By this criterion, we should use 5 factors, and in this case, the commonalities of the variables are presented in Table 4.
TAUS.1 | TAUS.2 | TAUS.3 | TAUS.4 | TAUS.5 |
0.750 | 0.701 | 0.617 | 0.503 | 0.651 |
TAUS.6 | TAUS.7 | TAUS.8 | TAUS.9 | TAUS.10 |
0.623 | 0.700 | 0.744 | 0.679 | 0.702 |
TAUS.11 | TAUS.12 | TAUS.13 | TAUS.14 | TAUS.15 |
0.846 | 0.801 | 0.613 | 0.569 | 0.683 |
TAUS.16 | TAUS.17 | TAUS.18 | TAUS.19 | TAUS.20 |
0.720 | 0.637 | 0.485 | 0.583 | 0.598 |
Analyzing the commonalities, it was observed that the question 18 (How often do you usually have the feeling of being accompanied when you are using the technologies (CCPT&O) should be excluded from the initial scale with 20 questions because of a commonality less than 0.5. Therefore, we decided to use the Screeplot as a basis for recompose the questionnaire scale
The third criterion used to find the number of factors was the Parallel Analysis. By this criterion, the number of factors found was equal to 3, and its commonalities are presented in Table 5.
TAUS.1 | TAUS.2 | TAUS.3 | TAUS.4 | TAUS.5 |
0.610 | 0.481 | 0.585 | 0.472 | 0.642 |
TAUS.6 | TAUS.7 | TAUS.8 | TAUS.9 | TAUS.10 |
0.592 | 0.611 | 0.586 | 0.649 | 0.684 |
TAUS.11 | TAUS.12 | TAUS.13 | TAUS.14 | TAUS.15 |
0.652 | 0.680 | 0.462 | 0.458 | 0.473 |
TAUS.16 | TAUS.17 | TAUS.18 | TAUS.19 | TAUS.20 |
0.597 | 0.591 | 0.195 | 0.418 | 0.405 |
For the number of factors equal to 3, we must remove questions 2, 4, 13, 14, 15, 18, 19 and 20 because they have a commonality less than 0.5, what would eliminate important issues as the goal of the research.
The last step of the study was to calculate Cronbach’s alpha,^{13} in order to measure the internal consistency of the scale. The value found was 0.910, which is considered good.^{13}
DISCUSSION
Samples from the so-called Main and Control Groups were randomly formed without pre-established concerns about the quantitative distribution of male and female subjects, as well as the distribution of age groups and education level. This resulted in a considerable level of variability, which reinforces the random characteristic of a data collection ensuring realistic results.
Thus, the significant differences in IAT scores for the “Main” and “Control” groups, formed by dependents and non-dependents of digital technologies respectively, were ratified by the Main group with an IAT score 60% higher than that of the Control group, securing the quality of the results obtained in the collection. A final validated scale was constructed, with the purpose of being used in clinical practice which fully met what was proposed, namely the evaluation of abusive use of technologies (Computer, cell phone, tablet, among others).
Factorial Analysis was then performed based on the results of the Bartlett sphericity tests that confirmed the correlation between the variables that constitute the questionnaire, in addition to the very satisfactory KMO index, equal to 0.877, a value above the statistically reference value of 0.8.^{13} The first of three criteria, Factorial Loads, signaled a high value of 13 factors for an initial scale with 20 questions, which made it a useless criterion.
The second criterion was the Screeplot which, despite pointing to 5 factors, suggested the withdrawal of only one of the 20 questions from the initial scale mentioned above. The indication of this withdrawal was due to the fact that the commonality of this question, extracted by the Screeplot Criterion was 0.485, thus below 0.5, which is a criterion for withdrawing a question from a scale. Withdrawing just one question leads to an excellent Cronbach’s Alpha of 0.910, corroborating with judicious scale construction.
As the third and last criterion, the Parallel Analysis pointed to only 3 factors, but with the suggestion to withdraw 8 questions out of a total of 20, which would compromise the consistency of the scale.
Thus, the Screeplot Criterion was used to recompose the TAUS Scale, now with 19 questions, pointing to a positive and consistent adequacy to obtain data on dependence on digital technologies.^{14}^{-}^{16}
As a limitation of the study, we came across an absence of specific validated instruments capable of investigating behavior using technologies on a day-to-day basis, which might have helped us in the preparation of the present scale. Future studies are recommended so that we can refine the research in all areas and especially on the subject of digital dependence.
CONCLUSION
The results obtained provided a validated version for the TAUS, with 19 questions appropriate to clinical and research contexts for clarity and accuracy.
Statistical results showed that the issues of the final version of the scale presented alignment among them, qualifying it as positive to measure the abusive use of technologies (CCPT&O). The final version of TAUS can be used, whenever it is necessary to carry out research projects related to the subject digital dependency.
With the results of new studies using the TAUS, we can better observe the clinical, cognitive-behavioral, social and professional effects resulting from the impact caused by the interference of the technologies (CCPT&O) in the daily life of individuals. We can also expand scientific knowledge, improve outpatient care and develop forms of conscious use of technologies such as prevention and reduction of physical and/or psychological damage in the population.
We recommend that the study be replicated in a larger sample and representative of the target population.