Is “not applicable” option in Internet Addiction Questionnaire affect its results? An exploration of results using four different methods (Preprint)
BACKGROUND Internet addiction problem is becoming a significant public health concern nowadays, especially during Covid-19 pandemic. Internet Addiction Test (IAT) is one of the most popular questionnaires for evaluating the internet addiction level. Using the same questionnaire in the similar population, the prevalence of IA was ranging from 28.3% to 66.6% with possibly due to the Likert scale used. Different methods on treating “not applicable” option in the Likert scale used in IAT questionnaire might increase the variability in results, and this might not be aware especially when comparing the results among studies. OBJECTIVE This paper aims to compare the differences of results in terms of total scores and prevalence of IA on using four different methods on dealing with the “not applicable” option on the IAT questionnaire. METHODS A total of 417 undergraduate students were recruited from a public university in Malaysia to fill in the IAT questionnaire with six Likert scale that include the following options from “not applicable”, “rarely”, “occasionally”, “frequently”, “often” and “always”. The total score and prevalence of IA for four different methods (Method 1, Method 2, Method 3 and Method 4) on treating “not applicable” option were be compared. RESULTS Out of 20 questions in IAT questionnaire, there were 11 questions had more than 80% of answer rate but no question with 100% answer rate. Only 17.5% of respondents answered all the questions. The Cronbach alpha test showed that four methods had good reliability value (0.910-0.945). The mean ± standard deviation of total scores of IA for Method 1, Method 2, Method 3 and Method 4 were significantly different, with 49.15±15.60, 40.50±17.50, 44.31±15.11 and 48.15±15.08 respectively (P<0.001). Method 1 had significant highest prevalence of IA (43.4%), followed with Method 4 (40.0%), Method 2 (28.3%) and Method 3 (17.3%) (P<0.001). CONCLUSIONS Four methods on treating the “not applicable” option in the IAT questionnaire showed significant different results, with Method 1 showed significant highest mean total score and prevalence of IA compared to other methods. Researchers are advised to choose an appropriate method on treating the “not applicable” questions, and also be aware of this when comparing the results among different studies.