Derivation of cut-off value for a 10 item opinion based ordinal survey questionnaire

• Published in: International Journal Of Community Medicine And Public Health Vol. 5; no. 3; p. 1030
• Main Authors: Arumugam, Balaji, E., Suganya, Nagalingam, Saranya
• Format: Journal Article
• Language: English
• Published: 24.02.2018
• ISSN: 2394-6032; 2394-6040
• DOI: 10.18203/2394-6040.ijcmph20180756

Background: Questionnaire is the most common tool for data collection in most of the study designs. In spite of the advantage of using a standardized questionnaire for generalising the study findings, the major disadvantage is that all the population is never the same and its similarity exists only with the presence of heterogeneity, hence, any universal cut-off value cannot be used for the diverse population. This makes it clear that determination of cut-off value for any content validated questionnaire to the specified study population is essential, in order to make the study tool more effective.Methods: This study was done to determine the cut-off value of psychometric scale of selfie addiction, which was content validated. A detailed mathematical model was used to determine the cut off value. Item analysis was done. Discrimination index, weightage of each item and correction factor was calculated to determine the cut-off value.Results: The total weighted score, total raw score and correction factor are 28012.62 and 31,046, 0.9 respectively. The total adjusted scale cut-off is 30.43 (rounded as 30). The cut-off value based on crude mid-value is 25 and the cut off value determined for the standardised population is 21.Study population with cut-off value of >30 are considered to be a selfie addict, and those who have obtained a total score ≤30 are considered to be normal (non-selfie addict).Conclusions: Present study is one of its kind, in determining the cut-off value for a content validated psychometric scale without any gold standard. The above derived cut-off value of 30 for the psychometric scale of selfie addiction is valid for the specified population, as the Cronbach’s alpha, discrimination index and the correction factor is above 0.75.