Antonio Reis de Sá Junior1, Arthur Guerra de Andrade2, Laura Helena Andrade3, Clarice Gorenstein4, Yuan-Pang Wang5. 1. Department of Health Sciences, Federal University of Santa Catarina, Araranguá, SC, Brazil; Institute & Department of Psychiatry (LIM-23), University of São Paulo Medical School, São Paulo, SP, Brazil. 2. Institute & Department of Psychiatry (LIM-23), University of São Paulo Medical School, São Paulo, SP, Brazil; Department of Neuroscience, Medical School, ABC Foundation, Santo André, SP, Brazil. 3. Institute & Department of Psychiatry (LIM-23), University of São Paulo Medical School, São Paulo, SP, Brazil. 4. Institute & Department of Psychiatry (LIM-23), University of São Paulo Medical School, São Paulo, SP, Brazil; Department of Pharmacology, Institute of Biomedical Sciences, University of São Paulo, São Paulo, SP, Brazil. 5. Institute & Department of Psychiatry (LIM-23), University of São Paulo Medical School, São Paulo, SP, Brazil. Electronic address: gnap_inbox@hotmail.com.
Abstract
BACKGROUND: This study examines the response pattern of depressive symptoms in a nationwide student sample, through item analyses of a rating scale by both classical test theory (CTT) and item response theory (IRT). METHODS: The 21-item Beck Depression Inventory-II (BDI-II) was administered to 12,711 college students. First, the psychometric properties of the scale were described. Thereafter, the endorsement probability of depressive symptom in each scale item was analyzed through CTT and IRT. Graphical plots depicted the endorsement probability of scale items and intensity of depression. Three items of different difficulty level were compared through CTT and IRT approach. RESULTS: Four in five students reported the presence of depressive symptoms. The BDI-II items presented good reliability and were distributed along the symptomatic continuum of depression. Similarly, in both CTT and IRT approaches, the item 'changes in sleep' was easily endorsed, 'loss of interest' moderately and 'suicidal thoughts' hardly. Graphical representation of BDI-II of both methods showed much equivalence in terms of item discrimination and item difficulty. The item characteristic curve of the IRT method provided informative evaluation of item performance. LIMITATION: The inventory was applied only in college students. CONCLUSION: Depressive symptoms were frequent psychopathological manifestations among college students. The performance of the BDI-II items indicated convergent results from both methods of analysis. While the CTT was easy to understand and to apply, the IRT was more complex to understand and to implement. Comprehensive assessment of the functioning of each BDI-II item might be helpful in efficient detection of depressive conditions in college students.
BACKGROUND: This study examines the response pattern of depressive symptoms in a nationwide student sample, through item analyses of a rating scale by both classical test theory (CTT) and item response theory (IRT). METHODS: The 21-item Beck Depression Inventory-II (BDI-II) was administered to 12,711 college students. First, the psychometric properties of the scale were described. Thereafter, the endorsement probability of depressive symptom in each scale item was analyzed through CTT and IRT. Graphical plots depicted the endorsement probability of scale items and intensity of depression. Three items of different difficulty level were compared through CTT and IRT approach. RESULTS: Four in five students reported the presence of depressive symptoms. The BDI-II items presented good reliability and were distributed along the symptomatic continuum of depression. Similarly, in both CTT and IRT approaches, the item 'changes in sleep' was easily endorsed, 'loss of interest' moderately and 'suicidal thoughts' hardly. Graphical representation of BDI-II of both methods showed much equivalence in terms of item discrimination and item difficulty. The item characteristic curve of the IRT method provided informative evaluation of item performance. LIMITATION: The inventory was applied only in college students. CONCLUSION:Depressive symptoms were frequent psychopathological manifestations among college students. The performance of the BDI-II items indicated convergent results from both methods of analysis. While the CTT was easy to understand and to apply, the IRT was more complex to understand and to implement. Comprehensive assessment of the functioning of each BDI-II item might be helpful in efficient detection of depressive conditions in college students.