Contents lists available at ScienceDirect Learning and Individual Differences journal homepage: www.elsevier.com/locate/lindif Internal structure of academic self-concept through the Self-Description Questionnaire II-Short (SDQII-S) Igor Esnaola a, ⁎ , Paula Elosua a , John Freeman b a UPV/EHU, Spain b Queen’s University, Canada ARTICLE INFO Keywords: Academic self-concept Adolescence SDQII-S Bifactor models Measurement invariance ABSTRACT Modelling academic self-concept through second-order factors or bifactor structures is an important issue with substantive and practical implications. To contribute to the discussion, this research analyzes the structure of academic self-concept through the Self-Description Questionnaire-II short-form (SDQII-S) and assesses its in- variance across gender and age in a sample of secondary students (N = 517). Results show a better fit for the incomplete bifactor model and the estimation of several statistical indices support the multidimensional nature of the data as well as the reliability of the domain scores. In terms of invariance regarding age and gender, partial invariance is concluded. Girls showed higher latent means in verbal self-concept, while boys had higher values in mathematics self-concept; as the student’s educational level increased, domain-specific and general factor means decreased. Practical implications associated with the use of domain scores are discussed. 1. Introduction Academic self-concept (ASC) is an important domain of general self- concept related to learners’ knowledge and perceptions about them- selves in overall academic domains (Wigfield & Karpathian, 1991). The substantial relationship between ASC and desirable educational out- comes, such as interest, persistence, coursework selection or achieve- ment, has been demonstrated (Arens, Yeung, Craven, & Hasselhorn, 2011; Parker, Marsh, Ciarrochi, Marshall, & Abduljabbar, 2014). Since Shavelson, Hubner, and Stanton’s seminal work (1976), the multi- dimensional nature of ASC is unquestioned (Byrne, 1996; Byrne & Gavin, 1996); however, the hierarchical structure is still under debate. 1.1. Models of academic self-concept Shavelson et al.’s initial proposal (1976) defined a general ASC with separate domain-specific factors related to mathematics, English, sci- ence or history (Model 1 in Fig. 1) but since the model failed to explain the pattern of correlations among the domain-specific factors, it was modified into the “Marsh/Shavelson model” (Marsh, 1990). This new model distinguished between general mathematics-related and general verbal-related self-concepts which are defined as factors that influence specific domains and subjects. Moreover, general ASC, represented as a first-order factor, was assumed to be simultaneously influenced by mathematics self-concept (MSC) and verbal self-concept (VSC) (Model 2 in Figure 1). More recently, and based on the bifactor model theory, Brunner, Lüdtke, and Trautwein (2008) proposed an incomplete nested-factor model, which combines a multifaceted and hierarchical structure. Ac- cording to the model, the factor representing general ASC influences all manifest measures of ASC as well as specific domain factors (Edwards & Bagozzi, 2000), and as the model is incomplete some items do not make a group factor (Model 3 in Fig. 1); ASC is uncorrelated with MSC and VSC, and the correlation between domain-specific factors is negative (Marsh & Hau, 2004). The validity of the bifactor model is supported by a cross-cultural investigation involving 15-year-old students from 26 countries (Brunner, Keller, Hornung, Reichert, & Martin, 2009), and by a study conducted with elementary school students in which writing and reading group factors were defined instead of a verbal domain (Schmidt et al., 2017). Most studies investigating the structure of ASC have used Self- Description Questionnaires (SDQ). The questionnaire has good psy- chometric properties (Leach, Henson, Odom, & Cagle, 2006) and it has been validated cross-linguistically (e.g., Chinese: Kong, 2000; French: Guérin, Marsh, & Famose, 2003; Spanish: Inglés et al., 2012). As the SDQII (designed for adolescents) may be too long (102 items; Marsh, Ellis, Parada, Richards, & Heubeck, 2005), the Self-Description Ques- tionnaire II short-form (SDQII-S) was developed (Marsh et al., 2005). SDQII was used by Brunner et al. (2009) to propose the bifactor structure of ASC. There is no doubt about the contribution of the cited https://doi.org/10.1016/j.lindif.2018.02.006 Received 13 October 2016; Received in revised form 27 January 2018; Accepted 5 February 2018 ⁎ Corresponding author. E-mail addresses: igor.esnaola@ehu.eus (I. Esnaola), paula.elosua@ehu.eus (P. Elosua), freemanj@queensu.ca (J. Freeman). Learning and Individual Differences 62 (2018) 174–179 1041-6080/ © 2018 Elsevier Inc. All rights reserved. T