Language Testing in Asia (Oct 2022)

A structural equation modeling of English tests’ social and educational consequences: exploring target, leverage, risk, and critical variables

  • Mahmood Khosravani,
  • Morteza Rostamian,
  • Hamid Ashraf,
  • Khalil Motallebzade

DOI
https://doi.org/10.1186/s40468-022-00177-2
Journal volume & issue
Vol. 12, no. 1
pp. 1 – 19

Abstract

Read online

Abstract Though the emergence of standardized testing promoted classical approaches such as summative achievement tests, it brought about serious criticisms regarding the stakes of these tests and their consequences on different dimensions of education. Exploring these consequences attracted researchers’ attention in education. A mass of studies are conducted, but due to the dynamic and intricate nature of the consequences, they suffer from some severe deficiencies, mainly in methodology and epistemology. Thus, the current research adopted an innovative approach to employ structural equation modeling in SPSS and matrix of crossed impact multiplications in MicMac to explore factors of test consequences of English module of Iranian universities entrance exam. Four groups of participants were selected through convenience sampling from different Iranian stakeholders for scale development, reliability estimation, exploratory factor analysis (EFA), and crossed impacts analysis, respectively. Reliability calculation was carried out (r = 0.89), and assumption of EFA for factorability (KMO = 0.796 and p = 0.001) was met. EFA was run, and 19 variables (factor loading ≥ 0.60) in three major factors were identified. Also, crossed impact analysis showed that these factors fall in an unstable system of effects, inside input factors (critical and environmental), bidimensional variables (risk and target), output variables, and leverage variables. Integrating structural equation modeling and structural interpretations resulted in development of a validated model and uncovered the nature of the variables within it. These information can feed subsequent interpretations, including policy makings and future studies for educational and management purposes. The findings have educational and statistical implications for different stakeholders.

Keywords