THE IMPACT OF THEORETICAL CONSTRUCT VALIDITY ON CAUSAL STRUCTURE MODEL WITHIN BANKING INDUSTRY – a study of applied multivariate data analysis – ŞTEFAN C. LIŢĂ * Abstract The current analysis is based on Hopkins’ data (1997) used to investigated the causal relations between strategic planning and financial performance within banking industry. The correlation matrix for the seven items developed to capture the construct of strategic planning intensity was used to conduct a confirmatory factor analysis and then, a structural model was fitted to data in order to replicate the model used in the original paper. The results of confirmatory factor analysis showed that: (a) the 7 original items do not measure the specified latent variable, but two separated factors that explain 49% of the variance, (b) the latent variable “strategic planning intensity” might be regarded as one-factor structure measured by 5 items (V1 – determining banks’ mission, V2 – developing major long term objectives, V5 – evaluating strategic options, V6 – implementing strategic options, V7 – controlling the implemented strategic option). Although the results of structural equation modelling were not very clear, the general impression was that strategic planning intensity could be indeed regarded as a mediator between institutional factors (managerial and organisational) and financial performance, because: (a) managerial factors have a strong direct effect (0,82) on strategic planning intensity, and also an indirect effect (0,41) on financial performance, mediated by planning intensity, (b) organisational factors have a direct effect on financial performance (0,77) and also an indirect effect (0,086) mediated by planning intensity. The paper ends with few recommendations regarding alternative strategies that might be used to improve the structural model. Cuvinte-cheie: analiză factorială confirmatorie, modelare prin ecuaţii structurale, planificare strategică, factori manageriali şi organizaţionali, performanţa financiară. Keywords: confirmatory factor analysis, structural equation modelling, strategic planning, managerial and organisational factors, financial performance. 1. INTRODUCTION The purpose of this paper is to illustrate how to apply multivariate data analysis (especially confirmatory factor analysis – CFA and structural equation modelling – SEM) to investigate the factorial validity of a theoretical construct and the validity of a causal structure. * The paper was written in 2009 while I was affiliated to the Operational Research Group, Department of Management, London School of Economics and Political Science. Rev. Psih., vol. 57, nr. 3, p. 214–226, Bucureşti, iulie – septembrie 2011