Split Ballot Split-ballot multitrait-multimethod designs' effect of response scale on assessment of emotional intelligence competencies

Most of the research on emotional and social competences has been done in English-speaking countries. Most of the survey measurement instrument of behavioural competencies use a five-point Likert scale, without considering that the response modality may affect the findings. Batista-Foguet, Saris, Boyatzis, Guillén & Serlavós (2009) have shown using a Split-Ballot design that an eleven-point response scale with mainly European MBA students, that the eleven-point scale lead to composites with greater reliability, validity and less invalidity, than the commonly used 5-point scale. A complete four-group Split_Ballot (SB-MTMM) design for two methods and their replications (exact repe¬titions of the same items are included in the design after sufficient time has elapsed to avoid memory effects) is used to compare both response modalities. On this paper, the SB-MTMM design allows to evaluate the occasion effects as well as measurement bias, reliability, and validity simultaneously, while reducing response burden a usual drawback of MTMM studies.