ABSTRACT
Survey interpretation often involves the effective use of advanced statistical techniques-example, multiple regression, item response analysis, analysis of variance, factor analysis, reliability analysis, and even structural equation modelling-but these sophisticated tools are not required to produce an interesting and compelling set of findings. Once the survey practitioner and/or survey team has a thorough understanding of the descriptive statistics and item-level means, the next stage in the data interpretation process is a conceptual level analysis. Besides the use of existing models or frameworks, the other main option for a level conceptual survey analysis, albeit less often employed, is to generate a new model based on the relationships inherent in the dataset at hand. Most practitioners will concentrate on the highest and lowest items, but if a midpoint cut-off value is needed, the best method for determining one is to compute an overall average rating across all survey items.
