Blackwell Reference Online

Extract

Variability is one of the key characteristics of a set of scores. Measures of variability are used to provide an indication of the extent to which characteristics of people (e.g., income, education), events (e.g., attendance at concerts), or things (e.g., capacity of concert venues) differ from one another. Of the statistical indices of variability (e.g., range, semi-interquartile range), the variance and standard deviation are the only ones that reflect the exact value of each score. In this respect, they are analogous to the mean, the measure of centrality from which they are derived. For this reason, the variance and standard deviation are the measures of variability used in parametric statistics (e.g., analysis of variance (ANOVA), correlation, general linear modeling, hierarchical linear modeling, regression and regression analysis, structural equation modeling). The variance and standard deviation depict variability of a set of scores in terms of their distances from the mean, or deviation scores. A deviation score is computed by subtracting the mean of the set of scores from the score. Therefore, like the mean, deviation scores can only be computed for interval scales (e.g., Fahrenheit or Celsius temperatures) or ratio scales (e.g., height, weight, Kelvin temperatures), in which differences between scores are meaningful and consistent. Because of the way that deviation scores ... log in or subscribe to read full text