Statistics for measuring the efficiency of electronic information retrieval

As electronic information searching with different software strategies and interactive systems becomes the only way to access desired holdings, the need to measure the efficiency of individual search strategies becomes critical to the research process that drives our choice of systems and instruction. Formal experimental research depends on the choice and use of an appropriate statistical technique. The relationship, between two dichotomously categorized variables distributed on an underlying continuous distribution is a tetrachoric correlation. This statistic was described at least as long ago as the turn of the century (Pearson, 1901, Philosophical Transactions of the Royal Society of London, 195A, pp. 1–47). It is a special case of a Pearson r and may be considered as an estimate of the product moment correlation (Guilford & Fruchter, 1978, Fundamental Statistics in Psychology and Education, 6th ed.). There are a number of ways that a researcher could use this statistic in a research design to measure information searching efficiency. The use of the tetrachoric correlation for experimental research and hypothesis testing of the efficiency of computerized information searching is described and illustrated. © 1996 John Wiley & Sons, Inc.