Over the past decade many colleges and universities have placed increased emphasis on having students develop statistical and data analysis skills in a range of disciplines. Some institutions now require that all students complete at least one course with a strong component of data analysis, whether the data are from chemical experiments, the census, or some other source. As chemists, one of our concerns should be to ensure that students view data analysis as an integral part of any quantitative experiment, and, as far as possible, do not treat this process as a black box. The authors of A Method of Visual Interactive Regression, a spreadsheet application, have developed a visual approach to linear least-squares curve fitting that drives home the idea of minimizing the sum of the squares of the deviations in order to find the best fit to a set of data that are being described by a linear relationship. For many students these visualizations are likely to persist a great deal longer than the mathematical derivations of the equation for the slope and the intercept. The visualizations will provide a useful connection between a set of equations and the buttons on a calculator or the insertion of a trendline in a spreadsheet. |