Introduction

Most of the experiments which you have done in the past have consisted of individual trials, each with its own discrete set of numerical data.   In those cases, a simple numerical average of the results of several identical trials was often the best tool to reduce the effects of many sources of experimental error.

As we study the equations of motion however, we often deal with experimental trails which produce a large number of related data points.  During the course of a one-second free fall for instance, we may measure the location of  an object at forty or fifty different times.  We could treat each of those distance-time measurements as a separate “trial,” comparing each individual distance value to the one predicted by our theory at that time.  However, that approach is not only time consuming, but it overlooks the fact that each data pair is a part of a larger, overarching relationship.  It is the characteristics of that relationship, not the “accuracy” of the individual data pairs, which the experiment was designed to explore.

The best tools for examining the overall patterns found in large data sets are the data plot and the statistical regression.
 

 Return to Main Data Analysis Page