GraphPad Statistics Guide

Graphing tips: Frequency distributions

Graphing tips: Frequency distributions

Previous topic Next topic No expanding text in this topic  

Graphing tips: Frequency distributions

Previous topic Next topic JavaScript is required for expanding text JavaScript is required for the print function Mail us feedback on this topic!  

At the bottom of the frequency distribution analysis dialog, you can choose among several ways to graph the resulting data. These are all shown below, using 'frequency distribution' sample data set.

Graphs of frequency distributions

If you don't create a cumulative distribution, Prism gives you three choices illustrated below: XY graph with points, XY graph with spikes (bars). or a bar graph

The last two graphs look very similar, but the graph on the right is a bar graph, while the one in the middle is an XY graph plotting bars or spikes instead of symbols. The graph in the middle has X values so you can fit a Gaussian distribution to it. The graph on the right has no X values (just category names, which happen to be numbers), so it is not possible to fit a curve.

The term histogram is used inconsistently. We use the term to mean a graph of a frequency distribution which is usually a bar graph. Some people use the term histogram to refer to any bar graph, even those that don't plot frequency distributions.

Graphs of cumulative frequency distributions

If you choose a cumulative frequency distribution that tabulates the actual number of values (rather than fractions or percents), Prism can only create one kind of graph:

If you choose to tabulate the results as fractions or percentages, then Prism also offers you (from the bottom part of the Parameters dialog for frequency distributions) the choice of plotting on a probability axis. If your data were drawn from a Gaussian distribution, they will appear linear when the cumulative distribution is plotted on a probability axis. Prism uses standard values to label the Y axis, and you cannot adjust these. This graph is very similar to a Q-Q plot.