Bar graphs along with pie charts are the most common types of graphical representation for qualitative data. Like a pie chart, a bar graph breaks categorical data down by group. Unlike a pie chart, it represents these amounts by using bars of different lengths. While a pie chart most often reports the amount in each group as percentages, a bar graph uses either the number of data points in each group (also called frequency) or the precentage in each group (called relative frequency). ^[1,p.96,101] Bar charts are used for surveys where each person surveyed is allowed to choose more than one answer. The resulting data set would not sum to one, and a pie chart representation would not be possible.^[1,p.106]

The following tips help to evaluate a bar chart for statistical correctness:^{[1,p.102,106,337]}

• Making sure that categories of grouping for a numerical variable are equivalent.
• Making sure that the scale, starting point and range of the axis of the bar graph is an appropriate representation of the information.
• Making sure how many variables are shown in the bar graph, and which should sum to one.
• Considering the units being presented by the height of the bars and what the results mean in terms of those units.
• Checking total sample size if relative frequencies are given. Or dividing each bar by the total sample size to get percentages for easier comparison, when frequencies are given.
• Checking for overlapping boundaries of numerical groupings and clarifying how borderline values are treated.

A graph can be misleading through the choice of scale on the frequency/relative frequency axis (where the amount in each group is reported) and/or its starting value. The scale of a graph is the quantity used to represent each tick mark on the axis of the graph. The scale can make a big difference in terms of the way the graph or chart looks. Stretching the scale out or starting an axis at the highest possible number makes differences appear larger; squeezing down the scale or starting the axis at a lower value than needed makes differences appear smaller than they are.^{[1,p.39,42,104]}

The bars in a bar graph don't connect, unlike the histogram, since the bars represent distinct categories without particular order.^[1,p.108]