Bar Charts

Bar charts are easily read and quite easy to manipulate.  Simple bar chart display one variable at a time and are thus univariate displays.  A simple bar chart displays the frequencies of different values of a variable as bars.  The higher the bar, the more cases fall in that category.  Bar charts, therefore, work best when the number of values (and thus bars) is relatively small.  For example, I can usually describe religion very well with a simple bar chart.

The bar for Protestant is the highest, indicating that more cases named this religion than any other.  Protestant, therefore, is the mode.  High bars for Catholic and Jewish/Muslim indicate that large numbers of cases named these religions.  Smaller numbers named Hinduism, Buddhism, or Animism.  Because bars can be put in sequence, bar charts are particularly useful with ordinal data.

There are many forms of bar charts and some of these display two or more variables.  For example, I could show the mean income of cases in these different religions.

In this bar chart, the height of the bar indicates the mean income level in $100,000 increments.  As you can see, Buddhists enjoy the highest mean income level -- nearly $200,000.  All the other religious groups are between about $70,000 and $100,000.

SPSS instructions for a simple bar chart

SPSS instructions for a bar chart displaying more than one variable