Histograms show a distribution of marks without identifying individual values. They contain less visual 'noise', especially for large data sets.
Histogram class widths should be adjusted to give the smoothest possible outline.
In a histogram, the area above any part of the axis is equal to the proportion of data values in that region.
Histograms bins usually all have the same width. However the effect of combining two bins or splitting a bin helps explain the fundamental relationship between histogram area and proportion.
Joining the tops of adjacent histograms gives a frequency polygon.
A histogram is usually drawn from a frequency table. Histograms can be drawn in Excel, but not easily.
The histograms of large data sets tend to be fairly smooth. A smooth curve called a probability density function may be used to summarise it.
One particular class of curves called the family of normal distributions is often used as a 'model' for mark data. Two 'parameters' can be adjusted to match the curve to actual data.
Normal distribution curves have similar properties to histograms. In particular, the proportion of values in any range equals the area under the curve.
Mark data are usually counts and are therefore discrete.
If discrete data cover a large range of counts, a histogram can be used. If there are only a small number of possible values, a bar chart is better.
Some assessment information is categorical rather than numerical. Bar charts can also be used to display categorical data.
These are alternative displays of discrete and categorical data.
Information is given about drawing bar and pie charts in Excel.