axis

this refers to a straight line, especially in the context of a graph. It constitutes a reference line that provides an indication of the size of the values of the data points. In a graph there is a minimum of two axes – a horizontal and a vertical axis. In statistics, one axis provides the values of the scores (most often the horizontal line) whereas the other axis is commonly an indication of the frequencies (in univariate statistical analyses) or another variable (in bivariate statistical analysis such as a scatterplot).

Generally speaking, an axis will start at zero and increase positively since most data in psychology and the social sciences only take positive values. It is only when we are dealing with extrapolations (e.g. in regression or factor analysis) that negative values come into play. The following need to be considered:

Figure A.4 Illustrating axes

• Try to label the axes clearly. In Figure A.4 the vertical axis (the one pointing up the page) is clearly labelled as Frequencies. The horizontal axis (the one pointing across the page) is clearly labelled Year.
• The intervals on the scale have to be carefully considered. Too many points on any of the axes and trends in the data can be obscured; too few points on the axes and numbers may be difficult to read.
• Think very carefully about the implications if the axes do not meet at zero on each scale. It may be appropriate to use another intersection point but in some circumstances doing so can be misleading.
• Although axes are usually presented as at right angles to each other, they can be at other angles to indicate that they are correlated. The only common statistical context in which this occurs is oblique rotation in factor analysis.

Axis can also refer to an axis of symmetry -the line which divides the two halves of a symmetrical distribution such as the normal distribution.