So far we have considered only the presentation of results for a single variable. In practice, most studies are concerned with the association between two or more variables.

The appropriate statistical method depends on the situation.

- Two variables measured on the same subject/object (e.g. peoples' heights and weights) - are these related? Usually use correlation or regression;
- One variable measured at selected values of another variable - (e.g. children's heights at different ages) Show simple trend - usually use regression;
- One variable measured at equally spaced times (or distances) but successive values highly correlated or pattern complicated (e.g. stockmarket prices) Use methods of time series analysis.

A scatterplot is constructed by marking the scales of the two variables, (x, y) along horizontal and vertical axes. Each pair of measurements is plotted with a cross at the point using the measurements as coordinates. The information from a third discrete variable can be incorporated into the plot by using different plotting symbols for each category of the third variable.

The figure below shows a scatterplot of income against age for some hypothetical data. The number 2 is used to indicate a point where two data values coincide.

Further information on sex is incorporated into this scatterplot by using diamonds for males and squares for females.

Student | Height(cm) | weight(kg) |
---|---|---|

1 | 167 | 60 |

2 | 170 | 64 |

3 | 160 | 57 |

4 | 152 | 46 |

5 | 157 | 55 |

6 | 160 | 50 |

First step - draw a **scatter plot**:

The MINITAB command for a scatter plot is:

## Progress check |

... Previous page | Next page ... |