For the graph to represent your data most effectively, consider the following guidelines.
- The data must include one pair of columns of numeric or date/time data
- The columns must have the same number of rows.
- The sample size should be medium to large
- You should consider the sample size for the scatterplot and for the type of graph in the margins. Although there are no formal guidelines for the amount of data needed for a scatterplot, larger samples more clearly indicate patterns in the data. Histograms and boxplots work best when the sample size is at least 20. A dotplot is best when the sample size is less than approximately 50. If the sample size is 50 or greater, a dot may represent more than one observation.
- The sample data should be selected randomly
- In statistics, random samples are used to make generalizations, or inferences, about a population. If your data were not collected randomly, your results may not represent the population.