For the graph to represent your data most effectively, consider the following guidelines.

- The sample size should be approximately 10 or greater for each group
- Usually, the larger the sample size, the smaller and more precise the confidence interval. If the confidence interval is too wide, try to collect a larger sample. If you have multiple Y variables, the sample size should be approximately 10 or greater for each Y variable.
- Groups should have approximately the same sample size
- Large differences in sample sizes between groups (or multiple Y variables) can affect the widths of the intervals, and may produce misleading results. If sample sizes for the groups (or multiple Y variables) are approximately the same, you can be more confident that differences in the widths of the intervals is primarily due to differences in variation.
- 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.