For the graph to represent your data most effectively, consider the following guidelines.
- Individual symbols are most useful when the sample size is less than approximately 50
- If the sample is too large, the individual data points on the plot may be too densely packed together and the distribution may be difficult to assess. To assess the shape of a sample distribution with more than 50 observations, consider using Boxplot or Histogram instead.
- Confidence interval bars are most useful when the sample size is approximately 20 or greater
- Usually, the larger the sample size, the smaller and more precise the confidence interval. If the confidence interval is too large, try to collect a larger sample.
- 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.