Understanding Bar Graph Confidence Intervals
The bar graphs presented in the Results Analysis section include 95% confidence intervals to illustrate the degree of precision available in your results. For example, in the following graph 54.2% (160/295) of the respondents indicated they will vote Democrat vs. 45.8% (135/295) Republican.
Because the survey is based on the results of only 295 respondents, the actual percent of people who will vote Democrat could be somewhat higher or lower than 54.2%. Confidence intervals tell you how much higher or lower the percent could be.
The I-bar shows, and the tip of each bar illustrates, the spread between the lowest and highest value you are likely to see if you were to survey the entire population. In the example above, you can be 95% certain that the actual percent of people who will vote Democrat will be between 48% and 60%. Furthermore, somewhere between 40% and 52% of people will vote Republican. As you increase the number of respondents the range of uncertainty shrinks.
Each bar graph group is followed by the text "Confidence:" and a percentage.
This number is the largest confidence interval found on any of the bars in the group and can be used as a summary measure of precision. The more precise, non-symmetrical confidence intervals are illustrated separately on each bar.
Some bar graph groups are followed by the text "Average Score:" and a number that represents the weighted average of all options chosen by the respondents. For example, if you asked respondents to rate their satisfaction on a scale including Very satisfied, Satisfied, Neutral, Dissatisfied, and Very dissatisfied, and half responded Very satisfied and half responded Satisfied, the average score would be 1.5--half chose the first option (score=1) and half chose the second option (score=2); so, the average score is 1.5.
When a statistically significant correlation between the answers of any two questions is found, the report will include a note highlighting the correlation. This information can be used to gain insight into what factors drive key measures such as overall satisfaction.
The answers to two questions are correlated when they tend to move together. For example, if you ask respondents to rate their overall satisfaction with your company and also ask if they are likely to purchase from your company again, the answers to these questions will probably show a strong correlation. That is, when satisfaction is high, the likelihood of repeat purchase is high. This is a positive correlation. Some question pairs have negative correlation. For example, the time a person spends on hold when calling for support usually has a negative correlation with overall satisfaction. Correlation is presented as a number from -1 to 1 where -1 is perfect negative correlation, 0 is no correlation, and 1 is perfect positive correlation.
View Survey Results
Understanding Probability Density Function
Understanding Cumulative Distribution