Mean Diamond Confidence Intervals

gugiyu — Tue, 23 Apr 2013 21:39:42 GMT

I'm doing a simple Oneway Analysis, ~10 groups each with # of data points ranging up to 20.

I noticed that there are 2 95% CIs that are calculated:

1. One set for each group in the Means and Std Dev report below the chart, where CIs are calculated off of the Std Dev of each individual group

2. One set for the Anova, where the Std Dev of all groups are assumed to be the same (calculated off of the Std Dev of the entire population).

The chart shows the mean diamonds, which shows #2 option above.

However, 2 of my groups have significantly higher std dev, and I'd like that to be reflected in my chart and CIs.

My questions are:

1. Is there a way to get the Upper/Lower CIs from the "Mean & Std Dev" report to show up on the graph?

2. Is there a reason not to do this?

Thanks in advance.

topic Mean Diamond Confidence Intervals in Discussions