15 supervisors submit estimates of time to complete a task. I would like to
plot the values, and, a normal distribution curve - then, using the chart,
draw a consensus as to the most accurate time for the task. What would
comprise such a chart and, would there be a way to model the estimated times
to produce the consensus (formulas, functions, VBA, etc)?

Thanks,
Phil

Phil Hageman
15 supervisors submit estimates of time to complete a task. I would like
to
plot the values, and, a normal distribution curve - then, using the chart,
draw a consensus as to the most accurate time for the task. What would
comprise such a chart and, would there be a way to model the estimated
times
to produce the consensus (formulas, functions, VBA, etc)?

Thanks,
Phil

Hi Phil,

I am not sure that you will necessarily get a "normal curve". I might be
inclined just to plot a histogram.

Tools | Data Analysis | Histogram.

Just food for thought.

Regards,
Kevin

I agree that it's usually a good idea to first "look at the data." And, if
you want a better histogram, arrange the fifteen values in a column on a
worksheet, and use my free add-in to obtain the frequency distribution and
histogram. Visit the "Better Histogram" page at www.treeplan.com.

- Mike

www.mikemiddleton.com

++++++++++++++++++

"Kevin H. Stecyk" wrote in message
...
Phil Hageman
15 supervisors submit estimates of time to complete a task. I would like
to
plot the values, and, a normal distribution curve - then, using the
chart,
draw a consensus as to the most accurate time for the task. What would
comprise such a chart and, would there be a way to model the estimated
times
to produce the consensus (formulas, functions, VBA, etc)?

Thanks,
Phil

Hi Phil,

I am not sure that you will necessarily get a "normal curve". I might be
inclined just to plot a histogram.

Tools | Data Analysis | Histogram.

Just food for thought.

Regards,
Kevin

Michael R Middleton wrote...
I agree that it's usually a good idea to first "look at the data." And, if
you want a better histogram, arrange the fifteen values in a column on a
worksheet, and use my free add-in to obtain the frequency distribution and
histogram. Visit the "Better Histogram" page at www.treeplan.com.

- Mike

www.mikemiddleton.com


Hi Mike,

Thank you for commenting. I found your website very interesting and have
bookmarked it.

Best regards,
Kevin

Mike and Kevin,
Thanks for your replies - I'll set up a test case and see if the histogram,
plus a mean and standard deviation provide the consensus I'm looking for.
FYI, John Walkenback discribes a normal distribution/histogram combination
chart in his book "Excel Charts" (pg. 281), but he's a little sketchie on
its construction. This may a good model, but will take some work to figure
out.
Regards,
Phil

"Kevin H. Stecyk" wrote:

Michael R Middleton wrote...
I agree that it's usually a good idea to first "look at the data." And, if
you want a better histogram, arrange the fifteen values in a column on a
worksheet, and use my free add-in to obtain the frequency distribution and
histogram. Visit the "Better Histogram" page at www.treeplan.com.

- Mike

www.mikemiddleton.com


Hi Mike,

Thank you for commenting. I found your website very interesting and have
bookmarked it.

Best regards,
Kevin