Hello,
I'm using Excel 2002 Analysis Toolpak to perform regression analysis.
I've successfuly performed the regression analysis, with a 95% Confidence
Interval entered (upper and lower limits), on different data. What I need to
do now is the following:
1. determine end points for each set of data (where it intersects with the
upper/lower limits and 95% CI.
2. draw a graphical line showing both the regression line and the upper and
lower CI lines, with just the regression analysis data calculated.
3. compare the data, within a lot at different conditions, and between lots
for the same condition to determine if they are statistically equivalent or
not

I'm a little perplexed as to what to do next, and appreciate any help you
can provide. I can draw the regression line with the % residuals calculated
for each x,y value (assuming this is the correct thing to do here), but I
only see one data point each for the upper and lower Confidence Interval, so
I am not sure how to apply this to draw a CI line. I also don't know what
data to use or how to use it to determine the similarity/difference between
each data set. Also, can a regression analysis be performed with an endpoint
for multiple lots (A, B) for one condition (1 or 2) ? Is this done the same
way, just expanding the values under x?

Here are some example numbers to help with the discussion:
y1: 0, 1, 2, 3, 6, 9, 12, 24
y2: 0, 1, 3, 7, 14, 28
x1A: 97, 97, 96, 92, 90, 92, 90, 88, 86
x1B: 93, 93, 91, 88, 87, 85, 84, 82, 81
x2A: 97, 96, 89, 82, 79
x2B: 95, 94, 85, 80, 75
Specification is 80%, lower alert limit is CI 95% of specification

Thanks for your help.
MH

Hi,

I'm not a statistician, but it sounds like you want something like this:

To make a line for the LCL and UCL add a two series like this and plot them:

Reg LCL UCL
2 3 4.5
3 3 4.5
4.3 3 4.5
5 3 4.5
....

Since you already have the regression line plotted all you need to do is
plot the LCL and the UCL series.

--
If this helps, please click the Yes button.

Cheers,
Shane Devenshire


"MH" wrote:

Hello,
I'm using Excel 2002 Analysis Toolpak to perform regression analysis.
I've successfuly performed the regression analysis, with a 95% Confidence
Interval entered (upper and lower limits), on different data. What I need to
do now is the following:
1. determine end points for each set of data (where it intersects with the
upper/lower limits and 95% CI.
2. draw a graphical line showing both the regression line and the upper and
lower CI lines, with just the regression analysis data calculated.
3. compare the data, within a lot at different conditions, and between lots
for the same condition to determine if they are statistically equivalent or
not

I'm a little perplexed as to what to do next, and appreciate any help you
can provide. I can draw the regression line with the % residuals calculated
for each x,y value (assuming this is the correct thing to do here), but I
only see one data point each for the upper and lower Confidence Interval, so
I am not sure how to apply this to draw a CI line. I also don't know what
data to use or how to use it to determine the similarity/difference between
each data set. Also, can a regression analysis be performed with an endpoint
for multiple lots (A, B) for one condition (1 or 2) ? Is this done the same
way, just expanding the values under x?

Here are some example numbers to help with the discussion:
y1: 0, 1, 2, 3, 6, 9, 12, 24
y2: 0, 1, 3, 7, 14, 28
x1A: 97, 97, 96, 92, 90, 92, 90, 88, 86
x1B: 93, 93, 91, 88, 87, 85, 84, 82, 81
x2A: 97, 96, 89, 82, 79
x2B: 95, 94, 85, 80, 75
Specification is 80%, lower alert limit is CI 95% of specification

Thanks for your help.
MH