Re: Excel--mac/ibm

[Doug Craigen <dcc@cyberspc.mb.ca>]

On Thu, 27 Feb 1997, Richard W. Tarara wrote:

> First, I assume you transposed x and y below, but accepting that and

Yes, I too noticed that I needed to transpose the columns.


> running the data through a program I wrote using algorithms from
> "Introduction to the Theory of Error" by Yardley Beers,  using equal
> weighting of all points, I get:
>
> slope = 4.752 +/- .02425
>
> y-intercept = 1.0057e-16 +/- .00184
>
> Therefore, within the uncertainties, the two answers from the two versions
> of Excel are indistiguishable.  The point here is that there IS an
> uncertainty associated with any fit and in this case whatever discrepency
> there is between versions is within that uncertainty.  If you can show a
> discrepency that is outside the uncertainties, then there room for concern.

MATLAB 4.2c.1 on a PC gives the following for fitting this data with a
first order polynomial:

a = 4.7520
b = -1.01 * 10^(-18)

It gives the y-fit values as:
y = -1.01 * 10^(-18)
    0.1188
    0.2376
    0.3564
    0.4752
    0.5940

and the associated individual errors as:

delta =  0.0032
         0.0029
         0.0028
         0.0028
         0.0029
         0.0032

The value on the y-intercept is 0.0032, similar in magnitude, but almost
double what you cite above.  I was surprised that I couldn't get an error
on the slope out of MATLAB. 

The manual says the fit is performed using a Vandermonde matrix and then
solving the least squares problem.  I won't attempt to summarize the
maual's three page description of how the latter is solved using a matrix
left division, especially since it ultimately leads to a reference to the
appropriate LINPACK routines.

> > From: Donald E. Simanek <dsimanek@eagle.lhup.edu>
> >
> > x      y
> >
> > 0.002  0.000
> > 0.117  0.025
> > 0.238  0.050
> > 0.356  0.075
> > 0.472  0.100
> > 0.597  0.125
> >
> > IBM Excel gives a = 4.751429,  b = 0.0050502
> > MAC Excel gives a = 4.7514  ,  b = -0.0002

|++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++|
| Doug Craigen                                                 |
|                                                              |
| If you think Physics is no laughing matter, think again .... |
|    http://cyberspc.mb.ca/~dcc/phys/humor.html                |
|++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++|