st: Fwd: Dangers of using MS Excel for statistical work [SEC=UNCLASSIFIED]

[Stas Kolenikov <skolenik@gmail.com>]

This is follow up on what I posted yesterday -- if you want to keep an
eye on this discussion, these points are good to know, too.

-- Stas Kolenikov, PhD, PStat (SSC)
-- Senior Survey Statistician, Abt SRBI
-- Opinions stated in this email are mine only, and do not reflect the
position of my employer
-- http://stas.kolenikov.name




---------- Forwarded message ----------
From: Jan Werner <jwerner@jwdp.com>
Date: Sat, Apr 13, 2013 at 11:45 AM
Subject: Re: Dangers of using MS Excel for statistical work [SEC=UNCLASSIFIED]
To: SRMSNET@listserv.umd.edu


The reason ISO certification doesn't allow accounting packages that
use spreadsheets has nothing to do with accuracy and everything to do
with the lack of traceability.

Using a stat package interactively, while perfectly fine for
exploratory purposes, also does not meet ISO requirements for
documentation of every step of a procedure.

If you expect your work to meet ISO standards, you must use syntax
driven statistical procedures that are guaranteed to produce the same
results from the same initial data set while documenting what was
done.

Jan Werner
____________

Art Kendall wrote:
>
> These links show both human factors reasons and technical computation
> reasons to use Excel as a spreadsheet and to use a stat package for stat.
>
> This is one of my soapbox posts, apologies to those of you who are tired
> of my making this point.
>
> Using spreadsheets for stat *can be* like using a hammer to drive a screw.
>
>
> IF I RECALL CORRECTLY any accounting system that uses computer
> spreadsheets cannot pass ISO certification.
>
> I do use Excel Pro or tables in WordPerfect several times a week.
>
> Sometimes it is practical to enter data via a spreadsheet and then read
> it into a stat package e.g., SPSS, for checking double entry and quality
> checks. [This is because spread sheets are usually available, but it not
> cost effective to provide  stat package to every person who will do data
> entry.]
>
> For an overview of the issues
> A video can be found at
>
> http://www.spss.com/events/event.cfm?E_ID=2921&Country=US
>
> A pdf can be found by clicking
> <The Risks of Using Spreadsheets in Data Analysis
> <http://forms.cognos.com/?elqPURLPage=4206&offid=wp_spssrc_risks_of_using_spreadsheets_for_statistical_analysis_imw14297>>
>
>
>
> For more technical review see.
>
> http://www.pages.drexel.edu/~bdm25/excel-intro.pdf
>
> http://www.pages.drexel.edu/~bdm25/excel2007.pdf
>
> http://www.pages.drexel.edu/~bdm25/excel-rng.pdf
>
> an instructional link is
> http://www.umass.edu/statdata/software/handouts/excel.html.
>
> Art Kendall
> Social Research Consultants
>
> On 4/13/2013 12:44 AM, Thomas Lumley wrote:
>>
>> Stas,
>>
>> Doug Bates (who knows from nonlinear least squares) has commented that
>> some of the more difficult NIST examples are unrealistic either
>> because of essentially-zero residual error or particularly bad
>> starting values.
>>
>> The other problems in the article seem more serious, but I would be
>> careful of drawing strong conclusions about the nonlinear least
>> squares routines solely from the NIST examples. And I speak as someone
>> with well-known anti-Excel and pro-R biases.
>>
>>    -thomas
>>
>>
>>
>> On Sat, Apr 13, 2013 at 4:04 AM, Stas Kolenikov <skolenik@gmail.com
>> <mailto:skolenik@gmail.com>> wrote:
>>
>>     Thought this would be useful to know. Background: NIST (National
>>     Institute of Standards) provides a suite of tests
>>     (http://www.nist.gov/itl/sed/gsg/strd.cfm) that push the limits of
>>     computational accuracy in routine statistical computation, e.g.,
>>     having a regression model with a poorly conditioned matrix, or a
>>     nonlinear regression problem with a sharp long ridge near the maximum
>>     of the objective function that a poorly calibrated optimization
>>     algorithm may miss. More serious software packages use these data sets
>>     to gauge how they are doing against them
>>     (http://www.stata.com/support/cert/nist/).
>>     <http://www.stata.com/support/cert/nist/%29.>.. but Excel
>>
>>     developers seem
>>     to be oblivious to these standards.
>>
>>     -- Stas Kolenikov, PhD, PStat (SSC)
>>     -- Senior Survey Statistician, Abt SRBI
>>     -- Opinions stated in this email are mine only, and do not reflect the
>>     position of my employer
>>     -- http://stas.kolenikov.name
>>
>>
>>
>>
>>     ---------- Forwarded message ----------
>>     From: Geoffrey Brent <geoffrey.brent@abs.gov.au
>>     <mailto:geoffrey.brent@abs.gov.au>>
>>     Date: Fri, Apr 12, 2013 at 1:18 AM
>>     Subject: Dangers of using MS Excel for statistical work
>>      [SEC=UNCLASSIFIED]
>>     To: anzstat@lists.uq.edu.au <mailto:anzstat@lists.uq.edu.au>
>>
>>
>>     Apologies if this has been circulated already and I missed it, but I
>>     found this article about serious weaknesses in Excel's statistical
>>     functionality rather alarming. This looks to go well beyond "rare bug"
>>     type issues: for instance, when given 27 NIST standard test problems
>>     on nonlinear least squares, it gave badly wrong answers for 11 (but
>>     still claimed to have converged to a correct solution).
>>
>>     http://www.pages.drexel.edu/~bdm25/excel2007.pdf
>>     <http://www.pages.drexel.edu/%7Ebdm25/excel2007.pdf>
>>
>>
>>     The article is based on Excel 2007, but it mentions that some of those
>>     problems had been identified > 15 years earlier and remained unfixed.
>>
>>     Geoffrey Brent
>>
>>     Senior Research Officer
>>
>>     Business Survey Methodology  |  Statistical Services Branch |
>>     Australian Bureau of Statistics
>>
>>     (P) (03) 9615 7685   (M) 0422 65 35 26
>>
>>     (E) geoffrey.brent@abs.gov.au <mailto:geoffrey.brent@abs.gov.au>
>>      (W) www.abs.gov.au <http://www.abs.gov.au>
>>
>>
>>     -----------------------------------------------------------------------------------------------
>>
>>     Free publications and statistics available on www.abs.gov.au
>>     <http://www.abs.gov.au>
>>
>>
>>     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>     To subscribe/unsubscribe SRMSNet:
>>     http://listserv.umd.edu/cgi-bin/wa?A0=srmsnet&D=0&F=&H=0&O=T&S=&T=1
>>
>>     SRMS website: http://www.amstat.org/sections/srms/
>>     ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>
>>
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ To subscribe/unsubscribe
>> SRMSNet:
>> http://listserv.umd.edu/cgi-bin/wa?A0=srmsnet&D=0&F=&H=0&O=T&S=&T=1
>>
>> SRMS website: http://www.amstat.org/sections/srms/
>> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ To subscribe/unsubscribe
> SRMSNet:
> http://listserv.umd.edu/cgi-bin/wa?A0=srmsnet&D=0&F=&H=0&O=T&S=&T=1
>
> SRMS website: http://www.amstat.org/sections/srms/
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
To subscribe/unsubscribe SRMSNet:
  http://listserv.umd.edu/cgi-bin/wa?A0=srmsnet&D=0&F=&H=0&O=T&S=&T=1

SRMS website:  http://www.amstat.org/sections/srms/
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/