> Hi
>
> Several weeks ago, I requested some material to be used in a presentation.
> I put the request on both Allstat and Radstats. I have had many useful
> pointers, which I enclose below. I am very grateful to those who took the
> time to email me back and send me material by email and regular post. You
> have been of great help. I would like to supplement those information with
> others that I found also very useful.
>
> Chatfield, C. 1991. Avoiding statistical pitfalls. Statistical Science, 6,
> 240-268.
>
> Statistical resources,
> <http://www.stat.psu.edu/~resources/Topics/sampdist.htm>
>
> Nelder, J.A. 1999. Statistics for the millennium. The Statistician, 48
> (2), 257-269.
>
> Finney, D.J. 1991. Ethical aspects of statistical practice. Biometrics,
> 47, 331-339.
>
> Nester, M.R. 1996. An applied statistician's creed. Applied Statistics, 45
> (4), 401-410.
>
> Lindley, D.V. 2000. The philosophy of statistics. The Statistician, 49
> (3), 293-337.
>
> Sahai, H. and Khurshid, A. 1999. A bibliography on statistical consulting
> and training. Journal of Official Statistics, 15 (4), 587-629.
>
> Lindsey, J.K. 1999. Some statistical heresies. The Statistician, 48 (1),
> 1-40.
>
> Sprent, P. 1998. Statistics and mathematics- trouble at the interface? The
> Statistician, 47(2), 239-244.
>
> Hand, D.J. 1998. Breaking misconceptions - statistics and its relationship
> to mathematics, The Statistician, 47(2), 245-250.
>
> Senn, S. 1998. Mathematics: governess or handmaiden. The Statistician,
> 47(2), 251-259.
>
> Bailey, R.A. Statistics and mathematics: the appropriate use of
> mathematics within statistics. The Statistician, 47(2), 261-271.
>
> Bennett R.J. and Haining, R.P. 1985. Spatial structure and spatial
> interaction: Modelling approaches to the statistical analysis of
> geographical data. Journal of the Royal Statistical Society, 148 (1),
> 1-36.
>
> Drew, D., and Demack, S. 1998. A league apart: statistics in the study of
> 'race' and education. Researching racism in Education: Politics, theory
> and practise, Connolly and Troyna (eds). Buckingham Open University Press.
>
> Atkinson, A.C. and Fienberg (eds.). 1985. A celebration of statistics: The
> ISI centenary volume. Springer-Verlag, New York.
>
> Brook R.J., Arnold, G.C., Hassard, T.H., and Pringle R.M. (Eds.). 1986.
> The fascination of statistics. Marcel Dekker, New York.
>
> Hand, D.J. and Everitt, B.S. (Eds.). 1987. The statistical consultant in
> action. Cambridge University Press, Cambridge.
>
> Huff, D. 1954. How to lie with statistics. Norton & Company, New York.
>
> Coomaren
> --------------------------------------------------------------------------
> --------------------------------------------------------------------------
> ----------
> Dear Coomaren,
>
> I hope your presentation tomorrow goes off OK: far too late to be of
> immediate use, I received the following via the Datateach list. Since I
> don't know if you see this list, I thought I would forward it.
>
> Julian Wells
>
> -----Original Message-----
> From: Booth, Charles [mailto:[log in to unmask]]
> Sent: Monday, November 27, 2000 6:18 PM
> To: [log in to unmask]
> Subject: Interpreting & Evaluating Published Research
>
>
> I've had a very useful and gratifying response to my email
> on interpreting and evaluating quantitative research - my
> sincere thanks to all who took the time and effort to email
> me. As promised, here is a digest of responses:
>
> Richard Makadok recommended Cook & Campbell's
> "Quasi-Experimentation: Design & Analysis Issues for Field
> Settings, " and the following two articles:
> Leamer (1983), "Let's take the con out of econometrics,"
> American Economic Review
> McAleer et al (1985), "What will
> take the con out of econometrics?" American Economic Review
>
> Stephen Morris recommended "Statistics for the Terrified",
> commercially available at
> http://www.conceptstew.co.uk/s4t.html
>
> Allan Reese drew my attention to the importance of the
> graphical presentation of data, and recommended:
> Tufte E. "The Visual Display of Quantitative Information"
> Cleveland W. "The Elements of Graphing Data"
>
> Kim Boal suggested Dick Mason and Burt Swanson (eds)
> "Measurement for management decision"
>
> Patrick Butler recommended his own Business Research
> Sources: A Reference Navigator, published by McGraw-Hill in
> 1999, shortly to be available at
> www.businessresearchsources.com, plus also Business
> Information: How to Find It, How to Use It, by Michael
> Lavin
>
> Bob Stephens recommended "Reading Statistics and Research"
> by Huck, Cormier, and Bounds (as did Mike Rehg and Tom
> Sigerstad) and also "Research Methods and Organization
> Studies" by Alan Bryman.
>
> Heidi Neck pointed me towards Girden, R. 1996. Evaluating
> research articles from start to finish.
> Thousand Oaks, CA: Sage Publications.
>
> Instead of a textbook, Maryellen Kelley suggested I get the
> students to compare a number of studies on the same
> subject, comparing methods, measures, and results.
>
> Milo Schield drew my attention to his website, where useful
> materials are available: www.augsburg.edu/ppages/schield
> (In a few months, this will change to
> www.augsburg.edu/ppages/~schield)
>
> Elie Giesler pointed me towards E. Geisler, Methodology,
> Theory, and Knowledge in the Managerial and organizational
> Sciences, Greenwood Press, 1999 (foreword by Ev
> Rogers).ISBN:1-56720-307-8 and E. Geisler, The Metrics of
> Science and Technology, Quorum Books, 2000, ISBN:
> 1-56720-213-6
>
> Ken Reed suggested "Reading and understanding multivariate
> statistics" edited by Laurence G.Grimm and Paul R. Yarnold
> Published: Washington, D.C. : American Psychological
> Association, c1995
>
> [log in to unmask] suggested I investigate Kennedy's "A Guide
> to Econometrics".
>
> Frank Winfrey suggested the Cummings and Frost book,
> "Publishing in the Organizational Sciences" (Sage) and
> recommended (as did Vincent Daly) that Sage would be an
> appropriate publisher to check out.
>
> Jim Underwood drew my attention to How to Think About
> Statistics by John L. Phillips, Jr. Pub. W. H. Freeman &
> Co., NY
>
> Thanks also to Jamie Hendry and Robin Rice for their
> supportive comments.
>
> Many thanks again to all who took the troiuble to contact
> me.
>
> Best wishes
>
> Charles
> ----------------------------------------
> Booth, Charles
> Email: [log in to unmask]
> "University of the West of England"
>
>
> --------------------------------------------------------------------------
> --
> Dear Coomaren,
>
> I'm sure that "very busy" describes part of the reason for lack of
> response
> -- but perhaps also it is due to uncertainty about what is required.
>
> I gather that your audience are grown-ups with some experience of using
> statistics, but this still leaves open a range of possibilities: does
> "using
> statistics" mean, say, calculating averages and other measures of central
> tendency, or does it extend to (e.g.) doing significance tests or
> regression
> analysis "cook book" fashion without really understanding the
> justification
> -- and hence limitation -- for this?
>
> And are they practitioners in real life (e.g. from your address, forest
> managers) or are they academics in other disciplines who have picked up a
> bit of self-taught stat.s along the way?
>
> Another question is how "radical" a critique are you looking for, and of
> what, precisely?
>
> At one end of the scale there are old-established texts like Darrell
> Huff's
> "How to lie with statistics", which should cover the case of naive users
> of
> basic summary statistics.
>
> At the other end of the scale there is RadStats own "De-mystifying social
> statistics".
>
> Somewhere in the middle might be Tufte's book on the visual presentation
> of
> data.
>
> Note that most RadStats members -- as far as I can see -- are more into
> the
> critique of data collection (both methods, and of *what* is sought) than
> the
> critique of statistical techniques as such (though there is some material
> in
> this vein in the "De-mystifying..." text, e.g. pointing out the
> implication
> of regression techniques in eugenicist ideas).
>
> Nor are they -- again, as far as I can see -- interested much in the
> philosophical problems of probability.
>
> Don't know if this helps, but feel free to come back for more!
>
> Best wishes,
>
> Julian Wells
> OU Business School
> The Open University
> Walton Hall
> Milton Keynes
> MK7 6AA
> United Kingdom
> +44 1908 654658
>
> --------------------------------------------------------------------------
> -------
> Coomaren,
>
> You may like to see a chapter I wrote recently in a book. If you cannot
> get a
> copy email me and I will photocopy my original and send it to you. I do
> not
> have an electronic copy.
>
> Drew, D and Demack,S. (1998?) A league apart: statistics in the study of
> race
> and education, in P Connolly and B Troyna (eds) Researching Racism in
> Education: Politics, Theory and Practice. Buckingham: Open University
> Press.
>
> This chapter is about the way the worlds of three groups of people
> interect;
> ethnic monority young people at school, policy makers and researchers who
> endeavour to reflect the realities of the worlds and the problems of the
> first two groups.
>
> Topics covered include;
> Statistical procedures and discussion of methodology;
> Model building with three case studies (educational attainment at 16,
> school
> transitions from 16-19 and ethnicity and school effectiveness).
>
> The aim is to show how the choice of methods and models affected the
> studies
> in each case and how basic issues like non response bias, problems of
> cross
> sectional studies, sampling, operationalisation of variables and
> contextualising the analysis are very important. The actual results of the
>
> studies are presented and discussed along with the issues which really
> concerned the researchers (me for example) when analysing the data. The
> three
> examples were all from published work of my own or the work of close
> colleagues in the field of educational research. I hope that we were able
> to
> give a real flavour of the problems of actually doing research. The models
>
> described in non technical terms in the three case studies are simple
> linear
> regression, logit models and multi level models.
>
> Best wishes
>
> Dr David Drew
>
> --------------------------------------------------------------------------
> ---
> Try Darryl Huff's "How to lie with statistics". I think it was published
> around 1957 but it's just as valid today as ever.
>
> Regards
>
> Paul Hewson
> Senior Worst Value Officer
> Chief Executive's Scapegoat Unit
> Torbay Madhouse
>
> email [log in to unmask]
> --------------------------------------------------------------------------
> -----
> Hi Coomaren
>
> An area I have always found interests an almost 'NON-STATISTICAL' audience
> (like the one you have to talk to) is whether assumptions for a
> statistical
> test produce something real, relevant or even practical.
>
> For example, if they are using the Chi-Square goodness-of-fit test. Does
> pooling some of the categories (to make expected frequencies at least 5)
> produce a real or practical category?
>
> Basic example: Colour of Cars
>
> Blue White Yellow Red Orange
> Observed 20 30 2 1 3
> Expected 25 25 1 1 4
>
>
> Usually you would have to pool Yellow, Red and Orange cars into one
> category to satisfy the assumption. So does this category mean anything?
> Probably not.
>
>
> All the best for the talk.
>
> Mike
> Mike Steele
> Lecturer in Statistics & Mathematics
> School of Mathematical and Physical Sciences
> James Cook University
> Cairns QLD 4870
> Telephone: (07) 4042 1225
> Fax: (07) 4042 1284
> Web: http://swarms.jcu.edu.au
>
>
> --------------------------------------------------------------------------
> ---
> Dear Coomaren,
>
> You may like to look at the website for our software:
>
> http://www.conceptstew.co.uk/s4t.html
>
> 'Statistics for the Terrified' was designed for exactly the audience you
> describe.
>
> Jill Szuscikiewicz
>
>
> --------------------------------------------------------------------------
> ---
> Dr. Vencatasawmy:
>
> This is in response to your posting in radstat concerning material for
> presentation. I am the site editor for SecondMoment.org. It is a site for
> academics and industry analysts in the field of applied analytics and is
> set
> up as a magazine/discussion group. It sounds like some of the articles on
> the site might be useful, especially the section on "Lies, Damned Lies and
> Statistics" which points out misuse of statistics. Our feature articles
> cover some of the state of the art in statistics. It is written for
> industry
> analysts, but can be followed by the general public. Please let me know if
> you have any questions. The site is at www.secondmoment.org .
>
> Yong Kim
>
>
> --------------------------------------------------------------------------
> It happens that I am addressing a group of senior doctors this Friday,
> with the title "Research Methods and Research Culture". I would be happy
> to provide you with the notes I am preparing for that talk.
>
> You might also look at the notes circulated onthe ALLSTAT list a couple
> of weeks ago, responding to the suggested training needs of research
> students, laid down by a UK research council (funding body for
> studentships). The Research Council have greatly modified their proposals
> in the latest web posting (www.esrc.ac.uk)
>
> R. Allan Reese Email: [log in to unmask]
> Associate Manager Direct voice: +44 1482 466845
> Graduate Research Institute Voice messages: +44 1482 466844
> Hull University, Hull HU6 7RX, UK. Fax: +44 1482 466436
> ====================================================================
> English is becoming an aggregate of vocabularies only loosely in
> connection with each other, which yet have many words in common, so
> that there is much danger of accidental ambiguity, and you have to
> bear firmly in mind the small clique for whom the author is writing.
> Willam Empson, Seven Types of Ambiguity
> -----------------------------------------------------------------------
> This should help you with (1), and might give you some pointers for (2)
>
> Dorling, D. and Simpson, S. (eds) Statistics in Society: the arithmetic
> of politics, Arnold: London.
>
> Mary Shaw
> ----------------------
> Dr. Mary Shaw, Research Fellow
> School of Geographical Sciences
> University Road
> Bristol BS8 1SS
> Tel. 0117 928 9000 Ext. 3843
> Fax. 0117 928 7878
> [log in to unmask]
>
> --------------------------------------------------------------------------
> -
> Dear Coomaren,
>
> As a member of both Radstats and RSS I am surprised and rather appalled at
>
> the lack of response you have had. Rad stats in particlar has as part of
> its mission to tackle 'the mystifying use of technical language to
> disguise
> social probalems as technical ones'. (Who did you try to contact at
> Radstats)
>
> To this end they have a number of publications which you may find useful.
> I
> can't think of a recent one which would serv your purposes but 'reaxing
> between the numbers' compares and criticises particular approaches to
> modelling educational data. Contact Chistina Pantazis (editor of
> newsletter)
> [log in to unmask] Also perhaps Jay Ginn of the troika (organising
> committee) - e mail above.
>
> There are many statistics texts of the type you mention . One useful ,
> for
> memeory , is Mostellor and Tukey. You may think of producing one yourself!
>
> Hope you get some success,
>
>
> Russell Ecob
>
>
> ***************
> ******************************
> Russell Ecob
> 36 Prospecthill Road
> Glasgow G42 9LE
> Scotland, UK
>
> +44(0)141-649-9387
> www.ecob-consulting.com
> [log in to unmask]
> mobile: 07788-145542
>
> --------------------------------------------------------------------------
> --
> Dear Coomaren,
> I'm not sure whether this is exactly what you had in mind, but it is very
> topical:
> http://madison.hss.cmu.edu/
> It gives some regression plots showing that the voting records in
> Palm Beach county in Florida were apparently severely aberrant given those
> in other Florida counties, with far more votes for Buchanan than expected.
>
> The text also gives some justifications for the selection of the
> variables to be regressed (i.e. votes for Buchanan against votes for
> Bush), and a brief discussion of problems with the simple regression model
> (e.g. heteroskedasticity). Makes interesting reading, anyway.
>
> All the best,
> Rachel Fewster
>
> Rachel Fewster ([log in to unmask])
> Department of Statistics, University of Auckland,
> Private Bag 92019, Auckland, New Zealand.
> tel: 64 9 373 7599 extn 3946 fax: 64 9 373 7018
> http://www.stat.auckland.ac.nz/~fewster/
> =====================================================================
> The Hunger Site: make a free donation at http://www.thehungersite.com
>
>
> --------------------------------------------------------------------------
> ----
> It's your audience, so you know best. If I faced them, I would present a
> talk
> focused on the results, and use the interest in these to indicate the
> value of
> the process that was used to get there. And of course, the process is
> statistical thinking, fact-based analysis, etc., e etc.
>
> Examples of results where the underlying analysis is known & valid,
> suggest
> you look at my web site - case history section.
>
> 1st ex. used an Ishikawa 'fishbone' chart to organize, then brain
> storming,
> the A2Q method (read scientific method, write large), then a DoE to find
> most
> beneficial improvements. SPC chart shows the effect of changes.
>
> 2nd example is 2 Pareto charts, pointing to most beneficial location for
> engineering analysis.
>
> If you like, I can ship you my Power Point slide show, with some
> additional
> examples.
>
> Feel free to use these - just give me, and Warner Consulting, Inc., credit
> for
> the projects.
>
> As for methods & thinking: For an audience unfamiliar with the topic, I
> would
> put emphasis on (a) planning the experiment ahead of (b) collecting data &
> analysis. Only by this path can they get to (c) lasting results. People
> frequently say they want to know what analysis 'tool' to use. I reply by
> asking what the question to be answered is. That comes first, and
> controls
> the rest. I tell my students that when they can set up a spread sheet (or
> software), put in some dummy values, and get calculated results, then they
> are
> ready to collect data. Extreme, but it make the point. Critical for a
> DoE.
> And DoE's are critical for solving most real problems.
>
> For an audience familiar with basic stat, I would emphasize the value of a
> designed experiment (2 level factorial design for starters) and the need
> for
> an orthogonal array.
>
> No one who hasn't done it understands the power of a properly done DoE.
>
> Jay
>
> Jay Warner
> Principal Scientist
> Warner Consulting, Inc.
> 4444 North Green Bay Road
> Racine, WI 53404-1216
> USA
>
> Ph: (262) 634-9100
> FAX: (262) 681-1133
> email: [log in to unmask]
> web: http://www.a2q.com
>
> The A2Q Method (tm). What do you want to improve today?
>
> --------------------------------------------------------------------------
> Dear Coomaren
>
> I think there would be useful material in Radstats newsletters. Has anyone
> offered to send you any? It may be that something is on the way, but has
> been delayed by the near collapse of the transport system in our country.
>
> Alison
>
> _________________________________________________________________________
>
> Alison Macfarlane Tel (44) (0) 1865 226706 (direct)
> Medical statistician 227000 (unit)
> National Perinatal Epidemiology Unit Fax (44) (0) 1865 227002
> Institute of Health Sciences
> Old Road
> Oxford OX3 7LF
> England
>
> Email [log in to unmask]
> _________________________________________________________________________
>
>
|