(American Statistical Association Consulting Section)
Some Extremes in
Client Types,
Their Characteristics
and the Expected Roles
of the Statistical Consultant
Gerald van
Belle
Department of
Environmental Health
The University
of Washington
Seattle,
Washington 98195
(evidently originally from Gerald van Belle 1982. Some aspects of teaching biostatistical consulting. in Rustagi J.G & D.A. Wolfe (eds) Teaching of Statistics and Statistical Consulting. Academic Press [pp 343-365])
| Observed Client Type | Expected Characteristic | Consultant Type |
|---|---|---|
| 1. Dempsey Dumpster | Wants to dump the whole problem into the consultant's lap. | Slave |
| 2. Rubber Stamper | Just wants p-value < 0.05 certified. Preferably p < 0.01. | Witch Doctor |
| 3. Type III Personality | Comes with wrong question. Does not want to consider correct answer. | Type III Statistician |
| 4. Airport Professor | Wants analysis yesterday for tomorrow's meeting in New Zealand; leaving tonight. | Infinite Statistician |
| 5. Dr. Northcote | Will take all available - and unavailable - time. | Dr. Parkinson |
| 6. Kitchen Statistician | Wants statistician's approval of paper's statistical analysis, severely criticized by referee. | Full Professor |
| 7. Dr. T. V. Savant | Has vague feeling of needing statistics for valuable dataset. Wants statistician to define the question as well as the answer and write it up. Will consider making statistician 5th author on this important paper. | Psychiatrist |
| 8. Innocent Abroad | Does not understand t-test but wants factor analysis on 20 cases with 50 variables. | Magician |
| 9. Tom Thumb | Everything's gone wrong in the experiment but the data are still "valuable and valid". | Incurable Optimist |
| 10. Hydra | Every answer generates 9 more (quick) questions. | Hercules |
| 11. Wallflower | Successively more complicated (simple) explanations to save the consultant's time. | Job |
| 12. Dr. Avis | Wants some advanced analysis that competitor used in paper on same topic. Is sure competitor didn't understand it either. | Dr. Hertz |
| 13. Tinkerer | Just wants to make a minor change and rerun analysis that took 5 weeks. | Sisyphus |
| 14. Encyclopedia Salesman | Just wants 5 minutes for answer to simple question. Can be handled over phone. | Gullible George |
| 15. Scientist | Considers statistician as coworker and collaborator in the scientific enterprise. | Statistician |
Readings on Statistical Consulting (NCSU)
STATISTICS 641/841
Statistical Consulting
Spring 2002 -- Jacquelin Dietz
ATLAS (Applied Technologies for Learning in the Arts and Sciences) at U Illinois Urbana-Champaign --interesting mix of talents
Here's a nice example of the wonders we are working with: an idle question about the date of first Norse settlement of Iceland led to an observation about recent DNA findings, and to a google search that led to a DNA listserv posting suggesting a search of PubMed, and noting that AJHG had full text linked... and sure enough:
mtDNA and the Islands of the North Atlantic: Estimating the Proportions of Norse and Gaelic Ancestry (Agnar Helgason, Eileen Hickey, Sara Goodacre2 Vidar Bosnes, Kári Stefánsson, Ryk Ward, and Bryan Sykes)
Am. J. Hum. Genet., 68:723-737, 2001
© 2001 by The American Society of Human Genetics. All rights reserved.
This stuff just wasn't in the calculus five years ago, and now it's at the center of what teachers and learners have and use.