Andrea M. Karkowski

Behavioral Sciences Department

Capital University

One College and Main

Columbus, OH 43209 USA

Phone: 614.236.6449

Email: akarkows@capital.edu

Courses

Publications

Grants

Presentations

Student Research

Talks & Colloquia

Organizations

Interests


Behavioral Sciences Statistics

10 Foundational Questions that Educated People Should be able to Ask and Answer when Evaluating Evidence*

Adapted from Neil Lutsky (NITOP 2007)

  1. What do the numbers show?
  2. How representative are the numbers?
  3. Compared to what?
  4. Is the outcome statistically significant (and what does it mean if it is, or isn't)?
  5. What is the effect size?
  6. Are the results of a single study or of a literature?
  7. What is the research design (e.g., correlational or experimental)?
  8. How were the variables operationalized?
  9. Who is in the sample?
  10. What was controlled (and what was not controlled)?

*This class helps students to ask and answer questions 1 - 5. Psychological Research Methods helps students to ask and answer questions 6 - 10. Experimental Psychology helps students to generalize their ability to ask and answer all 10 questions.

Quotes

"All statistics are social products, the results of people's efforts."

Joel Best, sociologist, 2001

"People lie more frequently using English than they do using statistics."

-- author unknown

"Most students do not really understand statistical material until they have been through it three times -- once for exposure, once for practice, and once for the dawn of genuine insight."

-- Robert P. Ableson, psychologist, 1995

"I'm a woman. I'm a black woman. I'm a poor woman. I'm a fat woman. I'm a middle-aged woman. And I'm on welfare. In this country, if you're any one of those things, you count less as a person. If you're all of those things, you just don't count, except as a statistic."

-- Johnnie Tillmon, welfare rights activist, 1972

"If there is a 50-50 chance that something can go wrong, then 9 times out of ten it will."

-- Paul Harvey, news caster, 1979

Statistical Limericks

Statistics I dream each night

No longer give me a fright

Within and between

I know what they mean

And choose the test that is right

Stats are like an old lover

Which test to do, fight over

t tests are best

When two groups we test

More than two levels, ANOVA

(The rhyme works with a Boston accent.)

Independent samples t

Assumes homogeneity

Of variance between

Different groups that mean

More precise results than z

Repeated measures, a wiz

Pretest and posttest it is

Can do a paired t

Or ANOVA for three

Or more levels, ace the quiz

Last Modified: 17 August 2009