"Modeling in Undergraduate Statistics"
with Michael Granaas, University of South Dakota
Hosted by: David Kline, The Ohio State University
In recent years much of the social and behavioral research community has been focused on flaws with Null Hypothesis Testing (Nickerson, 2000). Undergraduate statistics texts in the social and behavioral sciences have grown in size adding or expanding material on topics such as effects sizes, power, and confidence intervals. Meanwhile methodology in the social and behavioral sciences has been undergoing a "quiet revolution" making ever greater use of models and model fitting (Rodgers, 2010) rather than hypothesis testing. This change is evident in a number of graduate level texts but is almost entirely absent from undergraduate texts. I will introduce the basic differences between fitting models and testing hypotheses. This will be followed by open discussion and/or Q&A as appropriate to the audience present.
Having trouble viewing? Try: Download MPEG-4 Podcast (.mp4)
Comments
Paul Hewson:
Hear hear! Have you seen anything like Ben Bolker's book (for graduate ecology students press.princeton.edu/titles/8709.html)? I guess my next question is how much of that material is possible for undergrads?
Michael Granaas:
This book goes further than any I have seen so far. Clearly beyond the undergrad level, but this is what we need to prepare them for.
Michael Granaas:
Another presenter mentioned Daniel Kaplan's book www.amazon.com/dp/1448642396 which I had not heard of before. The ToC looks very promising.
Daniel Kaplan:
The second edition is posted at www.mosaic-web.org/go/StatisticalModeling/
-Danny Kaplan
Michael Granaas:
Today I learned that the latest draft of K-12 Science Education Standards from the FABBS includes recommendations regarding increased emphasis on models & modeling
www.nextgenscience.org/sites/ngss/files/Conceptual Shifts in the Next Generation Science Standards FINAL PUBLIC May Draft.pdf