Monday, July 29, 2013
Present: Marianne Burke, Kat Cheung, Abby Crocker, Kairn Kelley, Rodger Kessler, Connie van Eeghen
1. Start Up: Kat is in the Masters CTS program; contemplating the PhD. Discussion of the variability of soup quality in the Atrium Café vs. the FAHC cafeteria resulted in a preference for the FAHC kitchen, which has less variability in performance outcomes – recipes are pretty good too.
2. Discussion: Kairn: Power Calculations and You: How Standard Deviation Affects The Research Question
a. Kairn developed a plan to test the validity of dichotic word tests, in which the same subject gets the same test within 20 minutes. The pilot data resulted in pre to post-test differences that were much larger than expected, resulting in overlapping distributions and different means
b. The sample size is driven by the following characteristics of the testing procedure.
i. Competing word tests: 20 on the right + 20 on the left = 40 total (a commonly used test, from Kairn’s survey of interested audiologists)
ii. Dichotic digit tests: 40 right + 40 left (a commonly used test, from same survey)
iii. Dichotic consonant vowel tests: 30 right + 30 left = laterality index (R-L)/R+L (Kairn likes this test, but it is criticized as not being very reliable)
c. Given the variability in the pilot data (5 children, three tests each, 2 ears each), there are new questions:
i. For each test, is the group mean at T1 the same as the group mean for T2? Note that there is no clinical standard to determine a meaningful change in score.
ii. What is the SD of individual change in score T1 to T2?
d. Power calculation to replicate test-retest reliability published in the manual:
i. Expected effect size: 2
ii. Standard deviation of scores: 2.5
iii. Effect/SD = 0.8
iv. Sample size needed: 26 individuals for a difference of 2 or greater on a two tailed test with alpha of 0.05 and beta of 0.2
v. However, with multiple tests to analyze, a correction is needed to increase the sample size
e. Strategy: two separate populations (pediatric office and day camp) of 30 kids each, analyzing each group separately. There were some opinions offered that running two populations may be more than is needed for a dissertation, but this was left to the candidate and her committee to discuss.
a. July 31: Kairn: draft analytic plan for dissertation; also Sylvie Frisbie and the Blogspot (no Ben, Abby)
b. August 7: Marianne: Journal club on Banks DE, Shi R, Timm DF, Christopher KA, Duggar DC, Comegys M, McLarty J. Decreased hospital length of stay associated with presentation of cases at morning report with librarian support. J Med Libr Assoc. 2007 Oct;95(4):381-7. PubMed PMID: 17971885; PubMed Central PMCID: PMC2000787.
c. August 14 (no Abby, no Marianne); plan for new schedule Fall semester (week of Aug 26)
d. August 21: Abby: Opiate project update
e. August 28:
f. Future agenda to consider:
i. Peter Callas or other faculty on multi-level modeling
ii. Charlie MacLean: demonstration of Tableau; or Rodger’s examples of Prezi
iii. Journal article: Gomes, 2013, Opioid Dose and MVA in Canada (Charlie)
iv. Ben: Tukey chapter reading assignments, or other book of general interest
v. Summer plan: each week, one person will send out an article or prezi ahead for review or discussion by all. Alternatively, if a participant is working on a key document for their professional development, this is also welcome (e.g. K awards, F awards, etc.)
Posted by Connie at 7/29/2013 02:14:00 PM