Clinical Research Oriented Workshop (CROW) Meeting: June 17, 2021

 

Present:   Claudia Abbiati, Justine Dee, Juvena Hitt, Ben Littenberg, Liliane Savard, Adam Sprouse-Blum, Connie van Eeghen (7)

 1.                   Warm Up: all here…

2.                   Claudia’s research question about motor anomalies for people with autism:

a.       Overview of RQ

b.       Challenges

                                                   i.      Many variables: finger, larynx, lips

c.       Strengths

                                                   i.      Clearly stated question: pair wise testing

1.       Omnibus test: a test that looks for many outcomes at the same time with a p value that indicates if something, rather than a particular thing, is the variable that matters.  ANOVA is one of those tests: the F statistic and p value tell you that one group was different from the others. Test with p = .05 and a relatively small sample size

2.       Handy if there’s a big p value: you’re done

3.       Troublesome if there’s a small p value: what to do next?

d.       Options

                                                   i.      Create three hypotheses and test each independently, e.g. finger vs. larynx; larynx vs. lips; lips vs. finger

1.       Power for less than .05 to correct for multiple comparisons

a.       .05/3 = .0167

b.       Will need more subjects

e.       Study plan:

                                                   i.      40 subjects, 20 in each arm

                                                 ii.      Re-evaluate the power depending on effect size and variance or SD

f.        Interaction of variables leads to even more complexity in testing for outcomes

                                                   i.      This may not be a hypothesis test question; it may be a descriptive question

                                                 ii.      Based on prior literature, theory of mechanisms, the one hypothesis is… e.g. maximization of vocalization, with a list of secondary hypotheses

                                               iii.      OK to say that the final hypothesis is not determined and that preliminary work (a PILOT) will set that stage

g.       Manage the portfolio of ideas: do what is essential for the academic program, as well as what is essential for funded work.

3.                   Next week:  Jen will let us know; include Claudia