Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: [Phys-L] Surprise, Surprise! Active Participation beats Lectures

One of the comments on this information totally misunderstood the import.
They mistook the 1/2 STD for an error bar and went into a discussion of how
the metastudy showed nothing. Essentially the STD is calculated for the
grade distribution and is not an error bar. The proper term is that the
effect was 1/2 effect size where the effect size = increase in average/STD
of the grades. This particular measure has the problem that it can be
misleading when you have a class with closely clustered grades, but it can
be useful for comparing classes with a "normal" distribution of grades. For
those not familiar with the literature an effect size of 1.0 can be achieved
by individual tutoring, and an effect size of 0.5 is extremely large.
Actually some PER experiments show effect sizes larger than 1.0.

John M. Clement
Houston, TX reported in