BIOL 210 Lecture Notes - Lecture 24: Culturomics, Textedit, Effect Size
11/30/2017
Lecture 24
• Bring takehome with us to the exam
• S1/4 stew bio
• Whole golem is on the exam
• Heavier weight on 2nd half of the class
• Ioannidis says most papers are false because they can’t be replicated and they can’t be
replicated because they’re low power (with respect to effect size you predict, sample size
is inadequate)
• Effect size: how large will the difference be between your treatment and the control
o Perfect relationship has a slope of 1
o If your predicted effect is huge, you don’t need a big sample size
• If p=.05 the probability of a false positive is 1/20
• Universities came from religious institutions and have developed into autonomous
institutions
• Berlín: instead of teaching dogma, teach open (not specialists), the overlap of teaching
and research
• Go to web of science
o Find highly cited paper
o Create citation report
o Times cited
o Analyze results
o Sort by publication year
o Set it to 100 years
o Set min default number to 1
o Sort by selected field
o Click box that says analyze
o Save data to file
o Copy TextEdit into excel
• No questions on culturomics or google engrams on the final
find more resources at oneclass.com
find more resources at oneclass.com