Unformatted text preview: 9/29/13 1. Eﬀect size By increasing eﬀect size, we increase power. How could we increase eﬀect size? ▪ Just predict a larger eﬀect? ▪ No, this would be lying to yourself, and give you a false sense of power. Bad idea. ▪ Increase the expected diﬀerence between groups by changing your study design. For example: ▪ Use a stronger manipulaDon ▪ Compare more diverse groups (more extreme) UlDmately, changing eﬀect size may not be pracDcal or may change the meaning of your study 2. Standard [email protected] By decreasing standard deviaDon, we increase power. How could we decrease standard deviaDon? ▪ Collect data from a sample that is very homogenous (not much variaDon) ▪ Disadvantage is that your ﬁndings may not be generalizable ▪ Use standardized, controlled tesDng circumstances ▪ Reduces noise/error/variability Also may not be pracDcal. 3. Sample size By increasing sample size, we increase power. ▪ Probably the most common method for addressing issues related to power. When researchers apply for grants, they calculate power to determine the sample size they need to collect for the study. Must convince the funding commiyee that their study has enough power. May be more costly 19 9/29/13 4. Signiﬁcance level By using a more lenient signiﬁcance level (e.g., p < .10 or p < .25), we increase power. ▪ However, this raises the probability of a Type I error. ▪ Also, is harder to publish (because now you’re violaDng normal standards for the ﬁeld, and maybe ﬁnding things...
View
Full
Document
This note was uploaded on 03/24/2014 for the course PSY 21201 taught by Professor Bernard during the Winter '13 term at SUNY Stony Brook.
 Winter '13
 bernard
 Psychology

Click to edit the document details