PS296 Lecture Notes - Lecture 14: Confidence Interval, Null Hypothesis, Standard Deviation
Week 8 – Lecture 14
Confidence Intervals:
-the mean (Dbar) is the difference between the two related means (ex pretest and posttest means)
-the standard deviation is the standard deviation of the difference scores
-the CI represents the interval with a 95% likelihood of containing the true population difference
between the means
-if t is significant, the confidence interval should not include the population mean difference under the
null hypothesis (ex 0)
-if t is not significant, the confidence interval should include the population mean difference under the
null hypothesis
Effect Size:
-Dhat is estimated independently of n
d = .20 is small effect size
d = .50 is medium
d = .80 is large
Ex: Dhat = -1.073
Interpretation: Our estimate of the effect size (dhat = -1.07) based on the pretest standard deviation is
very large and shows that mean test scores were 1.07 standard deviations lower on the pretest than on
the posttest
-if you see a question using the words pretest and posttest, you know to run a related samples test
Document Summary
The mean (dbar) is the difference between the two related means (ex pretest and posttest means) The standard deviation is the standard deviation of the difference scores. The ci represents the interval with a 95% likelihood of containing the true population difference between the means. If t is significant, the confidence interval should not include the population mean difference under the null hypothesis (ex 0) If t is not significant, the confidence interval should include the population mean difference under the null hypothesis. Dhat is estimated independently of n d = . 20 is small effect size d = . 50 is medium d = . 80 is large. Interpretation: our estimate of the effect size (dhat = -1. 07) based on the pretest standard deviation is very large and shows that mean test scores were 1. 07 standard deviations lower on the pretest than on the posttest.