This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Independent samples ttest We “know” nothing! Use two samples. NHST: Null Hypothesis Significance Testing The “Prequel” – zscores and the normal distribution “NHST” – ztest, single sample ttest “NHST Part 2 – The Sequel” – dependent ttest, independent ttest “NHST Part 3 – ANOVA” – coming soon to a lecture near you “NHST The Final Installment” – Factorial ANOVA, power and effect size ttests • Single sample ttest: One set of scores from one group of people. • Dependent samples ttest: Two sets of scores from one group of people. • Independent samples ttest: Two sets of scores from two groups of people. We “know” nothing! Use Two Samples • Population 1 – Estimate μ 1 from sample 1 (M 1 ) • Examples: males, ethnic majority, introverts, treatment group • Population 2 – Estimate μ 2 from sample 2 (M 2 ) • Examples: females, ethnic minority, extroverts, control group NOTE! Your independent variable is your two samples: IV = gender (1 = males, 2 = females) IV = ethnicity (1 = minority, 2 = majority) IV = personality (1 = introvert, 2 = extrovert) IV = treatment (1 = treatment, 2 = control) Estimate variance Homogeneity of Variance assumption: Population 1 variance (σ 2 1 ) = Population 2 variance (σ 2 2 ) BUT we have TWO estimates of variance, one from sample 1 and one from sample 2. s 2 1 and s 2 2 NOTE! s 2 1 = s 2 2 Estimate variance BUT we have TWO estimates of variance, one from sample 1 and one from sample 2. s 2 1 and s 2 2 We use both of them to estimate the population variance. We pool the variance . Pooled variance s 2 pooled = (df 1 /df total )s 2 1 + (df 2 /df total )s 2 2 df 1 = degrees of freedom for sample 1 s 2 1 = variance of sample 1 df 2 = degrees of freedom for sample 2 s 2 2 = variance of sample 2 df total = df 1 + df 2 Comparison Distribution Distribution of Means from Population 1 (we will estimate using sample 1 mean and pooled variance) Distribution of Means from Population 2 (we will estimate using sample 2 mean and pooled variance) Comparison Distribution For the ttest for independent samples we calculate the difference between sample 1 mean and sample 2 mean. (M 1 – M 2 ) We will compare the difference between sample means to the Distribution of Differences Between Means (all possible differences between means). Comparison Distribution Distribution of Means from Population 1...
View
Full
Document
 Spring '11
 VictoriaHarmon

Click to edit the document details