{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

BUS310_L18_3.17.11

# BUS310_L18_3.17.11 - Inferential Statistics Hypothesis...

This preview shows pages 1–9. Sign up to view the full content.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Lecture 18: Inferential Statistics / Hypothesis Testing III Chapter 11 (Chi-Square Tests)
Outline Review Chapter 10: Hypothesis Testing with ANOVA: Compare means among 2 categories: Compare means among 3 or more categories: Are the means of three or more populations different? H1: Not all μj are equal H0: μ1 = μ2 = … = μc (the c population means are equal) Which means are different / better? ° Tukey-Kramer Procedure Are the means of population one and two different? H1: μ1 ≠ μ2 (the two population means are not equal) H0: μ1 = μ2 (the two population means are equal) Which mean is better? ° look up in descriptive statistics

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
New Material: Chapter 11 – Two categorical Variables: Testing for Equal Proportions 1. Both categorical variables have 2 categories: A 2X2 crosstab table a) Use approach in section 11-1. Use an existing table or create one with PHStat, ‘Descriptive Stats’, ‘Two-Way Tables and Charts’ b) Hypotheses: H0: π1 = π2 (the 2 populations have equal proportions/percents) H1: π1 ≠ π2 (the 2 populations have unequal proportions/percents) a) Obtain P-value from PHStat, ‘Two-Sample Tests’, ‘Chi Square Test for Differences in Proportions’ (Copy and paste your data into the worksheet created by PHStat) b) Conclusion: state your confidence in the alternative hypothesis , decide which hypothesis to conclude. 2. One categorical variable has c (>2) categories and one has 2 categories. a) Use approach in section 11-2. Use an existing table or create one with PHStat, ‘Descriptive Stats’, ‘Two-Way Tables and Charts’ b) Hypotheses: H0: π1 = π2 = … = πc (the c populations have equal proportions/percents) H1: Not all πj are equal ( the c population proportions are not all equal) a) Obtain P-value from PHStat, ‘Multiple-Sample Tests’, ‘Chi Square Test’ (Copy and paste your data into the worksheet created by PHStat). Make sure your table is a 2 X C table and not a C X 2 table.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Example with 5 groups: ‘BEER.xls’ 1. Run an ANOVA to get the p-value. We are almost 100% confident that there are differences among the mean prices for the 5 beer types in the population (e.g., beers available in the US), based on our samples.
Tukey-Kramer: Where are Differences in the Population? The population mean of Type 2 (Craft Ale) is higher than that of 1 (Craft Lager), 4 (Regular), 5 (Low Calorie) (meaning: Craft Ale is typically more expensive in these other types) Type 3 (Import) is typically more expensive than 4 (Regular) and 5 (Low Calorie) Type 1 (Craft Lager) is typically more expensive than 4 (Regular) 1. Tukey Kramer Test shows you which population means are different , with 95% confidence 2. If you want to know which population means are higher , look at the sample means: 4: Regular \$4.50 \$5.50 \$6.50 \$7.50 5: Low Calorie 1: Craft Lager 3: Import 2: Craft Ale 3. Combine the 2 sources of information for your conclusion: According to the Tukey Kramer Test, we can be 95% confident that:

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Example: Different Tests for the Population of Shidler Students Collect data for a sample of 30 Shidler students: GPA Numerical Gender
This is the end of the preview. Sign up to access the rest of the document.
• Spring '08
• Abou-sayf,F
• Statistical hypothesis testing, significant difference, PHStat, significant relationship, shidler students

{[ snackBarMessage ]}