1. (40) Use the data from the Canadian Tobacco Use Monitoring Survey, 2010. Test the hypothesis that, among males, the proportion of individuals who smoke every day is higher than the proportion among females. Use 3% significance level; use both the critical value and p-value approach. Show all your work! So, is gender an important determinant of smoking behaviour? Explain.
2. (60 points) Using the data on age at immigration and employment income:

(i)(30 points) Use ANOVA analysis to test the hypothesis that mean employment income is the same across all 5 groups (created based on age at immigration). Do the analysis manually (not the ANOVA command in Excel; you should use excel to calculate means, variances, etc. though). Make sure that you report each step/calculation of yours. State the null and alternative hypotheses and explain your decision about accepting/rejecting the null. Use 95% confidence level.

(ii)(10 points) Perform ANOVA analysis in Excel using 'Data - Data Analysis - ANOVA: one way'. Report the Excel output and use it to test the hypothesis of equality of the means. Explain your decision about accepting or rejecting the null.

Data Instructions

Question 1

Go to Odesi. In the Health category, choose the Canadian Tobacco Use Monitoring Survey 2010. Select the Annual, Person File. Add the following variables to the subset:

• Smoke cig daily, occasionally, never [in Variable Description, Smoking status]
• Respondent's sex [in Variable Description, Demographic Variables]

Download the subset and transfer it to Excel. Divide the observations into two groups based on sex: group one are males and group 2 are females. Compute the proportion of daily smokers (smoke every day; use the documentation file) for each of the groups [Hint: to make your calculations simpler, use the COUNTIF command in Excel].

Question 2

Go to Odesi.

In Social Surveys, select National Household Survey (NHS), 2011, Individuals File. Add the following variables to the subset:

• Age at immigration (in Variable Description, Place of birth and Generation Status
• Employment income (in Income)

Download the data and transfer them to Excel. Sort the observations by Employment income and delete observations with negative income as well observation with income of 200,000 and more (i.e. outliers and missing observations). You should be left with 720,989 observations. Divide your sample into 6 groups based on the age at immigration:

Group 1: 20-24 years old at immigration (category 5 in the age at immigration variable)

Group 2: 25-29 (category 6)

Group 3: 30-34 (category 7)

Group 4: 35-39 (category 8)

Group 5: 40-44 (category 9)

Delete the observations outside of these 5 groups.