STAT305/605 Fall 2020 SFU Midterm Due October 23rd 5:00PM PST
and inclusive? You may suggest modifications of the operationaliza- tion of the variable participation, or modification of the recruitment or the control or detail a matched-paired design (no marks for reference to new statistical tests or literature review: Valid arguments are re- quired, sound arguments are not required). Answer in two sentences.
(2 points)
Definitions for part c) of this Problem:
• A young person’s socioeconomic background includes a measure of the
income and occupation of their parents, and aspects of the community and household that they grew up in (Townsend et al. 1988).
• A factor is external to someone if it is something that a↵ects them, but also something that they have no control over.
• Post-secondary studies includes pursuit and completion of college or university degrees, and post-secondary attainment indicates the extent of such pursuits.
d) Suppose that your colleagues want to provide a one-sample t-test against the null hypothesis that the mean of the Y values listed in question a) of this Problem is equal to a given value μ. (They may want to test if the mean participation of the control class is di↵erent from a Canada-wide mean participation). A t-test with such a low value of m with m † 15 is only indicated if the data look normally distributed. The Shapiro-Wilk hypothesis test for normality provides a p-value for the null hypothesis that a collection of values Y1, . . . , Ym are normally distributed (Shapiro and Wilk, 1965). This test can be applied using the R code shapiro.test(Y), where Y is a variable in R specifying a vector with coordinates Y1, Y2, Y3, . . .. (Such a variable may be created in R with the code Y = c(10, 0, 4, …)). Does the Shapiro-Wilk test reject the null hypothesis that Y is normally distributed at alpha level 0.05, for the Y provided in part a) of this
Problem? And so, is the described test for the mean of Y indicated? (2 points)
9 of 9