March 13, 2012
• Tutorial Q and A cancelled because of lack of attendance.
• Students can go meet one-on-one instead
Theme: How work gets to us
• How often during the past month have you felt used up at the end of the day?
• Context: face-to-face interviews.
How can we turn this interval variable to into a continuous variable?
• Change the categories into days. Example: how many days did you feel
How to turn interval variable into a nominal variable. Example: How often
do you feel exhausted?
• Categories: Low, High
Hypothesis: As level of work stress increase, there will be higher level of exhaustion.
Null: Work stress has no impact to do with exhaustion.
Chi-square: test of independence
Work stress and exhaustion can be independent.
SOC202H1S Quantitative Analysis in Soc. Science. March 13 2012. Scott Schieman. SP. TG. 2
How often do you find your work stressful?
The categories of hardly ever and always are taken to
Chi-square test: example 1
• The relationship between tow nominal/categorical variables
• Cross-tabulation table of the frequency of join occurrences; observed
frequencies; marginal totals and appropriate percentages
Independent variable: work stress. Need percentage of independent variable. What
percentage of low work stress has high exhaustion?
In order not confuse yourself, say the percentage of that column.
• 75% of people with high work stress report high levels of exhaustion compared
to 27% of those with low work stress report high levels of exhaustion.
• The variables are not independent. If independent the levels of exhaustion
would be the same regardless of the amount of work stress.
Need to obtain expected frequencies: these are the values of the joint cell
frequencies if null hypothesis is true.
• The expected values if the null is true
• Observed subtracted expected to see the difference.
SOC202H1S Quantitative Analysis in Soc. Science. March 13 2012. Scott Schieman. SP. TG. 3
According to the chi-square chart
• The number 236.8 is the expect number of people in high stress and high
exhaustion category if null is true.
• The observed number in the survey is 417 in the high stress and high
• Better case against null hypothesis is a larger difference between observe and
Possible test question: The df in the chi square will be blotted out and student
must fill in the df. Then the numbers in the equation will change, student should be
able to deductively state the new df.
• Build a vas against the null hypothesis: The larger the differences between
expected and observed frequencies, the better the case against the null