MGSC 372 Lecture Notes - Lecture 10: Heteroscedasticity, Minitab, Probability Plot

• School: McGill University
• Department: Management Science
• Course: MGSC 372
• Professor: Brian Smith

Document Summary

The terms are more spread out on the right side of the chart when compared to the left hand side of the chart: heteroscedasticity. Possible transformation to help eliminate or reduce heteroscedasticity are: There are many possible transformation of the form where can be any real exponent. Salary vs. years of experience for a random sample of 50 social workers. Regression equation: salary = 20242 + 522exp + 53. 0expsq (cid:2870)=(cid:890)(cid:883). (cid:888)% But the residual plot shows signs of heteroscedasticity. Minitab output: model including logarithmic transformation and quadratic term. Summary of second order model ln(salary) vs. exp and expsq. Regression equation: ln(salary) = 9. 84 + 0. 0497exp + 0. 000009expsq (cid:2870)=(cid:890)(cid:888). (cid:886)% (cid:2870)=(cid:890)(cid:887). (cid:890)% The residual plot shows that the log transformation has significantly reduce heteroscedasticity. But the coefficient of expsq is not significant (p-value = 0. 98). Note: same (cid:2870) value and higher (cid:2870) value.