ECON 3740 Lecture Notes - Lecture 4: Null Hypothesis, Econometrics

• School: University of Guelph
• Department: Economics
• Course: ECON 3740
• Professor: David Prescott

Document Summary

D prescott, dept. of economics & finance, university of guelph. F-tests are used to test multiple restriction hypotheses in linear regression models. A multiple- restriction hypothesis involves more than a single = sign. Consider the following model of wages that is estimated using survey data (cross section data. ) (cid:1831)(cid:4666)(cid:1849)(cid:4667)(cid:3404)(cid:2010)(cid:2869)(cid:3397)(cid:2010)(cid:2870)(cid:1858)(cid:3397)(cid:2010)(cid:2871)(cid:1827)(cid:3397)(cid:2010)(cid:2872)(cid:1827)(cid:2870)(cid:3397)(cid:2010)(cid:2873)(cid:1827)(cid:2871)(cid:3397)(cid:2010)(cid:2874)(cid:1831)(cid:3397)(cid:2010)(cid:2875)(cid:1831)(cid:2870)(cid:3397)(cid:2010)(cid:2876)(cid:1831)(cid:2871)(cid:3397)(cid:2010)(cid:2877)(cid:1842) (cid:4670)(cid:1827)(cid:4671) Here, a and its powers refer to age measured in years while e measures years of formal education. The variable f is a dummy variable that takes the value 1 if the survey respondent is female and zero otherwise. P is also a dummy variable; it equals 1 if the respondent works in the public sector and is zero otherwise. It is common practice to test single-restriction hypotheses using t-statistics. Examples include: (cid:4670)(cid:883)(cid:4671) (cid:1834)(cid:2868): (cid:2010)(cid:2873)(cid:3404)(cid:882) (cid:4670)(cid:884)(cid:4671) (cid:1834)(cid:2868): (cid:2010)(cid:2876)(cid:3404)(cid:882) (cid:4670)(cid:885)(cid:4671) (cid:1834)(cid:2868): (cid:2010)(cid:2870)(cid:3404)(cid:3398)(cid:2010)(cid:2877) Note that both [1] and [2] have a single = sign.