144 Views

Download Presentation
##### Economics 310

**An Image/Link below is provided (as is) to download presentation**

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.

- - - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - - -

**Economics 310**Lecture 12 Heteroscedasticity**Economics 212**Lecture 23 - Heteroscedasticity**Heteroscedasticity**• Violation of classic assumption of constant variance of disturbance. • The variance of the disturbance may be different for some or all of the subpopulations. • Subpopulations with large variances are not as helpful in estimating our model as subpopulations with small variances.**Sources of Heteroscedasticity**• Error-learning models • Discretionary Income • Improved Data Collection techniques • Outliers**Example of Heteroscedasticity-Typing Test**Density Weeks WPM**Example Heteroscedasticity - Faculty Salaries**_______________________________________________________________11/19/1998_10:25_ FILE: box and whisker for faculty salaries, NO. OF VARIAB(MISS. CASES: 0) E LABEL: none ________________________________________________________________________________ BOX AND WHISKER PLOT VARIABLE: Full PLOT: ----------XXXXXX¦XXXXXX----------------- VARIABLE: Assoc. PLOT: ------XXX¦XXXX--------- o VARIABLE: Asst. PLOT: -----XX¦X------- o ¦--------------¦--------------¦--------------¦--------------¦ 35 51 66 81 97**Data for GLS Example**State pop autopc incomepc California 32.268 0.48 25.368 Florida 14.654 0.50 24.198 Indiana 5.864 0.54 22.633 Maine 1.242 0.46 21.087 Mississippi 2.731 0.46 17.561 New Hampshire 1.173 0.63 26.772 North Dakota 0.641 0.52 20.476 Rhode Island 0.987 0.51 24.613 Utah 2.059 0.40 19.384 Wisconsin 5.170 0.48 23.390**Results of unweighted regression**R-SQUARE = 0.4326 R-SQUARE ADJUSTED = 0.3617 VARIANCE OF THE ESTIMATE-SIGMA**2 = 0.23376E-02 STANDARD ERROR OF THE ESTIMATE-SIGMA = 0.48348E-01 SUM OF SQUARED ERRORS-SSE= 0.18700E-01 MEAN OF DEPENDENT VARIABLE = 0.49800 LOG OF THE LIKELIHOOD FUNCTION = 17.2196 VARIABLE ESTIMATED STANDARD T-RATIO PARTIAL STANDARDIZED ELASTICITY NAME COEFFICIENT ERROR 8 DF P-VALUE CORR. COEFFICIENT AT MEANS INCOMEPC 0.13807E-01 0.5590E-02 2.470 0.039 0.658 0.6577 0.6251 CONSTANT 0.18669 0.1270 1.470 0.180 0.461 0.0000 0.3749**Results of Weighted Regression**R-SQUARE = 0.0749 R-SQUARE ADJUSTED = -0.0407 VARIANCE OF THE ESTIMATE-SIGMA**2 = 0.11092E-02 STANDARD ERROR OF THE ESTIMATE-SIGMA = 0.33305E-01 SUM OF SQUARED ERRORS-SSE= 0.88738E-02 MEAN OF DEPENDENT VARIABLE = 0.48946 LOG OF THE LIKELIHOOD FUNCTION = 17.0600 VARIABLE ESTIMATED STANDARD T-RATIO PARTIAL STANDARDIZED ELASTICITY NAME COEFFICIENT ERROR 8 DF P-VALUE CORR. COEFFICIENT AT MEANS INCOMEPC 0.43115E-02 0.5357E-02 0.8048 0.444 0.274 0.2737 0.2123 CONSTANT 0.38555 0.1295 2.976 0.018 0.725 0.0000 0.7877