** Chapter 4.7 ** for W.H. Greene, Econometric Analysis 6th ed. *****************
* (c) Noel Roy 2003, 2008
*
*             THE NORMALITY ASSUMPTION AND BASIC STATISTICAL INFERENCE
*                 4.7.1 Testing a Hypothesis about a Coefficient
*
* The tutorial for this section will review the following tasks in SHAZAM:
*   •	Generating dummy variables (DUM command)
*   •	Excluding observations (SKIPIF command)
*   •	Calculating the covariance matrix of OLS estimates (OLS / PCOV command)
*
*===============================================================================
*
* Example 4.3 (p. 53) Earnings Equation
*
SAMPLE 1 753
READ (TableF4-1.txt) LFP WHRS KL6 K618 Age Educatn WW /SKIPLINES=37
*
* The data file consists of 753 observations on 19 variables. However, only 
* the first 7 variables are needed to estimate the earnings equation, so these
* are all we need to read. SHAZAM will read the first seven variables on each
* line, then drop to the next line, until 753 lines are read (unless an error
* is encountered). However, the block of variables to be read must be contiguous
* and the first set of variables on the line.
*
* LISTing a dataset this large is obviously impractical. We can do some checking
* of the data using a STAT command, to verify that the data "look right".
* The /ALL option lists STATs for ALL the data.
*
STAT /ALL
*
* The earnings equation only uses data for women with a labour force attachment
* (i.e., LFP is 1 rather than 0). Observations that do not satisfy this 
* condition can be skipped in estimation procedures by the SKIPIF command. An 
* observation is skipped if the expression contained in the SKIPIF command is 
* positive.
* By default, SHAZAM prints a warning when an observation is skipped. In this 
* data set, with 300 skips, it would be desirable to turn off the warnings.
*
SET NOWARNSKIP
SKIPIF (1-LFP)
*
* The dependent variable is the log of earnings, where earnings are defined as
* hourly wage (WW) times hours worked(WHRS). The explanatory variables are age
* (Age), age squared, education (Educatn), and a binary variable Kids reflecting the 
* presence of children under 18 in the household. These can be created where
* necessary using GENR commands. The variable Kids will be constructed using 
* the DUM(x) function, which takes a value of 1 for positive x and 0 otherwise.
*
GENR LNEARN=LOG(WW*WHRS)
GENR Age2=Age**2
GENR Kids=DUM(KL6+K618)
*
* The estimated covariance matrix for the OLS estimates will be printed if the
* OLS command is run with the PCOV option. The LOGLIN option corrects the
* calculated elasticities and the log-likelihood function for the fact that
* the dependent variable (but not the independent variables) is in LOG form.
*
OLS LNEARN Age Age2 Educatn Kids / PCOV LOGLIN
STOP
*
*===============================================================================
*
* Updated August 27, 2008