CDF Instruction

CDF( options ) distribution statistic degree1 degree2

Computes the marginal significance level of a statistic from one of four distributions.

Parameters

distribution	Choose from FTEST, TTEST, CHISQ, or NORMAL. Selects the distribution to use for the computation. The significance levels for TTEST and NORMAL are both for two-tailed tests.
statistic	the test statistic
degree1	degrees of freedom (TTEST, CHISQ), numerator degrees of freedom (FTEST)
degree2	denominator degrees of freedom (FTEST)

Options

[PRINT]/NOPRINT

TITLE="title for output" [none]

Variables Defined

%CDSTAT	the computed test statistic (REAL)
%SIGNIF	the marginal significance level (REAL)

Notes

CDF is a convenient way to display the significance level, but all the four distributions have functions (shown below) which compute those as well, so you can display the result some other way, such as with a DISPLAY or REPORT.

FTEST	%FTEST(statistic,degree1,degree2)
TTEST	%TTEST(statistic,degree1)
NORMAL	%ZTEST(statistic)
CHISQR	%CHISQR(statistic,degree1)

Examples

These are from two of the standard examples. Note that the statistic parameter can be an expression rather than just a variable, although both of these compute and save that. For instance, the end of the second example could be written:

cdf(title="LR Test for Children in School Dummies") chisqr 2*(logunres-logres) 4

* CHOW1.RPF

* Chow test by subsample regressions

open data states.wks

data(org=col,format=wks) 1 50 expend pcaid pop pcinc

set pcexp = expend/pop

linreg(smpl=pop<5000) pcexp

# constant pcaid pcinc

compute rsssmall=%rss , ndfsmall=%ndf

linreg(smpl=pop>=5000) pcexp

# constant pcaid pcinc

compute rsslarge=%rss , ndflarge=%ndf

* Full sample regression

linreg pcexp

# constant pcaid pcinc

compute rsspool=%rss

compute rssunr=rsssmall+rsslarge , ndfunr=ndfsmall+ndflarge

compute fstat = ( (rsspool-rssunr)/3 ) / (rssunr/ndfunr)

cdf(title="Chow test for difference in large vs small") ftest fstat 3 ndfunr

* PROBIT.RPF

open data probit.dat

data(org=obs) 1 95 public1_2 public3_4 public5 private $

years teacher loginc logproptax yesvm

* Linear probability model

linreg yesvm

# constant public1_2 public3_4 public5 private years teacher loginc logproptax

ddv(dist=logit) yesvm

# constant public1_2 public3_4 public5 private years teacher loginc logproptax

ddv(dist=probit) yesvm

# constant public1_2 public3_4 public5 private years teacher loginc logproptax

* Test whether "Children in School" dummies are significant.

* Use "Wald" test first.

exclude(title="Wald Test of Children in School Dummies")

# public1_2 public3_4 public5 private

* Likelihood ratio test. We already have the unrestricted model.

compute logunres = %logl

ddv(dist=probit) yesvm

# constant years teacher loginc logproptax

compute logres=%logl

compute lratio=2*(logunres-logres)

cdf(title="LR Test for Children in School Dummies") chisqr lratio 4

Sample Output

LR Test for Children in School Dummies

Chi-Squared(4)= 6.051680 with Significance Level 0.19532186