Version 7 Documentation Errata

Below is a listing of errors found in the RATS Version 7 User's Guide and Reference Manual. Most of these have been fixed in the PDF files shipping with the current versions, but are still present in the hard-copy manuals currently being shipped. See Version 6 Errata and Version 5 Errata for errors in the Version 6 and 5 manuals, respectively.

---

RATS 7.2 Supplement PDF

• STWISE Instruction:
  
  Lists the "general to specific" option as being a switch option:
  

  GTOS/[NOGTOS]
  

  This is incorrect--GTOS is actually a choice on the METHOD option.

  METHOD=GTOS
  

Reference Manual

• CMoment, pages 44 and 46:
  References to ZMATRIX should read ZXMATRIX (which is the correct option name) instead.

• CVModel, page 78:
  Typo in the CVMODEL instruction shown on that page. The V option should read V=VF, as shown below, rather than V=V.

  cvmodel(method=bfgs,v=vf,obs=tdim) rr

• DDV, page 96:
  Typo in formula for Pseudo-R2. Should be (Log L/Log Lc) rather than (Log Lc/Log L).

• DLM, page 120:
  Description of VHAT and SVHAT options should have included the word “errors” at the end of the argument descriptions.
  Page 123:
  Reference to deprecated “SCALE” option in description of the reserved variable %VARIANCE. This should refer to the newer VARIANCE=CONCENTRATED option instead.

• EQUATION, page 143:
  The description of the REGRESSORS option references the “MORE” option, which was the old name for the option. The text should just refer to “this option”, meaning the REGRESSORS option.

• HISTORY, page 234:
  Refers to the deprecated start parameter as the default value for the newstart parameter. Should instead refer to the default value of the FROM option.

• LINREG, page 276:
  LINREG no longer computes Q-statistics (to reduce computational overhead), so %QSIGNIF and %QSTAT should not be listed in the Variables Defined section. Note: Use CORR(QSTATS) to compute Q stats.

• MAKE, page 288, 289:
  On page 289, PANEL is a switch option, not a value option, so this should be shown as PANEL/[NOPANEL]. MAKE(PANEL) takes a panel-format series and produces a matrix as described. On page 288, MAKE does not define %NUMOBS. Beginning with Version 7.10, it does define %NOBS as the number of observations.

• MRESTRICT, page 308:
  “Variables Defined” list should include: %CDSTAT (the computed test statistic) and %SIGNIF (the marginal significance level).

• PRJ, page 361:
  The XVECTOR option description should refer to Xi, not Sigma.

• REPORT, page 382:
  Typo on the ALLOCATE instruction. Should read “allocate 2000:4” not “allocate 200004”

• SPECIFY, page 430:
  SPECIFY now has a NOPRINT option for suppressing output.

• STORE, page 447:
  
  1. “See Chapter 2 for more details.” Correction: Should say “See Chapter 2 of the User’s Guide for more details.”
  2. SKIP option documented incorrectly. It is actually a SKIPLINES option designed to skip header lines in a spreadsheet. See SKIPLINES description on DATA instruction for the correct description.

• THEIL, page 476:
  “Two additional parameters used in older versions, steps and data end, have been replaced by options of the same name.” Correction: The new options are actually called STEPS and TO.

• Section 2, page 511:
  First line refers to “built-in instructions”. Should say “built-in functions”.

• %SUM() Function, page 560:
  Rather than being limited to the ALLOCATE range, %SUM() will use the full defined range of the series (which may be longer than the ALLOCATE range), or the range set an SMPL instruction.

User’s Guide

• Chapter 2, Page 90:
  The line:

  open dbase dsn

  should read
  

  open odbc dsn

• Chapter 12, Page 420:
  Equation (12) is correct, but in the following paragraph, the indicator function case is incorrectly listed as being for u>0. This should read u<0 instead (to match the formula).

• Chapter 12, Page 445:
  The code is correct, but the formula for the A matrix has the 1 and the 0 in the second row in the wrong order. That second row should be (1,0), not (0,1).

• Chapter 12, Page 452:
  Several typos in the second block of sample code. The COMPUTE statements in SimpleRegimeF should set "SimpleRegimeF", not "F". Also, the right parentheses at the end of the two COMPUTE lines should be deleted. The left parenthesis before "pstar" on the next to last line should also be deleted.

• Chapter 12, Page 454:
  Note that MSVARF must be defined prior to this code by the user.

• Chapter 12, Page 456:
  The LINREG and COMPUTE statements in the middle block of code should reference nlags and nlags+1, rather than fixed values of 4 and 5, respectively. Revised code is:

  linreg g
  # constant g{1 to nlags}
  compute phi=%xsubvec(%beta,2,nlags+1)
  

---

Copyright © 2013 Estima. All Rights Reserved.
This file was last modified on 12/30/11