1996 Time Series Study
Updates & Errata
Current Dataset Release: March 2, 2001 Version
The 1996 Time Series Study codebook appendix section “Sampling Error Computation Models” identifies ‘2125’ as the variable for the the Sampling Error Stratum code. The correct variable is V960122. Thanks to Rosalia Greco for identifying the incorrect reference.
May 26, 2015
The variable label for V960005B contains an error. It should read as follows:
96PR: NEW TIME-SERIES WEIGHT (POST).
The following syntax can be used to correct the label in the actual data sets.
SPSS: var lab V960005B "96PR: NEW TIME-SERIES WEIGHT (POST)". SAS: label V960005B = "96PR: NEW TIME-SERIES WEIGHT (POST)"; run; Stata: label variable V960005B "96PR: NEW TIME-SERIES WEIGHT (POST)"
- Variables documenting the randomized order of the Post Presidential candidate thermometers (V961019,V961020,V961021) were omitted from the release data file. These randomization variables may be included in a future re-release.
- Variables V960693a,V960694a,V960695a,V960696a,V960697a were omitted from the release data for this study. These variables describe recent (last 6 months) occupation of nonworking homemaker/student spouses as folllows:
V960693a: 71-category collapse of Census occupation codes; V960694a: 14-category collapse of Census occupation codes; V960695a: prestige score for occupation (blanked for release); V960696a: industry of occupation V960697a: hours worked
23 cases are affected. Data (excluding V960695a) and documentation for these variables are provided here .In addition, the 0 inap label for variables V960693,V960694,V960695,V960696,V960697 is incorrect and should be:
Inap, 2,3,4,5,9 in Y2; 70,75,80,99 in Y17(1); 5,8,9 in Y17a
ANES 1996 codebook marginals exclude counts for cases coded “DK”. Most or all variables with marginals are affected. The codebook will be corrected in the next re-release of the 1996 study data.
Variables related to the special battery on Group Membership which were provided by analysts who have reviewed responses at mention level were merged with 1996 data using an incorrect link. Previously these variables were available in 1996 Pre-Post releases and are now available only in the mergeable Auxiliary File on Group Membership, which was corrected as of 2003-Aug-13. In addition, the documentation for variable V961454, which is included in both the new Auxiliary file and in the 1996 Pre-Post file, was incorrect; description of the variable has been corrected in the codebooks and variable labels for the Auxiliary File but are not yet corrected in the 1996 Pre-Post codebook or variable labels.
March 2, 2001
The 1996 Time Series Study data release of March 02, 2001 includes these corrections, among others:
- addition of descriptive variables for group variables
December 7, 2000
Variable V961322 in the 1996 codebook has the labels for codes 4 and 5 reversed.
Variable V961194 code 5 label is incorrect and should omit the word “not.”
Previous Dataset Release: May 21, 1999 Version
May 21, 1999
The 1996 Time Series Study data release of May 21, 1999 includes these corrections, among others:
- added dataset ID and version number
Previous Dataset Release: January 19, 1999 Version
January 19, 1999
The 1996 Time Series Study data release of January 19, 1999 includes these corrections, among others:
- one erratum was applied
Previous Dataset Release: March 1998 Version
January 19, 1999
V961114-961115 (Clinton gets things done) and V961116-V961117 (Dole moral) were switched in the 1st and 2nd releases of the 1996 data.
Users should amend their card files and codebook such that V961116-961117 document “Clinton gets things done” and V961114-961115 document “Dole Moral.”
The third release of the 1996 data (March 1998) differs from its predecessor in the following ways:
1. NEW TIME SERIES WEIGHTS HAVE BEEN ADDED. These weights are provided as an alternative to the time-series weight available in release 2; there is now a new weight for 1996 Pre variables and a new weight for 1996 Post variables.
2. CHANGES TO SOME MISSING DATA CODES. It was discovered that the CAPI application output of data did not always preserve cases which were coded “DK;” for several variables (V961189-961192) this resulted in a large number of blank or NA cases. The DK values have been restored to these cases.
In addition, the following lesser changes also were effected:
- One case with an incorrect 1994 case ID has been corrected.
- Several variable label corrections have been made.
- The Candidate List appendix has been placed in Appendix H and Appendix G is now the full technical report for the additional weight variables
- CPS 2-digit Occupation Master Codes, which are required for the occupation coding of R’s father, has been included with the Appendix E set of 1996 Master Codes. (This Master Code was omitted in the codebook included with release 2.)
- Codebook documentation for the new weight variables has been added.
Previous Dataset Release: May 19, 1997 Version
May 19, 1997
Five revised files were put on the ANES downloading page on May 19, 1997.
1. The revised version of appndcbk.txt includes the codes for Appendix E, Master Code 3 (Party-Candidate) that were added for the 1996 study. Some of these codes were inadvertently omitted from the earlier version of this file.
2. The revised versions of the SAS and SPSS data definition statements (nes1996.sas and nes1996.sps) include corrections to the variable labels for V960066 and V960067 (which should refer to, respectively, respondent gender and race, not interviewer gender and race).
3. The revised version of the raw (ASCII) data file (nes1996.dat) includes corrections to one case which had incorrect data in a number of 1994-related variables and another case which had incorrect respondent gender. Users of SAS might wish to run the following recode to fix the old version of the data:
if V960001=1378 then do; V960007=2239; V960009=0; V960010a=2; V960010b=12; V960106b=0; V960122a=372; V960131=4; end; if V960001=3377 then V960066=2;
4. The revised version of the Pre codebook (precbk.txt) includes new marginals for the variables changed in the data, as described above.