INSTALLATION INSTRUCTIONS: 1) Download the missing value converter zip file from http://www.ssc.wisc.edu/wlsresearch/data/utilities/ 2) Unpack the zip to your target directory 3) Start Stata and add that directory to your ado-path by typing "adopath + path to target directory" If you unzip the utility to your working directory, you can skip this step. 4) For help, type "help wls2stata" Note: Converting the entire WLS dataset (by typing "wls2stata _all") takes a considerable amount of time. Consider saving the resulting dataset. Note: The program is written for Stata 12 or later. PROGRAM DESCRIPTION: The Wisconsin Longitudinal Study uses negative integers in the range -1 to -31 to describe missing values in numeric variables and string representations of these numbers (e.g. "-1") to describe missing values in text variables. The Stata program wls2stata will convert these negative codes to Stata missing values in the case of numeric variables and an empty string ("") for text variables. Using _all as the variable list will perform this operation for all variables in the dataset, Since the number of negative codes used in the WLS exceeds the number of missing codes available in Stata, bracketing summary variables, which use codes -6 though -10, are NOT recoded to a Stata missing code. As of the WLS data release 13.01 all variables that are known to contain negative data other than missing codes were reviewed. Variables with valid data within the negative code range are processed so that no valid responses are overwritten with a missing code. Users can also add their own list of variables that should not be recoded through the norecode option. The negative codes are assigned missing value codes as follows: -1 -> .d (Don't know) -2 -> .i (Inappropriate) -3 -> .r (Refused) -4 -> .n (Not ascertained) -5 -> .p (Partial interview) -27 -> .s (Not asked of Proxy) -28 -> .t (Not asked via telephone) -29 -> .b (Left Blank on SAQ) -30 -> .m (Not part of MOSAQ) -11 -> .a (At most part time worker) -12 -> .c (At least full time worker) -13 -> .e (Amount Differs For Generic/Brand-Name) -14 -> .f (Amount Changes Over The Year (Vol)) -15 -> .g (Multiple Sessions) -16 -> .h (Never went to usual provider) -17 -> .j (Special Missing Code for Allocation Cases) -18 -> .k (Unable to Code (ICD-9 related questions)) -19 -> .l (Occupational scores not assigned to military occupations) -20 -> .o (Amt varies) -21 -> .q (Pays all costs not covered by the plan) -22 -> .u (Different payment arrangement) -23 -> .v (Attended a HS in Wisconsin, did not graduate from that HS) -24 -> .w (Volunteered, < 1 drink per day) -25 -> .x (Only drank at religious services) -26 -> .y (Unable to code, multiple people mentioned) -31 -> .z (Not yet used) In addition to recoding the values, wls2stata copies over value labels so that the system missing values have labels formerly corresponding to the negative codes. For syntax, options and other details type "help wls2stata" from within Stata.