/*****************************************************************************/ /* SPSS User Files are being provided for the convenience of analysts who */ /* may wish to use SPSS to analyze the MEPS data. However, users are */ /* cautioned that SPSS does not currently have the capability to produce */ /* appropriate standard errors for estimates from a survey with a complex */ /* sample design like the Medical Expenditure Panel Survey (MEPS). The */ /* annual national probability sample of households selected for the MEPS */ /* is a subsample of those participating in the National Health Interview */ /* Survey (NHIS) for the previous calendar year. The NHIS, an ongoing */ /* general health survey of the civilian noninstitutionalized population, */ /* is conducted by the National Center for Health Statistics, Centers for */ /* Disease Control and Prevention. The NHIS sample design, and thus that */ /* of the MEPS, uses complex sampling techniques including */ /* stratification, clustering, and differential sampling rates for the */ /* oversampling of selected subgroups of the population. Because of this */ /* sampling structure, software that accounts for this complex sample */ /* design must be used. Without the use of appropriate software that */ /* accounts for the complex sample design of the MEPS in conjunction with */ /* the survey's sampling weights (reflecting the survey's sample */ /* selection probabilities and nonresponse and other adjustments), the */ /* estimated standard errors of survey point estimates will generally be */ /* underestimated, and, thus, the inferential conclusions in an analysis */ /* (descriptive or analytical) may be misleading. It is recommended that */ /* sample survey software be used to analyze the MEPS data. Software for */ /* analysis of complex survey data include, for example, SUDAAN (Research */ /* Triangle Institute), STATA (Stata Corporation), VPLX (U.S. Bureau of */ /* the Census), and WESVAR (Westat, Inc.). A more comprehensive summary */ /* of software for the analysis of data from surveys with complex sample */ /* designs can be accessed through the website for the American */ /* Statistical Association's Survey Research Methods Section. Users may */ /* also reference the Agency for Healthcare Research and Quality (AHRQ) */ /* web document "Computing Standard Errors for Medical Expenditure Panel */ /* Survey (MEPS) Estimates." */ /*****************************************************************************/ FILE HANDLE INFILE NAME="H220IF2.DAT" /LRECL=77. DATA LIST FILE=INFILE / DUPERSID 1-10 (A) RXLKIDX 11-42 (A) EVNTIDX 43-58 (A) LINKIDX 59-74 (A) EVENTYPE 75-75 PANEL 76-77 . VARIABLE LABEL DUPERSID "PERSON ID (DUID + PID)" RXLKIDX "RECORD ID: EVNTIDX + LINKIDX" EVNTIDX "EVENT ID" LINKIDX "ID FOR LINKAGE TO COND/OTH EVENT FILES" EVENTYPE "TYPE OF EVENT RX IS LINKED TO" PANEL "PANEL NUMBER" . VALUE LABELS EVENTYPE 1 "1 OFFICE-BASED MEDICAL PROVIDER VISIT" 2 "2 OUTPATIENT DEPARTMENT VISIT" 3 "3 EMERGENCY ROOM VISIT" 4 "4 INPATIENT HOSPITAL STAY" / PANEL 23 "23 PANEL 23" 24 "24 PANEL 24" 25 "25 PANEL 25" . EXECUTE.