|
Hello Listserv Participants!
Does anyone have advise to share about how to "stack" imputed data sets? I am going to use Mplus to run a SEM analysis and to prepare, want to set up the data set so that each case has, for each variable, the original data value and each imputed value (five imputed data sets) one after the other in one row. I tried a variety of sorting and transposing schemes, and still no stacked data set. Please advise. Drop-down window advise esp. welcome. Thanks so much! Julia ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
|
Hello Julia,
Merging data/add variables is the proper method. You can find it in menu, under Data/merging files. First prepare your original data and imputed data sets - all files must have an ID variable (the same name in every file), data set sorted by ID and with unique values in ID. The imputed variables will have different names in every file, something like var1 in original, var1_1 var1_2 ... in imputed data sets. Then open your original data set, go to Menu/Data/Merge files/select Add Variables, ID is the key variable. Browse and select the imputed data set to be merged with original. This should create desired data set with both original and imputed variables. With syntax it is: MATCH FILES /file = OriginalData /file = ImputedData1 /by ID. EXEC. You can sort your variables in target data set using keep option: MATCH FILES /file = OriginalData /file = ImputedData1 /by ID /keep var1 var1_1 var1_2 ... var2 var2_1 var2_2 etc, list of all variables in interest in the order you want them in the merged file. The last is not a must, only a trick to have variables sorted in the way you will use them and make the work with final data set easier. HTH Jindra > ------------ Původní zpráva ------------ > Od: Julia Alexander <[hidden email]> > Předmět: stacking imputed datasets > Datum: 25.1.2010 19:24:24 > ---------------------------------------- > Hello Listserv Participants! > > Does anyone have advise to share about how to "stack" imputed data sets? I > am going to use Mplus to run a SEM analysis and to prepare, want to set up > the data set so that each case has, for each variable, the original data > value and each imputed value (five imputed data sets) one after the other in > one row. I tried a variety of sorting and transposing schemes, and still no > stacked data set. Please advise. Drop-down window advise esp. welcome. > Thanks so much! > Julia > > ===================== > To manage your subscription to SPSSX-L, send a message to > [hidden email] (not to SPSSX-L), with no body text except the > command. To leave the list, send the command > SIGNOFF SPSSX-L > For a list of commands to manage subscriptions, send the command > INFO REFCARD > > > ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
| Free forum by Nabble | Edit this page |
