Disjoint strings of responses

When the responses are not arranged in one continuous string in the record, instruct Winsteps to skip over or ignore the gaps.

 

Example: The 18 item string is in columns 40 to 49 and then 53 to 60 of your data file. The person-id is in columns 11-30. Data look like:

xxxxxxxxxxPocahontas Smith, Jrxxxxxxxxx1001001110xxx11001110

 

Method a: Delete unwanted "items" in columns 50, 51, 52 using an item delete file, IDFILE=.

  NAME1 =11 in original record

  NAMLEN=20 length in original record

  ITEM1 =40 in original record

  NI    =21 include deleted items

  IDFILE =DEL5052 file of deletions

 The contents of DEL5052 are:

  11 - 13  Cols 50-52 are items 11-13

 

Method b: Rescore "items" in columns 50, 51, 52 as missing values with RESCORE=.

  NAME1 =11  in original record

  NAMLEN=20

  ITEM1 =40

  NI    =21  include rescored items

  RESCORE=000000000011100000000 rescore 50-52

  CODES =01  (the standard)

  NEWSCORE=XX  non-numerics specify "missing"

 

Method c: Make the items form one continuous string in a new record created with FORMAT=. Then the item string starts in the 21st column of the new record. Reformatted record looks like: Pocahontas Smith, Jr100100111011001110

  FORMAT=(T11,20A,T40,10A,T53,8A) reformatting

  NAME1 =1  in the formatted record

  ITEM1 =21  in the formatted record

  NI    =18  the actual number of items