The number of columns taken up by each response in your data file (1 or 2 or more). If possible, enter your data one column per response. If there are two columns per response, make XWIDE=2. If your data includes responses entered in both 1 and 2 character-width formats, use FORMAT= to convert all to XWIDE=2 format. When XWIDE=2, these control variables require two columns per item or per response code: CODES=, KEYn=, KEYSCR=, NEWSCORE=, IVALUE=. Either 1 or XWIDE= columns can be used for RESCORE=, ISGROUPS=, RESCORE= and IREFER=
See Example 11, Item responses two characters wide.
Example 1: The responses are scanned into adjacent columns in the data records,
XWIDE=1 ; Observations 1 column wide
CODES=01
Example 2: Each response is a rating on a rating scale from 1 to 10, and so requires two columns in the date record,
XWIDE=2 ; 2 columns per datum
CODES=010203040506070801011 ; 2 CHARACTERS FOR EACH OBSERVATION
or
CODES=" 1 2 3 4 5 6 7 8 91011" ; 2 CHARACTERS FOR EACH OBSERVATION
Example 3: Some responses take one column, and some two columns in the data record. Five items of 1-character width, code "a", "b", "c", or "d", then ten items of 2-character width, coded "AA", "BB", "CC", "DD". These are preceded by person-id of 30 characters.
XWIDE=2 Format to two columns per response
FORMAT=(30A1,5A1,10A2) Name 30 characters, 5 1-chars, 10 2-chars
CODES ="a b c d AABBCCDD" "a" becomes "a "
NEWSCORE="1 2 3 4 1 2 3 4 " response values
RESCORE=2 rescore all items
NAME1=1 person id starts in column 1
ITEM1=31 item responses start in column 31
NI=15 15 items all now XWIDE=2
Example 4: Scores are entered as hundreds, but are to be analyzed as incrementing by 1
XWIDE = 4
CODES = " 100 200 300 400 500 600 700 800 9001000"
STKEEP = No ; analyze the ascending scores as ordinal levels
For more examples, see CODES= and Decimal, percentage and contiuous data.
