Multiple affective subtests

Here's a common example of CCs lines for a survey with three subscales:


*col (c5-c20)

*sub aff, title=(Anxiety)

*pol ++--+ -++++ ++-++

*col (c21-c35)

*sub aff, title=(Friends)

*col (c36-c50)

*sub aff, title=(Homesick)

*pol ---++ +++-- -+-++


Three 15-item affective subtests are defined by these lines.  Two of the three subtests, the first and the third, have a mixture of forward- and reverse-scored items.  There is no *pol line for the second subtest, which means that all items for this subtest are forward-scored.


Lertap will make three subtest scores, and also a total score.


What about possible score ranges for this example?  Each subtest has 15 items.  There being no Res= declaration on the *sub lines, Lertap assumes Res=(1,2,3,4,5), that is, five possible responses per item.


Lertap will score each item on a one- to five-point basis.  Why?  Because there are five possible responses.  The minimum score a person can get on an item is one; the maximum is five.  There are 15 items in each subtest.  Therefore, the score range for each subtest is 15 to 75, and, there being three subtests, the range for the total score will be 45 to 225.


What happens when a person doesn't answer an item?  What sort of score do they get?


For cognitive tests, a non-answer gets a score of zilch (zero).  However, for affective items, a non-answer will get a score equal to the mean of the item's response weights (note).  It is possible to defeat this scoring system by using the MDO control word on a *sub line.


Note that it is possible to achieve almost any sort of scoring for affective items, or, for that matter, cognitive items.  This is done by using *mws lines in the CCs worksheet.