Does a psychological scale need to be revalidated after reducing the number of response options from 9 to 4? I have used an established 33-item empathy questionnaire. 
I have modified the scale from a 9 point to a 4 point scale in order to force an answer and  reduce ambiguity. 
Do I need to retest the scale properties after this reduction in the number of response options?
 A: Whether you need to retest your instrument's validity and reliability now that you have changed the number of response options is a matter for you and your dissertation committee to decide; there is no standard statistical knowledge that can definitively answer the question for you.  You and/or your colleagues may find that, as you suggest, going from 9 to 4 reduces the arbitrariness with which respondents choose an option.  On the other hand, when you call the instrument "established" you are implying that it already had good psychometric properties the way it was, and so changing the method of measurement entails some risk.  Perhaps there are so few options now that the choice between, say, the first and the second has become more arbitrary.  Alternatively, perhaps the lack of a neutral option would seem unreasonable to some; the topic might require it.  If so, this would reduce reliability, and with it validity.
The bottom line is, if there is good evidence for reliability and validity at 9 options, the burden could easily be placed on you to show that these properties have improved, or at least not worsened, at 4 options.
