My questionnaire is about tracking survey of the higher graduated and it includes 4 sections and 163 items. In each section the nominal, dichotomous, interval and Likert scale is specified. We have also some multiple-choice and open-ended questions. What techniques can we use to reduce the number of items? how could we score these items? And for assessing the reliability, although I know it is different for distinct scales, but does it check all the sections together or each section separately?
Please comment to me with reference.