How to design a questionnaire

Lecture

What makes a good questionnaire?

• Discrimination
  "Do people who differ in the construct being measured obtain different scores on the questionnaire?"
  People with the same score on a questionnaire should be equal to each other on the measured construct.
  People with different scores on a questionnaire should be different to each other along the measured construct.
  The degree of difference between people should be proportional to the difference in scores. 
  
  

  Researchers tend to use 7 point versions these days to try and increase spread and avoid ceiling effects (Ceiling effects is when measurements of the dependent variable result in many very high or highest scores on the dependent variable, thus masking a potential effect of the independent variable. Ceiling effects may occur when giving examinations to students, when the exam is so easy that all the students correctly answered every test item).

• Validity
  Did you measure what you were supposed to measure? eg. does the empathetic concern scale really measure empathetic concern, or does it measure something else?
  1. Content Validity
  How represenative…