Summarizing Markless on surveys:
A major pitfall of surveys is that they “often project designers’ preconceptions out to potential respondents” and thereby dilute the value of the responses. General tips: Give clear instructions for answering the questions! Don’t use library/technology jargon! Some types of questions we can use:

Closed questions: Questions with several answer choices (with check boxes), and with a catch-all at the end.
e.g. I will use the LIS website to:
Find resources X
Get technology help X (etc….)
Anything else? Please say what! X (with space for answer)
Open questions: May best follow closed questions, i.e. “Why is this?”
Simple checklists (with check boxes), and a catch-all at the end:
e.g. Why did you use the website today? Check all that apply
Find resources X
Look for information X
Get help from a librarian X (etc…)
Other, please say what! (with space for reply) X
Likert scales: Highly likely/highly desired with boxes to highly unlikely/highly undesired
e.g. I will return to use the website in the future to do X
select one box
Highly likely  X  X  X  X  X  X  Highly unlikely
Thurlstone scale: tracks agreement or disagreement with the question
e.g. I found what I needed on the website today X yes X no
Semantic differential: quantitative measures of topics usually addressed through qualitative means
e.g. The library website is
select one box
Good  X  X  X  X  X  Bad
Guttman scales: statements arranged in sequence to gauge strength of respondents’ view
e.g. Shifting LIS communications to web based tools:
select one box
Web based commun. will catch on quickly/completely replace current tools X
Web based commun. will require some adjustment, but will mostly work X
Web based commun. will be hard to sell, may or may not work X
Web based commun. will never work for our organization X