SharePoint Surveys - Design for Analysis

By DANIEL ANTION posted 07-27-2011 07:22


Last week, I spent a good deal of time wiring up Data View Web Parts to display survey results. If you want some background on that, see SharePoint vs. Excel, Round 2, but the background isn’t important here. What I want to share here are the things I learned about crafting surveys from having to analyze survey results. Nothing here qualifies as a stunning breakthrough in survey design, and these suggestions are aimed at making the downstream job easier – that has to be balanced with making the survey experience a good one. Note: the survey referenced is designed to gather information about attendees coming to an off-site meeting.

Collect What You Need – We used to ask “Enter your name as you would like it to appear on your nametag”. That made it easy for the attendee, but it was then difficult to put the lists we use in alphabetical order. Now we ask for “Last Name” and “First name as you would like it to appear…”

Small Bites – We are offering a simple social event on one of the days. We set up a multiple choice question (allowing more than one answer) to handle whether or not the attendee is participating and whether or not his or her guest is participating. For a different, more complex event, we presented two questions, one for the attendee and one for the guest. Individual questions are much easier to analyze. If you do combine the questions and you include options like “I am coming to dinner” and “My guest is coming to dinner”, don’t make the mistake of adding the third choice: “Both my guest and I are coming to dinner” – it makes the analysis much more difficult.

Isolate Control Attributes – Certain questions control the analysis. Be sure that you isolate those, make the answers required and keep them simple. For example: “Are you planning to attend?” (Yes or No).  “Are you planning to bring a guest?” (Yes or No). If the person is not planning to attend, we don’t need to worry about considering the rooms they might need, meals they might eat, or sessions and events they might attend. Even more important is what happens when plans change. If a person is planning to bring a guest, he or she will likely complete the social calendar for the guest. If their plans change, they will change their answer to “No, I am not bringing a guest” but they probably won’t clear all the other selections. By keying off the guest attendance question, you can avoid putting a ghost on a tour bus.

Answers You Can Use – When analyzing a survey, all you have is the question and the answer; make sure the answers are worded so you can extract the information you need. Consider Golf Equipment: if our attendees or their guests are golfing, we can arrange to rent clubs and shoes. The multiple choice field includes answers like “Mens right-handed clubs”, “Mens left-handed clubs”, “Mens shoes”, “Womens right-handed clubs” and so on. The choices include “No equipment necessary”, but if neither person is golfing, they might skip the entire question. To avoid reporting that “Equipment needed includes <blank>”, I have to check 6 separate possible answers. Fortunately, I was able to check first to see if the answer contains(‘ens’), but I didn’t plan it that way. Better to be lucky than smart I guess.

One last tip: before you design your survey, design the report. Knowing what you want to report will make designing the questions easier.

#SharePoint #survey #analysis