Get the latest articles and downloads sent to your inbox in a monthly newsletter.

Get the latest articles and downloads sent to your inbox in a monthly newsletter.

Get to know your users better: Seven secrets for successful surveys

Donna Carroll, University of Leiden | Mar 13, 2012

With increasing tuition on the horizon for undergraduates at UK universities, students see themselves as customers of the education system and, as such, their expectations of academic provisions are growing. Academic libraries must ensure their services and resources remain needs-focused. At the University of Warwick Library we continually ask for feedback from our user community — often using online surveys — to inform future developments.

There are many factors to consider when creating an online survey, including deciding whether to collect open or closed responses, being careful not to use leading questions, and ensuring all options are available in multiple choice selections.

Aside from these basics, here are seven important tips we recommend in conducting effective surveys.

1. Choose the right tool Where do you start?
There are many survey tools out there, but it appears that you get what you pay for. So a decision really depends on your budget and what you need to achieve. Here’s a quick summary of my top three survey tools:

Cheap and cheerful: Google Docs allows you to create free forms that can be used as questionnaires and sent to an unlimited sample. The survey themes are interesting and varied, and questionnaires can be embedded into web pages or even e-mails so that respondents can reply without following a link. Data can be summarized at the end or extracted into an Excel spreadsheet. The biggest drawback is that you have to write your questions in one sitting (or draft it in another tool first) — you can’t edit the form after creating questions and closing it. You also can’t easily reproduce existing surveys.

Solid as a rock: SurveyMonkey is probably the most well known online survey tool. It is easy to use and free, although an annual $200 subscription allows you to create custom themes, use skip logic, and have an unlimited amount of surveys and responses. If you’re looking for something quick and easy that does the job, SurveyMonkey is it.

The whole shebang: SurveyGizmo’s free option offers slightly less functionality than SurveyMonkey’s, although it does allow an unlimited number of surveys and questions, and up to 250 responses per survey compared to SurveyMonkey’s 100. But if you can afford the $50 monthly subscription, it offers much greater customization, more question types, and better training and support.

2. Learn the language
Consider the terminology you use before sending your survey to make sure it’s free of librarian lingo. Run your survey questions by a non-librarian or someone from your target audience. We found it interesting that when we asked about “information skills,” people assumed we meant IT skills!

3. Personalize the message
We received a much better level of response by sending surveys as links within e-mails that had a personalized message to the recipient. (We used Outlook Mail Merge.) Also, people are more likely to respond when the e-mail is from someone they know. Embedding surveys within e-mails looks great when it works, but the downside is that you can’t guarantee they will be accessible from all e-mail platforms.

4. Try midweek delivery
It’s important to catch people at less hectic times. We tended not to send surveys on Fridays, when people are finishing up before going home for the weekend (and then forget about the survey by the next week), or on Mondays, when the surveys get lost in the pile of weekend mail.

5. Define responses for ease of filtering
Let people answer with check boxes or drop-down lists rather than open-ended text boxes whenever possible (depending on what you’re exploring). This allows you to quickly filter responses and make comparisons between different sections of your respondents without trawling through individual sets of unique data responses. For example, if you’re asking students what subject they’re studying, provide a list rather than a text box, so you don’t have different students typing both “biological sciences” and “bio sci,” resulting in two independent subjects in the results filter.

6. Limit to e-only delivery
Beware if you intend on increasing the accessibility of your survey by allowing paper responses. You may get more responses, but you’ll also increase your workload prior to analysis because someone will need to type them in. Only go this route if it is absolutely essential, or if someone offers their time to upload the information.

7. Supplement your survey results
A final thought: Surveys can only explore the perceived or explicit needs or thoughts of the respondents, and very often, what people say they want or would do, is not actually true! It’s always worth considering the latent needs that a survey can’t really explore.

Comments