Conduct a Survey
A survey can help you learn about the beliefs, attitudes, or behaviors of people associated with a problem. For example, you might use a survey to discover whether the attitudes and beliefs about education differ among students who stay in school and those who drop out. Or you might use a survey to explore whether students who put a high value on community involvement are highly engaged in volunteer activities.
Typically, surveys help you answer what, who, or how questions — such as “What kinds of exercise do you engage in at least once a week?,” “Who will you vote for in the next election?,” or “How likely are you to use public transportation?” Surveys are less useful in obtaining the answers to why questions. In an interview, for instance, you can ask, “Why did you vote the way you did in the last election?” and expect to get a reasonably well-thought-out response. In contrast, survey respondents seldom write lengthy answers to questions.
Conducting an effective survey usually involves the following activities.
Decide whether to conduct a survey. Your decision about whether to conduct a survey should be based on the role it will play in your essay, the amount of work required to do a good job, and the kind of information you are seeking. Surveys are useful if you want to collect evidence to support your assertion that a problem exists, or if you want to learn about the attitudes and behaviors of a large group of people (more than five or ten). If you simply want opinions from a handful of people, you can gain that information more efficiently by interviewing or corresponding with them.
Decide whom to survey. Most surveys collect information from selected members of a particular group to estimate the beliefs, attitudes, or behaviors of the group as a whole. For example, surveys completed by a hundred students might be used as evidence about the beliefs, attitudes, and behaviors of all students at a school. Similarly, “national” polls seldom survey more than a thousand people, yet they are used to provide an indication of the opinions of everyone in the country.
To select participants, you can choose people from your target group at random. For example, if you are interested in surveying students at your college or university, you could open your school’s telephone directory and pick every twentieth name. Or you can stratify your sample. For example, you could randomly select a specific number of first-year, second-year, third-year, and fourth-year students — and then make sure that the number of men and women in each group is proportional to enrollment.
Decide what to ask and how to ask it. Designing effective surveys can be challenging. A survey item, such as a multiple-choice question or a true/false statement, that seems perfectly clear to you might confuse someone else. To identify confusing items, ask a few classmates, friends, or family members to try out your survey. Ask them what they think about each survey item. Then rewrite any items that caused confusion, and test the survey again. Doing so will help you improve the clarity of your survey.
Understanding the strengths and weaknesses of the kinds of items that are frequently used in surveys is a good way to get started.
- Yes/no items divide respondents into two groups.
Example: Do you check your Facebook news feed more than once a day?
Yes
No - True/false items deal more often with attitudes or beliefs than with behaviors or events.
Example: People who have a Facebook account are in better touch with their closest friends than are people who don’t have Facebook accounts.
True
False - Ranking items forces respondents to place items along a continuum.
Example: Rank the following social-networking sites from best to worst using the numbers 1 (best) through 4 (worst):
___ Facebook
___ Google+
___ Twitter
___ Tumblr
- Likert-scale items measure respondents’ levels of agreement with a statement, their assessments of the importance of something, or how frequently they engage in a behavior.
Example: If you have a Facebook account, how often do you check it?
Never
Once a month
Once a week
Every day
Several times each day - Multiple-choice items indicate whether a respondent knows something or engages in specific behaviors. Because they seldom include every possible answer, be careful when including them.
Example: I use the following tools on social-networking sites:
___ Blogs
___ Messaging
___ Friends
___ “Walls”
___ Organization pages
___ Applications
- Short-answer items allow greater freedom of response, but they can be difficult to tabulate.
Example: What do you like best about social-networking sites?
Conduct your survey. The large number of surveys people are asked to complete has reduced the public’s willingness to respond to them. In fact, a “good” response rate for a survey is 60 percent. The following guidelines can help increase your response rate:
- Keep your survey short, preferably no longer than a single page.
- Make sure your survey is easy to read. Don’t crowd the questions onto the page or jam questions into an e-mail message.
- Explain the purpose of your survey in a brief statement. If you are handing out your survey in person or during a class, provide a brief statement on the first page. If you are distributing the survey via e-mail or the Web, include a brief statement in the message or at the top of the Web page. A clear explanation can increase a respondent’s willingness to complete the survey.
- Treat survey respondents with respect. If you are mailing your survey, address the envelopes by hand and use a cover letter to explain the survey.
- Make the survey easy to return. If you mail the survey, include a stamped, self-addressed envelope. If you are conducting your survey on the Web or via e-mail, provide directions for returning completed surveys.
Analyze your results. It’s usually best to tabulate survey responses using a spreadsheet or statistics program. These kinds of programs provide flexibility when you want to analyze your results. For example, they offer statistical tests that allow you to look for differences between groups of respondents in the average rating given to a survey item. (You can learn about these tests through the program’s online help.) If you prefer, you can also organize the results in a table in a word-processing program.
If you conduct a survey and use its results in your essay, include a copy of your survey questions in an appendix.