Can I ask you a question: How to conduct survey research
My first visit to Syracuse University in the cold and wet spring of 2012, I was stopped on my way into the bookstore by a young woman with a clipboard and pen. As I inwardly groaned about being interrupted on my way to buy t-shirts for my family members (I was determined to deck them out in orange as I had just been accepted into the graduate program), I begrudgingly agreed to take her short survey. I did this because I have rule. Whenever I’m stopped in the shopping malls or called at home to take a survey, I try to always oblige. I have this rule because after conducting research myself, experiencing the disappointing response rates infamous with surveys, and knowing the frustrations a researcher faces in obtaining a reliable sample size, I understand how important surveys can be for planning in Public Relations. These research tools provide the best method for making descriptive analysis of your larger target audience, and, like the old saying goes, ”knowing your audience” is vital to a successful campaign.
How to develop your own survey research
To get at the heart of your target audiences’ information, you will need to get three things right:
1. The right sample size
2. The right questions
3. The right delivery medium
The quality of a sample is judged by how representative it is of the entire population it’s drawn from. For the best results, a census, or giving every person in your universe an opportunity to respond, is ideal. As censuses are, unfortunately, often unrealistic, the next best thing is to concentrate on gathering a sample size that can guarantee the industry standard of a 95 percent accuracy rate. As PR research guru Dr. Don W. Stacks puts it, “… when we set a 95% confidence interval, we are saying that no more than five units (people or messages) will be missampled …” To reach this confidence level, there are calculators and graphs available, but as a rule of thumb, an actual-responding sample size of 384 will garner a 95 percent confidence interval with a population size of 500,000 or more. (I highly recommend checking out Dr. Stacks’ “Primer of Public Relations Research” for a more in-depth look into drawing a sample.)
Types of samples
There are two basic types of samples, probability and nonprobability. Probability, or scientific, samples are those that are randomly selected so that each person or unit in the population has an equal chance to participate. These include:
-Stratified random where the population is divided into distinct groups where a random sample is selected from each group
-Systematic random where the population is segmented into elements and each element has an equally random chance of being selected but combinations of elements have different probabilities of being chosen
-Random cluster where the population is divided into similar groups where a cluster is randomly selected
-Stratified cluster where the population is divided into similar groups and clusters of matching groups are made into separate strata where one cluster from each strata is randomly chosen
Nonprobability, or convenience, samples are those that participants are not randomly selected, and, therefore, each person in a population does not have an equal chance of participating. These include:
-Convenient/Accidental where a sample is drawn from whatever is first available
-Quota where the population is divided into subgroups and then a participant is selected by convenience from each subgroup
-Judgment where the researcher selects participants he thinks are most representative of the population
-Snowball where a current participant is asked to identify other potential participants like himself
How you ask and where you place questions in your survey’s questionnaire can affect your response rate. To receive the most accurate information that paints the landscape of your audience’s perceptions, it’s vital that you’re asking the right questions that get at the heart of the matter.
Types of questions
There are five types of questions that can be used in a questionnaire:
-Open-ended questions are those that allow the respondent to answer freely without constraints.
If you could improve the Aubia Communications Blog, how would you do it?
-Multiple-choice questions give several answer options, and be sure to always give the respondent an “other” selection.
What would you most like to see on the Aubia Communications Blog?
A. Posts about Public Relations in other countries
B. Posts about media relations
C. Posts about employee communications
D. Other _______________
-Yes or no questions serve a specific need for your research.
Would you recommend the Aubia Communications Blog to your colleagues?
-Ranking answers allows you to gauge what is most important to your audience.
Please rank in order from 1 being most important to 4 being least important what you would find useful on the Aubia Communications Blog.
___ Coverage of current events in Public Relations
___ In-depth how-to articles for Public Relations techniques
___ Review of academic Public Relations literature
___ Roundup of Public Relations resources
-Number scale questions, typically found in a Likert Scale, can be used to judge your audience’s values, and be sure to offer the respondent a neutral choice.
With 1 being Strongly Agree, 3 being No Opinion and 5 being Strongly Disagree, please rate the statements below.
___ The Aubia Communications Blog is a useful Public Relations resource
___ The Aubia Communications Blog provides valuable information
___ The Aubia Communications Blog has well-written posts
___ I would recommend the Aubia Communications Blog to colleagues
These five types of questions can be used in four categories:
-Non-threatening questions are those that are simple and easy for the respondent to answer. Specific questions asked in a more structured format, such as yes-or-no or multiple choice, should be placed at the beginning of the questionnaire.
-Threatening questions are those that can make the respondent uneasy about completely the questionnaire. These are personal questions that are best saved until the end of the questionnaire after some trust and comfort has been built into prior questions. It is best to keep these deeper questions in open-ended format, creating a safe environment for the respondent to answer in his own words.
-Knowledge questions test the understanding of a subject among your audience. Make sure your questions are clear to your intended audience and fall in between an easy and difficult range.
-Attitude and opinion questions are used to gauge the intensity of how your audience feels about a certain subject. Be sure to keep these questions separated so as not to cause a double-barrel effect, and do not use inflammatory language. This category is best asked with a number scale format.
Mediums for survey research
How you distribute your survey can have a lasting effect. It is most important to make sure your survey is reaching your intended audience where they are most conveniently located.
Telephone, Mail, Person-to-Person and Internet-Based (such as email) are the four mediums for survey research. Below are the advantages and disadvantages of each provided by the Universal Accreditation Board’s Accredited in Public Relations Study Guide.
|Telephone||-Speed of administration
-Ability to clarify, probe
-No geographic limitations
-No sampling limitations
-Accepted as industry standard
|-Requires professional phone bank
-Limited in length and scope
-Not good for rankings, paired
comparisons, evaluation of messages
|-Can be internally administered
-No cost differences based on geography
-Can use rankings, comparisons
-Can do message evaluations
-Can be included in existing
|-Slow to completion
-Considerable time/labor commitment for
-Limited control over who completes
-Little ability to clarify question
-Limited ability to use filter questions
-Cost can be high if response rates
-Often yields very low response
|Person-to-Person (Intercept)||-Ability to target a geographic location or demographic group
-Ability to clarify, probe
-Can use supporting materials
|-Limited in length and scope|
|Internet-Based||-Speed of data collection
-Lower cost of administration
-Excellent tool for known populations with access to technology
-Can combine data collection methods and gain advantages over telephone and mail methods
-Can disaggregate results easily for analysis
|-Best with known populations
-Requires access to technologies
-May be difficult to generalize
to larger groups or populations
-Requires disciplined administration
and ability to randomly sample from
within universe to qualify as formal
research; most often is informal
When you have your sample size set, your questions in order, and have chosen your medium, be sure to pretest the survey with a small select group. Take any feedback from that group and make any adjustments needed before releasing the full-scale survey.
Surveys are valuable research tools that help you generalize your larger target audience so you can figure out what they want. Have you conducted your own survey? What were the results?
Feel free to answer the italicized mock questions above in the comments. I always love to hear the opinions of readers.