Conducting Survey Research

Conducting Survey Research

ResearchUSA Surveys are conducted to obtain data that can be used to assess trends and conditions, advance understanding, test theories, develop policy recommendations or business strategies, and much more. Surveys can be conducted by Qualtrics email questionnaire (in writing) or by interview (by phone).

In our context, surveys are a way to assess business owner needs that goes beyond the information collected when creating a building and business inventory.

As in all science, survey measurement is not error free. Procedures used to conduct a survey have major effects on outcomes. Therefore, your goal as an analyst is to choose the most appropriate survey procedures that, when applied, will reduce error and maximize the likelihood to accurately describe what is being measured.

ResearchUSA offers its business survey expertise for Sampling, Question Design, Data Collection, Interpreting Data, and Communicating Results. Following best practices in each component will enable a community to make more informed decisions.
Anytown Example:
The leaders of “Anytown” are interested in improving the retail mix in their central business district by attracting new businesses in sectors they currently do not have. They ask respondent to identify the factors that make doing business in Anytown attractive. ResearchUSA helps you put together a survey research program that seeks to understand this research question. You envision surveying a sample of small business owners to obtain their opinion on why they operate their businesses in Anytown. Following the five key components, here’s how a survey research program can be implemented in Anytown.


Sampling consists of selecting a small portion of a population as representative of the whole population. It is this sample that you would actually survey. When sampling, you need to give all members of the population the same chance of being selected. Note: For small city downtowns (cities under 25,000), we advise that you attempt to contact ALL businesses and set a goal of achieving a 100-percent response rate.
Anytown Illustration:
Considering your budget, you decide to survey a sample of 200 small business owners by written email questionnaire. Since you do not want your entire sample to come from the same business sector and you want to give all businesses an equal chance of being chosen, you decide to randomly pick 20 small businesses from the 10 business sectors operating in Anytown’s central business district. You first sort all businesses by business sectors, then randomly pick 20 businesses from each of the 10 business sectors. After selection, you end up with a sample of 200 small businesses that reflect the different businesses operating in Anytown’s central business district.

Question Design

Survey questions must be carefully worded, double-checked by a fresh pair of eyes, and pretested to insure they are understood the way you intend them to be. A poorly worded question will greatly increase the chance of response error and limit the usefulness of the survey data you collect. Simply put, designing good survey questions involves selecting those needed to meet the research objectives, testing them to make sure they can be asked and answered as planned, then formatting them to make it easy for interviewers to ask them and for respondents to answer them.
Anytown Illustration:
Now that you have selected your sample, you are ready to design the survey questionnaire itself. Since this is your first time designing a survey questionnaire and you do not want to hire a professional survey designer for budgetary reasons, you decide to search the web and ask around about communities that pursued a survey research program similar to the one you are commissioned to do. You found that Othertown ran a similar survey research program and is willing to share the questionnaire with you. Since Othertown has different amenities, you restructure its questionnaire to reflect both the objectives of your research agenda and the different business environment of your community.

For a more comprehensive approach to data collection, you include both quantitative (forced-choice questions, such as yes or no, or true or false) and qualitative questions (open-ended questions that give respondents a chance to write their thoughts and feelings). After clearing the survey instrument of any possible confusion and receiving feedback from all your collaborators, you decide to test it by sending it to 10 small business owners of Anytown who you know personally and who are not part of your sample. After receiving honest feedback from your 10 “testers” and making any necessary changes, you are ready to send out the survey to your sample.

Mode of Data Collection

ResearchUSA surveys can be administered by telephone or email. Each of these survey alternatives has both benefits and shortcomings. Business owners are busy people and it’s difficult to reach them over the phone or email. Response rates can range between 5% - 25% and ResearchUSA recommends selection of 10 – 20 times as many records to call. ResearchUSA’s predictive dialer will detect no answers, busy signals, and other non-voice signals. Over 50% of call attempts are not answered. We can make several attempts to reach the business owner by trying different times of day, different days, etc.