Surveys. If you haven’t created one, you may have participated in at least one. But there’s a lot more to a survey than one might think.
At the core, surveys are a research method that aim to represent a population as a whole by gathering information from a sample population. To do this, a researcher must be thoughtful about the choice of questions as well as the choice of participants.
Without proper preparation and mindfulness, data and outcomes are easily skewed thereby defeating the purpose of the survey.
Surveys at Champlain are subject to the College’s Survey Policy and must be approved by the Office of Institutional Research & Assessment.
Survey projects that are part of a research study may be subject to approval by Champlain’s Institutional Review Board (IRB).
Champlain’s Institutional Review Board (IRB) is the body that reviews research involving human subjects conducted under the auspices of the institution, in order to oversee whether such work is being conducted according to the ethical standards of academic research. The rules governing IRB and human subjects research are controlled by the federal government, regulation 45 CFR 46.
If your research involves interacting with human beings in any significant way, it could qualify as human subjects research. Research in this context is understood to be systematic investigation, including research development, testing, and evaluation designed to develop or contribute to generalizable knowledge. Research methods that entail significant interaction with humans include but are not limited to surveys, interviews, participant observation, focus groups, and experimentation.
If your survey results are to be used for internal institutional purposes, then your project will likely not need IRB review. If your survey is part of a systematic investigation designed to contribute to knowledge in a field of study or a published paper, then your survey may need IRB review. If you are unsure, you can check with the Champlain College IRB Chair at irb@champlain.edu.
Find advice online about question types and how to write good survey questions, including from Nielsen, Qualtrics, Hubspot, Pollfish, and SurveyMonkey.
There are many good reasons to gather information through surveys; however, before embarking on a survey journey, it is important to ask yourself if conducting a survey is the best approach to obtaining the information you need.
A survey is not a good idea when | Alternative ways to collect information include |
|
|
While it may seem simple enough, creating a quality survey takes time. If you do decide to conduct a survey, you are required to submit a Survey Approval Request Form to the Office of Institutional Research & Assessment.
Additionally, you will want to pilot test your survey. This allows you to identify problematic instructions or questions. Questions that are misinterpreted will not be useful to you, so make sure to get the most out of every question.
Consider the following timeline:
Define a clear, attainable purpose for your survey supported by three to five goals. The purpose is a broad statement of the primary survey aim or outcome. The goals are specific and measurable ways to achieve that purpose.
State your purpose using an action verb, such as describe, explain, explore, identify, investigate, gauge, measure, assess, or test.
Goals help you determine your target audience, how best to distribute the survey, and how to design questions that are relevant and lead to actionable outcomes.
The purpose and goals will help you provide context and set expectations for your survey audience, which is vital to engaging them with the survey and obtaining valid results.
Writing good survey questions and organizing them sensibly makes it easier for respondents to answer truthfully and complete the survey. Your aim is to make completing the survey as easy as possible for the respondent.
To gather opinions and behaviors, consider what type of rating scale will produce the best, analyzable results from your multiple choice questions.
Important considerations for writing questions include the following:
Selecting the target audience is vital to obtaining the actionable insights you seek. The target audience is based on a set of demographics, behaviors, or characteristics.
Suppose you are surveying students. Do you want to include undergraduate/graduate, traditional/adult, part-time/full-time, residential/commuter, in-person/online? Students who studied abroad, had an internship, were in a club? This is just one example of “audience segmentation.” Targeting the correct, specific population is crucial to the success of your survey project.
The sample population must be of sufficient size to allow you to generalize your findings in a way that accounts for diversity and heterogeneity within the population. The sample of respondents needs to be representative of the population’s important characteristics to allow for valid inferences from the data. (A representative sample “accurately represents the characteristics of a larger population.”)
There are multiple ways to administer a survey, including one-on-one interviews, mail, and online. Chances are, you will be using an online platform. Considerations include user privacy and data security.
Self-administered formats include Google Forms as well as freeware such as SurveyMonkey. For a high-stakes survey, work with the Office of Institutional Research & Assessment. We will build the survey and administer it for you through Champlain’s licensed Qualtrics account.
Consider how publicity and incentives can bolster your survey’s response rate and the likelihood of obtaining results from a representative group. Whether you are emailing survey links, conducting a survey in class, or posting a QR code to provide survey access, you will need to promote your survey in order to get a good response rate and a representative sample.
In promoting your survey, it is helpful to emphasize the survey’s value to the respondent or to something they care about. You can advertise, for example, on social media, The View, digital signage, the Stall Street Journal, the Academic Brief, the People Center Brief, The Crossover, posters, flyers, or verbal announcements (at Faculty Senate or Staff Council). It may be appropriate (for a high-stakes, institutional survey) to send a pre-survey announcement (from a high-placed administrator) by mail or email.
Incentives can improve response rates, but also pose potential negative consequences. Incentives should be large enough to entice the respondent, but not so large as to overshadow the importance of the survey itself, or worse, constitute “undue influence.” A small incentive for each respondent who completes the survey (a $5 gift card to Zime) or a chance to win a larger prize (a $30 Amazon gift card, an Apple watch) through a lottery constitute acceptable incentives.
If incentives are used, promotional materials as well as the survey invitation and reminder communications must describe the prizes, their value, and an estimate of the odds of winning.
Surveys ask respondents for their opinions and personal information. As the initiator and owner of a survey project, you are responsible for making sure your project respects privacy and data security.
If you collect personal data, tell survey takers how you plan to use their data and how it will be stored, protected, and potentially destroyed. Sharing your privacy practices with survey takers may help ease their privacy concerns and increase your response rate. Describe your privacy practices as well as the anonymity or confidentiality of the survey results in the survey introduction and in the email inviting people to take your survey.
How data is stored, used and protected varies from platform to platform.
If the survey is built on the Qualtrics platform, it is secure; “Qualtrics complies with applicable data privacy laws in its role as a data controller of personal data it collects,” and is GDPR compliant.
However, confidentiality and privacy of data collected through Google Forms depends largely on the user’s settings, making user vigilance an essential aspect of data security. Google Forms should not be used for collecting sensitive information unless additional security measures are in place.
Once you have chosen the platform, you can choose from different options for distributing the survey, for example:
If you are unfamiliar with analyzing quantitative and qualitative data, ask for help from the Office of Institutional Research & Assessment.
There are four different measurement scales that shape data analysis:
Data analysis allows you to make inferences from the survey data that can lead to insights or actions. Analysis is done by conducting operations on the data. A good survey–which asks about measurable, observable characteristics, activities and behaviors–provides analyzable data. Use statistical analysis to determine if comparisons or trends have statistical significance rather than being due to chance.
Most likely, you will conduct descriptive analysis, which summarizes and describes the important characteristics of the data set, mainly frequency, central tendency (averages, including mean, median and mode), and dispersion.
Cross-tabulation allows you to break out the data according to sub-groups within the survey population for comparison. Also known as contingency tabulation, cross-tabulation is subject to tests of statistical significance including chi-square. Cross-tabs can be created in Excel using pivot tables.
Whereas quantitative data answers the “what,” qualitative data enables you to dig into the “why” and “how.” A word cloud is a quick and easy approach that uses an image, illustrating the most frequently used terms. More rigorous qualitative analysis methods include:
Once your data analysis is complete, share the story. A formal report should include:
Use visuals (charts, graphs and tables) to support text. You may wish to track how the findings for particular questions have changed over time (trend analysis). Take care to contextualize and not overinterpret data. Do not confuse correlation and causation.
Contact the Office of Institutional Research & Assessment for survey approval or help with survey design, administration and analysis
Contact Us