Subscribers often ask about how survey research works.
Before delving into the topic, I will first provide a definition of survey research posted by WhatIs.com. They define survey research as “the collection of data attained by asking individuals questions in person, on paper, by phone or online.” SnapSurveys further elaborates that there are four major benefits of survey research: 1) low cost, 2) samples allow you to describe large populations, 3) they are amendable to many delivery modes, and 4) they are dependable in getting honest and reliable responses. “
Surveys are quick, inexpensive ways to learn about public opinion, consumer thoughts, market forces, and other interesting things. When handled well, surveys offer helpful insights and ideas for future research. Thanks to surveys, people have invented new products, services, and even companies. It helps them stay in touch what people want, need, and don’t need.
Ensuring high data quality is essential when running a survey. Researchers put safeguards in place to help make sure data is accurate. Quality controls include monitoring the process, analysis, and verifying the most important parts. Also, computers can help control questionnaires, manage samples, and ensure that everything is presented well. Exact methods depend on the type of study.
Writing Good Surveys
Survey researchers learn to design questions that provide meaningful answers. In training, they learn to write clearly and pilot-test the questions. Quality control is essential (Lavrakas, 2008). The questionnaire should be clear, easy to read, and focused on the research question(s).
When the survey is on a computer, this helps standardize management and monitoring.
It’s also important to examine context and wording for questions and directions. Even the order of the questions matters, because it can influence results.
In general, they should be framed in a way that doesn’t elicit strong emotions. If people get emotional during a survey, that can affect the data quality. However, in personality or values tests, sometimes it’s okay if the questions provoke emotions.
Choosing Who to Survey
Many researchers are interested in sample design and coverage (Lavrakas, 2008). A sampling methodology describes how researchers choose who to survey. This should be done carefully to make sure that respondents are representative of the population. This way, the data is more reliable and researchers can make better conclusions.
When choosing samples, researchers should keep different subgroups in mind. Samples should try to include people from all subgroups, and they should avoid over- or under-representing a group. When done well, researchers can compare different subgroups and draw conclusions for them.
If a sample is chosen badly, a researcher may have to throw out the data.
Questionnaire Use and Misuse
Anyone can design a questionnaire nowadays and put it on the internet. Unfortunately, some untrained people throw together surveys and treat them like valid research studies. Because so many people are doing questionnaires poorly, this has hurt the reputation of these surveys in general.
One example of survey misuse is the many bad surveys that take place before major elections. A campaign might deliver a survey full of leading questions to try to influence respondents. Then, once they use the bad data, it looks like lots of people support the candidate’s positions.
In phone surveys, people may also use tone of voice to try to influence someone. This pressures someone to say what the speaker wants to hear.
Nowadays, survey companies are more careful about this. Many of them take measures to prevent errors like this. For example, making sure an online survey is anonymous helps people feel like they can be honest.
However, online surveys tend to have a low response rate. This means that respondents may only choose to answer because they have strong feelings about the subject. Thus, they may under-represent people who are more ambivalent.
To reduce this, researchers can email respondents before they run the study. They can also follow up to make sure the respondent didn’t feel rushed or distracted during the survey.
Often, readers prefer to take surveys about topics they understand well. Importantly, they don’t want their response options to be limited. Thus, it’s important to pre-test survey items in case there are options you missed. Personally, I find it helpful to allow people to enter their own response if one of the provided ones doesn’t cover it. This can share useful information not only about item integrity, but about ideas you didn’t think of before.
While surveys are limited, they remain a critical part of research. For marketers, they’re fast and often clear enough. They may not always share definitive answers, but they often help guide researchers.
One of the best aspects of surveys is the low cost (Eaden et al. 1999). They’re cheap, quick, and convenient. When done online, they’re easy to use and require no travel. Also, there’s no need to hire lots of people to oversee their use.
Online surveys are kept anonymous. This helps ensure that the responses are accurate. While taking the survey, the participant doesn’t need to worry about negative consequences. This helps them be honest and confident while they give their answers.
It’s also easy to collect and sort data. For paper surveys, high-speed scanners can help input data. Online, answers go into a database.
The sample sizes are often good (Eaden et al. 1999). Often, researchers underestimate the importance of a big enough sample size. You need to survey lots of people so you can make firm conclusions. Often, surveys can reach thousands of people. This way, it’s easier to make solid conclusions.
Online surveys can be powerful tools nowadays, especially when they reach many people. If researchers choose their samples carefully, they’re likely to succeed.
Eaden, J., Mayberry, M. K., Mayberry, J. F (1999). Questionnaires: the use and abuse of social survey methods in medical researches. Publication of the Fellowship of Postgraduate Medicine.
Lavrakas, P. J (2008). Quality control. Encyclopedia of Survey Research Methods. Sage publishers. Available at https://dx.doi.org/10.4135/978141293947.n422
Lavrakas, P. J (2008). Representative sample. Encyclopedia of Survey Research Methods. Sage publishers. available at https://dx.doi.org./10.4135/9781412963947.n469
Would you like to take part in some of our surveys? Visit our survey research page.