Best Practices In Online Questionnaire Design
By Kelly Kwon, Data Analyst, Polaris Marketing Research
Online survey capabilities have changed significantly in the last few years with more engaging templates, such as drag and drop features, and publishing formats, such as smart phone and social media applications. Although there are so many new styles and ways to capture online data today, it�s still important to apply �best practices� to ensure quality data.
Polaris often receives questions about online questionnaire design, so here are some best practices to frequently asked questions about constructing a successful online questionnaire:
Using odd-numbered scales give you a better result. When using an even-number scale (ie. Strongly Agree, Agree, Disagree, Strongly Agree), you are forcing your respondent to agree or disagree if they are undecided or neutral.
A human mind can effectively embrace up to 7 values, hence the phone number system. It�s not recommended to use a scale bigger than 7 points (with the exception of a Net Promoter Score � click here for more information). Also, numbered scales are difficult for people. For example, a scale marked with 1 to 5 results in less accurate data than a scale with labels.
Don�t Know Option
Not including a �Don�t Know� or �None� option can skew your data. It is important to include a �Don�t Know�, �N/A (Non Applicable)�, or �Other� in your scale, check-all-that-apply, select-one, or even open-ended type questions where people simply don�t have an opinion.
Including a �Don�t Know� option also increases the accuracy of response because you can easily take those respondents out when analyzing your data.
A matrix question is a concise technique for combining questions with common topics, and it is a favorite among researchers. By combining questions in a table, a respondent can complete the set of questions 50% faster than having to answer each question separately.
However, matrix questions raise a few concerns when used too much and incorrectly. When respondents see matrix questions, they find them tedious and complex to fill out. The faster speed of completion can lead to errors. They actually answer the questions quickly, but it doesn�t feel that way. One common result with matrix questions is �straight lining�, when respondents select the same choice for each question in a virtual straight line down the table.
If you plan to have a long matrix table, it�s recommended to have a quality control question in between the rows. (Please see highlight below for example.) Limiting the number of questions in the matrix table and actually having fewer matrices will result in better data quality and more engaged participants.
For matrix questions, placement of the error message is also important. Rather than placing a general error message at the top or bottom of the table, help the respondent point to the location of the error by placing the error message next to the question. Also, alternating colors of the rows can help reduce respondent fatigue and error in matrix tables.
Ranking questions and questions with a lot of check-all-that-apply options can be overwhelming to respondents, especially if they need to scroll down to see the rest of the question. Limit the amount of scrolling on a single page. Respondents� browser settings are different and one respondent might have to scroll down to click the �next� button whereas another respondent doesn�t. Also, disable the back button so that screened-out participants can�t go back to change their answers.
Using these tips will help you create a better online experience for your respondents as well as to get better quality data. And that�s a win-win we can all appreciate!
Kelly Kwon is a Data Analyst in the analytics department at Polaris Marketing Research Inc., where she handles a variety of data manipulation tasks involved in survey research. She has a bachelor's degree from the University of California, Berkeley in Molecular and Cell Biology and in Marketing from Georgia State University.
All Content � Copyright 2004-2011. Polaris Marketing Research, Inc. All Rights Reserved.
Send inquiries to email@example.com or call 1-888-816-8700.