There are two types of researchers – those who care about data quality and those who don’t. If your primary concern is generating leads and you don’t really care about the quality of the data in the survey report you’re promoting, then this blog is not for you. But if you do care about data quality, then here are five easy ways to maximize data quality in your next survey-based research study.
#1: Minimum of 500 survey completions
The more people who complete your survey, the higher your sample size and the lower your survey margin of error. You want your survey data to be reliable and actionable. CyberEdge recommends a minimum sample size of 500 respondents, which yields a respectable 4% margin of error. If your budget can’t accommodate 500 survey completions, then a sample size of 400 respondents will yield a survey margin of error of 5%. (Okay, so will a sample size of 320 respondents, but fewer than 400 respondents just plain looks bad.) CyberEdge will not conduct survey report projects on behalf of our clients with fewer than 400 respondents.
TIP: Check out our blog titled, Determining Minimum Survey Sample Sizes Based on Survey Margin of Error, for more information on this topic.
#2: Survey the right people (Duh!)
You’d think that this is obvious. Right? But CyberEdge has seen both tech vendors and (low-tier) IT research firms include survey responses from people who have nothing to do with the subject of the survey. For example, we’ve seen survey reports…
- … on IT security topics where the majority of the respondents do not work in the IT security department, including sales people!
- … on enterprise-only topics where the majority of the respondents are employed by organizations with fewer than 500 employees, including under 10 employees!
TIP: Be sure to include “filter questions” that politely eliminate respondents from the survey. And configure the setting in your online survey tool that prevents eliminated respondents from re-starting the survey (from the same IP address or web browser) just so they can obtain the survey incentive.
#3: Add ‘Don’t know’ responses to most questions
As a very strong rule of thumb, don’t assume that every respondent knows the answer to every question. Thus, include a “Don’t know” response to most questions so you’re not forcing respondents to guess at questions they don’t know the answer to.
#4: Randomize responses to most survey questions
Unless the responses must appear in a sequential order, configure your online survey platform to randomize responses whenever possible. This prevents “order bias,” where some fatigued survey respondents will select the first item on the list in an attempt to complete the survey as quickly as possible to obtain the survey incentive.
#5: Translate surveys for non-English speakers
Sure, most IT professionals in non-English speaking countries are conversant in English. But the percentage of such individuals varies by country. And there’s always the potential that a respondent who is not fluent in English will misinterpret survey questions and responses. So, if you’re going to conduct your survey in countries where English is not the national language, then translate the survey and host a separate instance of the translated survey for those respondents.
At CyberEdge, data quality is paramount. We hope you’ll share our believe and practice these good habits when conducting your own survey-based research project.
Share this Post