Survey news releases have their own unique needs, in part due to the statistical material they present. Here are some common mistakes made when reporting survey results:
- Overgeneralizing. Claiming “X% of widget vendors use the cloud” is an exaggeration of how representative most online surveys are: better to write “X% of widget vendors surveyed use the cloud.” Projecting incidence rates to an overall business population is much more difficult than projecting to consumers.
- Being overly precise. Even the most representative survey results are only accurate within a range of sampling error (e.g., plus or minus 3% at 95% confidence for a survey of 1,000): so the reported 50% might be anywhere from 47% to 53% in the overall population, and sampling theory says that there is a 1-in-20 chance that it’s any number at all. Reporting the results of your online survey to one or two decimal places simply comes across as amateurish: don’t write “15.67% of respondents prefer strawberry” but round to a whole number instead (16%, in this case).
- Claiming sampling error. More news releases than not include references to sampling error, yet this only applies for probability samples (typically telephone surveys) not most online surveys. The American Association of Public Opinion Research says this claim is misleading for online panel surveys. Instead, report a credibility interval.
- Reporting on questions with too few respondents. Sometimes the questionnaire, which had a great set of questions for respondents who answered other questions in a certain way, simply doesn’t collect enough results for that section to be useful. For instance, one survey discussed the 25% of Internet-using centenarians who use online dating; when you did the math on the percent of centenarians who use the Internet, it was clear this was referring to a single respondent out of the four Internet-using centenarians!
- Failing to disclose the basics. In our review, we found 4% of the 2022 releases analyzed didn’t even disclose the number of people surveyed – as basic a requirement as you could have. Additionally, 23% failed to disclose the dates when responses were collected. And 57% failed to specify the mode of the survey (e.g., whether it was online, over the telephone, or in person). Altogether, 58% of news releases failed to specify all three of these three basic requirements. Answer these common questions in the release to spare the busy blogger or reporter the need to contact you.
Finally, here are some general mistakes that are especially applicable to survey news releases:
- Not linking to resources. For resources that you won’t be attaching and distributing with the release (it often costs more to embed graphics and PDFs for distribution by a wire service), provide links in the release to exhibits, an infographic, the white paper, the methodological FAQ, and the web page you want bloggers to link to in their story.
- Scheduling. Quite a few survey releases tied to specific holidays or events are published just a day or two prior to the event, providing too little time for reporters to find them, research them, and write about them.
- Missing the point. It’s easy to get caught up in the details and forget to state the meaning and impact of the survey results—the ‘why’ the survey was conducted.
Your survey is ready to be published! You are well on your way to improving the amount of coverage you receive.
This is an excerpt from the free Researchscape white paper, “Best Practices for Newsmaker Surveys”. Download your own copy now.
Author Notes:
Jeffrey Henning
