If Truth Matters, Prove It with a Survey
The New York Times’ recent brand and marketing campaign is all about Truth. Well, if Truth really matters, then they desperately need a good old fashioned survey to back their flimsy claims in a recent article about how small business owners view the candidates running for president.
In a front-page article of the business section last month, the newspaper claimed that owners of small and medium-size businesses do not the like what they are hearing from candidates on the left. Their headline read:
Employers Are Leery of Warren and Sanders
Then they made claims in the article like:
“Small businesses routinely thank the administration for hacking through a regulatory thicket.”
“In their view, sweeping plans to remake the health care system or slash the cost of higher education will mean higher taxes for businesses and the middle class, no matter what candidates promise.”
How do they know? I wondered. I read the article to see how they were gathering their data and using it to document their claims. It turns out they talked to five business owners.
Is five enough to make such sweeping claims? I hope you agree the answer is no. Take a look at our Interactive Graph for Choosing Sample Size. Zoom all the way down to a sample size of 5, and you’ll notice that the margin of error is ± 43.83%, which is so large we can not even show it on the graph.
If The Times really wants to substantiate the claims in their article, here is what we suggest:
- Conduct a survey of at least 300 small business owners, which would give them a maximum margin of error of about ± 6%.
- Better yet, make the sample size 600 small business owners. This group is easy to find and sample, and it would bring the margin of error down to ± 4%.
- Stratify the sampling and weighting of final data by region, industry, and employer size (they can use Census data for that) to ensure that the sample matches the full universe of small businesses.
- Design careful and multiple survey questions that measure attitudes about business regulations, health care, higher education, and specific views being attributed to particular candidates.
- Measure perceptions of all candidates (not just Warren and Sanders) to establish legitimate points of comparison.
The Times says that it has “an unwavering commitment to the truth.” The problem is they cannot know the truth they pretend to communicate in this article until and unless they do a real (and rigorous) survey.
—Joe Hopper, Ph.D.