“75% lean” or “25% fat”? Same meat, different verdict
Copywriting advice is easy to find and hard to check. A few studies held the facts still and changed only the words, or only the typeface, and measured what happened. Here are three that matter for a clinic website, with how big each study was and what it did not test.
1. “75% lean” or “25% fat”
In a study published in the Journal of Consumer Research in 1988, two researchers at the University of Iowa gave 88 students a taste of the same ground beef, cooked in the same skillet.1 Half were told it was “75% lean”, half that it was “25% fat”. Both statements describe the same meat. Everyone rated it on 7 point scales.
The label changed the ratings. On the scale from fat to lean, the gap between the two labels was:
- 2.32points when people only read the label and never tasted (an earlier study, separate group)
- 1.10points when they read the label, then tasted
- 0.60points when they tasted first, then read the label
So the words mattered most before any experience, and experience diluted them without erasing them. On the one scale about taste itself, the label made almost no difference once people had tasted.
Why this matters for a clinic: a visitor to your website has not had the treatment yet. On that page they only have your words. That is my inference from the study, not something it tested on clinics.
What this study does not show
- 88 students, one food, one university, 1988.
- The “label only” numbers come from an earlier study with a different group, which the authors flag.
- It measured ratings, not purchases. And both frames were true. It says nothing in favor of wording that misleads.
2. Hard to read feels hard to do
In 2008, two psychologists at the University of Michigan published a short report in Psychological Science.2 Students read identical instructions for an exercise routine, printed either in Arial or in Brush, which participants rated clearly harder to read.
- 8.2minutes the routine was expected to take, in Arial
- 15.1minutes expected for the same routine, in Brush
- 4.5 vs 2.9willingness to make it a daily habit, on a 7 point scale
A second study repeated this with a recipe for a Japanese roll: 22.7 minutes of expected preparation in Arial, 36.2 in Mistral, another typeface rated harder to read. People remembered the instructions equally well in both versions. What changed was how demanding the task felt.
On a clinic website, the obvious candidates are the texts that describe something the patient has to do: how to book, how to prepare for a treatment, what aftercare involves.
What this study does not show
- The samples were very small: 20 students in the exercise study, 27 in the recipe study, 19 in a third.
- It compared Arial with unusual typefaces that readers found clearly harder, not two ordinary website fonts.
- It measured estimates and stated willingness, not whether anyone did the exercise.
- I looked for a direct replication of these studies and did not find one. That is not proof there isn’t one.
3. What happened when the hype came out
In 1997, John Morkes and Jakob Nielsen ran a study in the SunSoft usability laboratories in Menlo Park, California, published by Nielsen Norman Group.3 51 web users were split across five versions of the same small travel website about Nebraska. The information was the same. The writing was not:
- Control: promotional style, with exaggeration, subjective claims and boasting.
- Objective: the same text with the promotional language removed.
- Concise: about half the words.
- Scannable: headings, bulleted lists, highlighted keywords.
- Combined: objective, concise and scannable together.
They combined task time, errors, memory of the content, memory of the site structure and satisfaction into one usability score. Against the promotional control:
- +27%objective version, hype removed
- +58%concise version, half the words
- +124%all three changes together
Removing the promotional language on its own significantly improved only satisfaction. Speed, errors and memory also moved in the right direction, but those differences were not statistically significant.
What this study does not show
- About ten people per version.
- A usability score, not bookings or sales.
- A travel site in 1997, tested in a lab with set tasks.
What I take from this for a clinic
- Words set expectations before any experience, and a website visitor has none yet. Choose true framings deliberately.
- Anything a patient has to do (book, prepare, recover) should be set in plain, easy to read type, because difficulty reading can be mistaken for difficulty doing.
- Plain, short, factual copy tested better than the promotional version of the same facts.
- None of these studies measured bookings. They measured judgments and usability. I use them as reasons to fix the text, not as a promise of a result.
Sources
Each source was opened and read before this page was written. Where a study is small, old or not peer reviewed, it says so.
- Levin, I. P., & Gaeth, G. J. (1988). How consumers are affected by the framing of attribute information before and after consuming the product. Journal of Consumer Research, 15(3), 374 to 378. doi.org/10.1086/209174 Design: lab experiment, peer reviewed. 88 University of Iowa students, compared with a label only group from Levin (1987).
- Song, H., & Schwarz, N. (2008). If it’s hard to read, it’s hard to do: Processing fluency affects effort prediction and motivation. Psychological Science, 19(10), 986 to 988. doi.org/10.1111/j.1467-9280.2008.02189.x Design: three small experiments, peer reviewed short report. 20, 27 and 19 students.
- Morkes, J., & Nielsen, J. (1997). Concise, scannable, and objective: How to write for the Web. Nielsen Norman Group. nngroup.com/articles/concise-scannable-and-objective-how-to-write-for-the-web Design: between subjects usability experiment, published by the firm, not peer reviewed. 51 web users across five versions.