Overview
Writing a message in ALL CAPS generally defies digital etiquette. It can come across as shouting. It can convey anger. It may even signal ignorance of or apathy toward internet norms.
Most of us would never use all caps in a social media post, let alone a professional context like email. But what if the attention-grabbing power of this formatting technique could make your messages stick? Might using ALL CAPS selectively increase recall of key information?
In this email experiment, we tested whether formatting important details with ALL CAPS makes those details more memorable. We also tested whether such a benefit comes at the cost of forgetting the unformatted text, or affecting the sender’s image, specifically how warm (i.e., nice) and competent (i.e., smart) they seem.
The Experiment
823 people were recruited from Prolific to take part in an email study. Participants read a brief email from a manager or junior colleague about 4 cities to explore for market research. Unbeknownst to participants, we randomized whether 2 of the 4 cities were formatted using ALL CAPS or none of the cities were formatted with caps.
The specific text that participants read is below (ALL CAPS formatting was randomized). Participants were told that, "The following email is from your junior colleague at a market research company."
From: Brian, Intern
Subject: Cities
Hello,
I was reviewing the report you sent and thought of four cities we could investigate for next month's market research. HOUSTON and ATLANTA seem most promising. The others are Miami and Chicago.
Best,
Brian
Subject: Cities
Hello,
I was reviewing the report you sent and thought of four cities we could investigate for next month's market research. HOUSTON and ATLANTA seem most promising. The others are Miami and Chicago.
Best,
Brian
Participants were then asked a survey question to measure how warm and competent the email writer seemed. “To what extent do you think each of the following words describe this person? (1 = Not at all, 7 = Very much)” with words that included “Nice” and “Smart.”
For the final outcome, we tested recall of the four cities by having participants answer three minutes of unrelated survey questions, then asking them to recall as many of the four cities as they could.
For the final outcome, we tested recall of the four cities by having participants answer three minutes of unrelated survey questions, then asking them to recall as many of the four cities as they could.
Results
Using all caps DOES increase memorability of key information, albeit slightly. Participants recalled 15.1% more of the first two cities when they were formatted using ALL CAPS compared to their unformatted versions (p = 0.006). However, there was a trade-off; participants remembered 11.6% fewer cities in the latter part of the email, both of which were always left unformatted (p = 0.006). Thus, the total number of cities recalled was statistically indistinguishable across our ALL CAPS and unformatted conditions (avg. = 2.35 vs. 2.34, p = 0.849).
There was no trade-off, however, for perceived warmth or competence. Participants were no less likely to perceive our email sender as less nice (p = 0.772) or less smart (p = 0.982) when he used ALL CAPS vs. unformatted text. Furthermore, none of the results differed by age or gender, nor did the seniority of the email sender matter.
Conclusion
Formatting the key details of your email with ALL CAPS slightly enhances those details’ memorability. But it comes at the cost of making the remaining, unformatted text more forgettable. Fortunately, using all caps does not make you seem less warm or competent, at least when used sparingly.
Although the effect is small, other formatting techniques may further enhance the recallability of your emails. Check out our studies on the use of bold formatting and bulleted lists for more.
Methods Note
To test for significant differences in likability between our experimental conditions, we used an independent samples t-test. For statistically significant results, the difference between the averages of the two groups would be large, and the corresponding “p-value” would be small (p < 0.05). This indicates that if we were to replicate the study with the same population, we would likely obtain a similar effect. Additionally, we utilized OLS regression analyses with interaction terms to explore potential significant interactions between the main results and participant demographics, such as age and gender.
Data and survey materials used for this experiment are available upon request.