Does Anyone Care About Mistakes in Your Profile?
A study examines the effect of typos and other errors.
Posted Oct 14, 2014
But I’m a writer and editor by trade—and, at that, I’m one who cares about such things like the order of the question mark and exclamation point when combined. (The question mark comes first.) So maybe I’m in the minority. Recently I talked with my friend Spencer Greenberg, an applied mathematician and self-optimizer, about how much typos matter, and exactly how they matter. We wondered if sloppy writing might signal lower social class, and we ran a couple of quick experiments to find out.
For the first, we composed two versions of an online dating profile. One had zero typos and one had five:
I’ve lived in three diferent cities over the last six years. I love traveling, exercise, and reading a great book. On a Friday night, I’m equally likely to be found at home watching movies as I am to be at a bar with friends. I love listening to the Beatles, especially Abbey Road, and my current TV obsession is Game of Thrones. I like great food, and meeting new people. But my friends and family come first. I’m looking for someone with great sense of humor, who is honest, and doesnt play games. It’s all about the chemistry. Let’s meet an see how it goes.
Three hundred people on Amazon’s Mechanical Turk read one or the other profile and judged it as if they were single and the profile was written by someone of the gender they found most attractive. They then answered several questions on a 1-to-7-point scale: How educated is the person? How intelligent? How high-status? How wealthy? How respected? How hard-working? How kind? How likable? How trustworthy? How friendly? How good-looking? How easy would it be for this person to get a job? How interested would you be in dating this person? Finally, they decided—yes or no—whether they would go on a date with this person. (We worried that the profile was so riddled with clichés as to be unbearably insipid, but 75% of people disagreed with that assessment. I should have known.) All of the measures were also combined into a composite “positive perception” score.
What we found: Typos did not have a statistically significant effect on any of the measures, even with 5 typos in about 100 words. The closest we came to significance was on the perception of education: Typos lowered it from about 3.5 to about 3.3.
This outcome surprised us. Maybe people just aren’t noticing the typos, we thought. So we ran another experiment. This time, the profile was longer, about 200 words, but we included 13 typos on the flubby version:
I’ve lived in three diferent cities over the last six years. I lov traveling, exercise, and reading a great book. On a Friday night, I’m am equally likely to be found at home watching movies as I am to be at a bar with friends. I love listening to the Beatles, especially Abbey Road, and my current TV obssesion is Game of Thrones. My last big trip was to Mexico, where I spent all my time on the beach or eating tacos. Someday I want to check out France and finally see the eiffel Tower up close. I like great food and meeting new people. But my friends and family come first Currently I work in sales. It’s alright, I guess, but I want to try something new. Something were I’m not sitting at at desk all day! I’m looking for someone with great sense of humor, who is honest, and doesnt play games. You should like animals because I’ve got an amazing dog named Ralph. You should also want to do something with your life, but still know where you came form. I could make a big list of what a perfect date would be like, but really Its all about the chemistry. Let’s meet an see how it goes.
We also asked a few more questions, about the target and about the participants: Is the profile-writer white, African-American, or other? How would you rate your own socioeconomic status relative to other Americans? How negatively do you react to people making a lot of typos and spelling and grammatical mistakes? And, how many mistakes do you recall seeing in the profile you read?
The people who saw the 13-typo version of the profile estimated that they’d seen, on average, 4.6. Based on this study, people tend not to notice most typos in dating profiles (unless they read actual profiles much more closely than they read this one, but usually they’re not told they will have to answer questions about a profile). Still, seeing typos in this case had a few statistically significant effects: The sloppy writer was rated as slightly less intelligent (3.0 < 3.5), educated (2.8 < 3.3), able to get a job (3.1 < 3.5), appealing as a date (2.7 < 3.3), respected (3.0 < 3.3), trustworthy, (3.2 < 3.4), likable (3.6 < 3.9), and hardworking (3.1 < 3.3). Summing up all the scores, the sloppy writer was rated slightly less positively overall than the clean typist (40.4 < 44.0). See below for a chart of percentage reduction in ratings. (Red bars are strongly significant, orange bars somewhat significant, and grey bars non-significant.) People’s ratings of their own social class (based on income, education, and job status) did not have much effect on their ratings of the profile writer.
The ratio of people who said they would date this person dropped from 71% to 53%. Based on a linear regression model, that's about 3.5% for each typo noticed. But remember that only a third of the typos were noticed. So, while we finally got people to react to typos, it was only by adding 13 to a single paragraph, many more than I’ve ever seen in a real profile.
We think mistakes worse than mere typos—genuine misspellings, grammatical errors, profusions of LOL—would have stronger effects stronger than those we found. And the effect could certainly be stronger in a resume or cover letter. But it appears that making typos in your personal dating profile does not significantly affect the impression you make, unless you type it while wearing ski gloves.
One possible exception might be if you hope to date a writer or editor, or someone as uptight about language use as I am. But who would want to do that?
A methodological note: To conduct this study, we used GuidedTrack, a new platform designed by Spencer that makes it easy to create your own online experiments, surveys, and training programs. You can take the experiment yourself here to see what it's like. If you have a GuidedTrack account you can access the experiment code here. The raw data are available upon request, and we welcome questions or suggestion about possible next steps.
Report typos anonymously to EditMob.