First, methodology: I downloaded the United States Social Security Administration 2025 data set . This contains data on US names, by year of birth, from 1880-2024. In the 1980s, you started getting a Social Security card at birth, so what's in the data since then is different than previously. In particular, if you only get a SSN when you start to work, and child mortality was higher in earlier data, you don't QUITE get accurate comparisons. In addition, you only get names given to more than 5 people in that year, for privacy reasons.
I downloaded the zipped version and unzipped it into text files so that I could use command-line tools on it. There are individual files for every year.
I then collated every instance of the text strings Kermit, Grover, and Elmo into individual files. I did this by using the command "grep" over the contents of the directory. grep takes every instance of whatever string you give it, and grabs the whole line that string is in. So it would grab every instance of the name, and all the other info on that line, for me. I passed the results to a single csv file.
So for example, I used the command:
grep Kermit *.* | cat >> kermit.csv
and that would give me a file called "kermit.csv" with every single line containing the string "Kermit" in the whole directory.
At this point I needed to do some data cleaning, because what the data looks like is:
I needed to turn, for example, "yob1925.txt:" into "1925," to end up with a nice csv file. That's easy enough to do in a text editor. Then I opened it up in a spreadsheet. I like to use Numbers personally for this kind of thing because you can put multiple tables next to each other on the canvas.
Now I had to make some decisions: am I combining M & F instances of the name Kermit in the same year? Also, does "Kermitt" or "Kermith" count as "Kermit"? Most of the name charting things you can get online a) separate by assigned sex and b) are restricted to the exact spelling, so I decided I was going to use only M data (matching the characters' genders) and only the exact spelling (to match the data usually out there). There's 5 years with F Kermits, 14 years with F Grovers, and 26 years with F Elmos. All gone.
Elmo had the most name variations I had to filter out:
Now I have this nice filtered table that has all the years in which each name appears in the data. It looks like so:
Then I charted each data set.
We also need to look figure out what's going on when the graphs get a little weird. For example, this in Elmo data:
That means there's not 10 years of data between 1990 and 2000; the next time it got to 10 lines of data was 2003. And indeed:
This means there were fewer than 5 Elmos in those years so they aren't in the original data set. We know what's going on there, and in this case I'm just going to let it be instead of filling in the years with 0s or 4s. The trend is clear enough.
Here we go, real charts for each of those names, sex M, exact spelling, from their first appearance in the data until 2024:
[ID: Kermit began trending downward in the 1910s/1920s, had a brief leveling out the 1940s, and then went steadily downward. It has a brief resurgence around the introduction of the character Kermit the Frog in the late 50s/early 60s but then resumes downward until it is nearly gone.]
The Grover chart is fun because I made all the charts top out at the same count (500 individuals) for these, and Grover was more popular than that at its peak, so the tops are cut off on its individual graph.
Grover is falling sharply through the 1920s, levels out until around 1950, then falls sharply until it nearly vanishes. No major change in downward trend in 1969.
Elmo peaked around 1920 and then fell off a cliff. The downward slope gentled out around 1930 but kept falling steadily until around the 1980s, where it leveled out at a nearly-nonexistent count.
Sesame Street had absolutely nothing to do with the dropoff in these names. ALL of them were already on the downslope.
In 1969, there were 116 Kermits, 105 Grovers, and 35 Elmos in the data. Those are not popular names. For reference, the most popular name for boys in 1969 was Michael (85,159); 100th was Henry (3,057); 1000th was Titus (60).
In 1974, 5 years later, there are 71 Kermits, 74 Grovers, and 18 Elmos. Percentage-wise, those are huge drops, but RAW NUMBERS-wise, they aren't huge drops at all. (We are ignoring for the second that Elmo hadn't debuted yet.)
Now we're not ignoring that anymore:
1979 Elmos (pre-debut): 15. 1981 Elmos (post-debut): 16. 1985 Elmos (Elmo starting to become popular character): 12.
In fact, the most Elmos in the Social Security data in ANY year since 1980 is 18 (1983 and 1994). This is not a popular name. At all. At any time since the character debuted whatsoever. The most popular name for boys in 1980 was Michael, again (68,687); 100th was Billy (2,991); 1000th was Mohammed (82).
For funsies, here's all of them from 1880-2024. Any years with no data for a name have been set to 0 for the purposes of this graph and this graph only:
On this, you can see that these names all peaked around the same time, and all declined since, with slightly different patterns.
