Hilary: the most poisoned baby name in US history

[This article was first published on Not So Standard Deviations » R, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

I’ve always had a special fondness for my name, which — according to Ryan Gosling in “Lars and the Real Girl” — is a scientific fact for most people (Ryan Gosling constitutes scientific proof in my book). Plus, the root word for Hilary is the Latin word “hilarius” meaning cheerful and merry, which is the same root word for “hilarious” and “exhilarating.” It’s a great name.

Several years ago I came across this blog post, which provides a cursory analysis for why “Hillary” is the most poisoned name of all time. The author is careful not to comment on the details of why “Hillary” may have been poisoned right around 1992, but I’ll go ahead and make the bold causal conclusion that it’s because that was the year that Bill Clinton was elected, and thus the year Hillary Clinton entered the public sphere and was generally reviled for not wanting to bake cookies or something like that. Note that this all happened when I was 7 years old, so I spent the formative years of 7-15 being called “Hillary Clinton” whenever I introduced myself. Luckily, I was a feisty feminist from a young age and rejoiced in the comparison (and life is not about being popular).

In the original post the author bemoans the lack of research assistants to perform his data extraction for a more complete analysis. Fortunately, in this era we have replaced human jobs with computers, and the data can be easily extracted using programming. This weekend I took the opportunity to learn how to scrape the social security data myself and do a more complete analysis of all of the names on record.

Is Hilary/Hillary really the most rapidly poisoned name in recorded American history? An analysis.

I will follow up this post with more details on how to perform web-scraping with R (for this I am infinitely indebted to my friend Mark — check out his storyboard project and be amazed!). For now, suffice it to say that I was able to collect from the social security website the data for every year between 1880 and 2011 for the 1000 most popular baby names. For each of the 1000 names in a given year, I collected the raw number of babies given that name, as well as the percentage of babies given that name, and the rank of that name. For girls, this resulted in 4110 total names.

In the original analysis, the author looked at the changed rank of “Hillary.” The ranks are interesting, but we have more finely-tuned data than that available from the SSA. The raw numbers of babies named a certain name are likewise interesting, but do not normalize for the population. Thus the percentages of babies named a certain name is the best measurement.

Looking at the absolute chance in percentages is interesting, but would not tell the full story. A change of, say 15% to 14% would be quite different and less drastic than a change from 2% to 1%, but the absolute change in percentage would measure those two things equally. Thus, I need a measure of the relative change in the percentages — that is, the percent change in percentages (confusing, I know). Fortunately the public health field has dealt with this problem for a long time, and has a measurement called the relative risk, where “risk” refers to the proportion of babies given a certain name. For example, let’s say the percentage of babies named “Jane” is 1% of the population in 1990, and 1.2% of the population in 1991. The relative risk of being named “Jane” in 1991 versus 1990 is 1.2 (that is, it’s (1.2/1)=1.2 times as probable, or (1.2-1)*100=20% more likely). In this case, however, I’m interested in instances where the percentage of children with a certain name decreases. The way to make the most sensible statistics in this case is to calculate the relative risk again, but in this case think of it as a decrease. That is, if “Jane” was at 1.5% in 1990 and 1.3% in 1991, then the relative risk of being named “Jane” in 1991 compared to 1990 is (1.3/1.5)=0.87. That is, it is (1-0.87)*100=13% less likely that a baby will be named “Jane” in 1991 compared to 1990.

(Note that I’m not doing any model fitting here because I’m not interested in any parameter estimates — I have my entire population! I’m just summarizing the data in a way that makes sense.)

So, for each of the 4110 names that I collected, I calculated the relative risk going from one year to the next, all the way from 1880 to 2011. I then pulled out the names with the biggest percent drops from one year to the next.

bigdrop

#6?? I’m sorry, but if I’m going to have one of the most rapidly poisoned names in US history, it best be #1. I didn’t come here to make friends, I came here to win. Furthermore, the names on this list seemed… peculiar to say the least. I decided to plot out the percentage of babies named each of the names to get a better idea of what was going on. (Click through to see the full-sized plot. Note that the y-axis is Percent, so 0.20 means 0.20%.)

names

These plots looked quite curious to me. While the names had very steep drop-offs, they also had very steep drop-ins as well.

This is where this project got deliriously fun. For each of the names that “dropped in” I did a little research on the name and the year. “Dewey” popped up in 1898 because of the Spanish-American War — people named their daughters after George Dewey. “Deneen” was one name of a duo with a one-hit wonder in 1968. “Katina” and “Catina” were wildly popular because in 1972 in the soap opera Where the Heart Is a character is born named Katina. “Farrah” became popular in 1976 when Charlie’s Angels, starring Farrah Fawcett, debuted (notice that the name becomes popular in 2009 when Farrah Fawcett died). “Renata” was hard to pin down — perhaps it was popular because of this opera singer who seemed to be on TV a lot in the late 1970s. “Infant” became a popular baby name in the late 1980s for reasons that completely defy my comprehension, and that are utterly un-Google-able. (Edit: someone pointed out on facebook that it’s possible this is due to a change in coding conventions for unnamed babies. This would make more sense, but would also make me sad. Edit 2: See the comments for an explanation!)

I think we all know why “Iesha” became popular in 1989:

“Khadijah” was a character played by Queen Latifa in Living Single, and “Ashanti” was popular because of Ashanti, of course.

“Hilary”, though, was clearly different than these flash-in-the-pan names. The name was growing in popularity (albeit not monotonically) for years. So to remove all of the fad names from the list, I chose only the names that were in the top 1000 for over 20 years, and updated the graph (note that I changed the range on the y-axis).

names_trimmed

I think it’s pretty safe to say that, among the names that were once stable and then had a sudden drop, “Hilary” is clearly the most poisoned. I am not paying too much attention to the names that had sharp drops in the late 1800s because the population was so much smaller then, and thus it was easier to drop percentage points without a large drop in raw numbers. I also did a parallel analysis for boys, and aside from fluctuations in the late 1890s/early 1900s, the only name that comes close to this rate of poisoning is Nakia, which became popular because of a short-lived TV show in the 1970s.

At this point you’re probably wondering where “Hillary” is. As it turns out, “Hillary” took two years to descend from the top, thus diluting out the relative risk for any one year (it’s highest single-year drop was 61% in 1994). If I examine slightly more names (now the top 39 most poisoned) and again filter for fad names, both “Hilary” and “Hillary” are on the plot, and clearly the most poisoned.

more_names_trimmed

(The crazy line is for “Marian” and the spike is due to the fact that 1954 was a Catholic Marian year – if it weren’t an already popular name, it would have been filtered as a fad. And the “Christin” spike might very well be due to a computer glitch that truncated the name “Christina”! Amazing!!)

So, I can confidently say that, defining “poisoning” as the relative loss of popularity in a single year and controlling for fad names, “Hilary” is absolutely the most poisoned woman’s name in recorded history in the US.

Code for this project is available on GitHub.

(Personal aside: I will get sentimental for a moment, and mention that my mother was a at Wellesley the same time as Hillary Rodham. While she already knew that she wanted to name her future daughter “Hilary” at that point, when she saw Hillary speak at a student event, she thought, “THAT is what I want my daughter to be like.” Which was empirically the polar opposite of what the nation felt in 1992. But my mom was right and way ahead of her time.)

Update: This seems to be an analysis everyone is interested in. For perhaps the first time in internet history, Godwin’s Law is wholly appropriate.

names_adolf


To leave a comment for the author, please follow the link and comment on their blog: Not So Standard Deviations » R.

R-bloggers.com offers daily e-mail updates about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.
Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

Never miss an update!
Subscribe to R-bloggers to receive
e-mails with the latest R posts.
(You will not see this message again.)

Click here to close (This popup will not appear again)