Casual Fridays: Are people with similar names more likely to pair up?

I don't know if you have some good phonebook based name database, but it might be easy (easier?) to use the social security database, which has the incidence of the top 1000 names for each sex by year of all babies getting social security numbers (which is most people).
Lisa

Haven't you introduced a participation bias in your survey by telling everyone the goal? People with like-named spouses or like-named grandparents may be more eager to join.

Sloppy? or am i just no fun?

PS - my wife's name is not willamina

I answered the survey, but I don't think very highly of the hypothesis.

On an interesting side note, one of my ex-boyfriends has the same last name as me. When people said that was convenient, I used to joke that if we got married we would use the double-barreled version.

Why not ask about parents' names, too?

Were you thinking about letter-name-effect research like this when you hypothesized that there's be relations among the names?:

Nelson & Simmons (2007). Moniker malidies: When names sabotage success. Psychological Science, 18(12), 1106-1112.

http://en.wikipedia.org/wiki/Name_letter_effect

This was harder than I expected. I didn't know whether to put the short or long version of my name, or the English or Polish version of my Grandmother's name.

My immediate family may go against the grain if your hypothesis is accepted -- the marriage that didn't last was between Michael and Michelle, but the less similarly named pairs have all been together over 20 years now.

How do you expect to assess the similarity between names? That seems to me to be the major obstacle even if you had a random sample of couples.

If you have a good way of answering that, and you find that people are somehow more likely to pair up with people with similar names, then that could be because people usually pair up with others of similar cultural backgrounds.

CogProf:

We discuss that study here as well.

Cathy:

Good points -- I trust you did the best you could. It's a weakness of this study design, but there's always a balance between keeping it simple and "casual" and getting good results.

Maria:

How do you assess similarity? Good question. I was initially planning on comparing how many of the first letters are the same (i.e. is the first letter the same? Is the second letter the same? And so on.).

Then, depending on how ambitious I was feeling next week I might consider looking at how many other letters are the same in each name. Any other suggestions?

I agree that if we do find an effect it may because people seek others from similar backgrounds. Or because the popularity of names follows trends (i.e. in the 1920's names like "Vern" and "Verna" were popular, and now names like "Austin" and "Ashley" are popular.)

What about middle names, Dave? My maternal grandmother and grandfather both had the same middle name -- Lee. Turns out that "Lee" was a long-time "family" name (middle or first) in my maternal grandfather's family history -- I looked back and for many generations, Lee was almost always there in some form. I think it's kind of awesome that they ended up together, both with the same middle name! (My maternal grandmother actually goes by "Lee" instead of her given first name.)

My immediate thought was, "nope. no similarity" but when you involve the grandparents, I realized that my husband's name and one of my grandfather's name is similar. Never gave that a thought before. Hmmm.

I just realized that I abbreviated my maternal grandmother's name. At least one of your answers should be Rosabelle, not just Rose.

A couple in my extended family were named Denise and Dennis. They didn't last quite 10 years.

Another configuration; one of my uncles married a woman with the same name as his sister; my daughter did the same thing, marrying a man with her uncle's and her brother's name, and my nephew married a woman with my name.

Certain names seem to be prevalent in specific cultural groupings, even to specific decades of those groupings. So there would be a high probability of duplications around those names.

It does seem a bit of a flaw not to mention how the relationships of the grandparents were truncated, as far as longevity of relationships go. Surely a breakup due to incompatibility would mean one thing to your hypotheses, where the relationship ending because of the death of one of the partners would mean something else again.

Divorce was considered a source of shame back in the day, which may also futz up the results....

My husband and I have names starting with the first two letters, and so do his parents. I didn't think anyone else had similar letters or sounds as I started to fill it out, but then I realized that my maternal grandparents have matching sounds. If you just go by letters, though, it's not going to work. (E.g., Karl and Cara, George and Josephine, etc.)

One way you could run the analysis:

A. Compare first letter / letter phrase (e.g "sh"):
1 - different first letter
2 - similar first letter (e.g. Curt and Karen)
3 - same first letter

B. Compare next four non-vowel letters
1 - none the same
2 - one the same
3 - more than one, less than all the same
4 - all the same

C. Compare the order of the letters that are the same from B.
1 - none the same / responded to B. 1 or 2.
2 - one two in a row
3 - three in a row / two groups of two in a row
4 - four or all in a row (if all is less than 4)

D. Compare the total length of the name to the standard deviation of random names
1 - more than 2 Stdev difference in length (eg. Poe and Christopherschompskito)
2 - less than 2 but more than 1 stdev in length
3 - within one letter, but not same length (this might not be here, depending on stdev)
4 - same length

etc. Of course, you'd compare each name in each group. For example, I am Mike and Shira and my grandparents are Bob and Sonya. Sonya is similar to Shira in A. and D.

A friend of mine, in geeky anticipation of his first child, used a computer program to compare male and female first names by how they *sound*, by running the text through an audio processor.

Download spreadhseets here:

http://slacy.com/blog/2007/05/the-ultimate-baby-name-explorer/

Looks like what you're really after is the masculine and feminine of the same name (e.g. Vern and Verna, Dennis and Denise). My husband's name is Justin, but aside from the first letter, the names are really not that similar.

My parents have very similar names. My mom's name is Sally and my dad's nickname is Sam (no one calls him Francis not even his mother).

There are a lot of computational ways to compare words for similarity. You might want to start with a measure of edit distance, which counts the number of character additions, deletions, and substitutions needed to transform one word into another. It might behoove you to use Soundex to transform the written words into a sequence of phoneme representations before applying an edit distance metric, since it seems to me that people think of words as being "similar" more based on how they sound than how they're spelled.

You may have set yourself up for more work than you realized...

My links got eaten. They were to the Wikipedia pages for "edit distance" and "Soundex", so you can find them yourself.

Hi dave,

I was born in Israel and my parents in Iraq.
i hope the cultural difference wont lower the effect.

:)

lyleobt -- thanks for the links. I fixed them (you need to put urls in quotation marks).

I have a feeling there won't be an effect, but if there is, what will it say about people whose names have no opposite-sex counterpart? Would they be more likely to be single? Or gay? With all those Jameses out there, is it now advantageous to be called Jemima?

Friends of ours are named "Andrea" and "Andreas" (in Germany), they just started their family with a first child. From my experience here, this is an extraordinary coincidence and always a funny story to tell.

I'm curious about the results!

what about sibling names? my husband has the same name as my brother and i have a friend in the same position.

Just to point out that the last question about the length of the marriage of the paternal grandparents asks how long the "maternal" grandparents were married (or is it the other way around! )

The test doesn't ask for any assessment of how happy the marriage was. Which seems important, given the greater focus on marriages staying together back in my grandparents' time.

I've just realized that my answers may also be misleading -- my paternal grandmother's name is Virginia, but she may have once gone by Jenny, and my grandfather was John.

And i've just realized that I should have paid attention to the date. ;-)

Think about last names, too. My last name is very similar to my wife's "maiden" name, which she continues to use. (No, they're not identical, and no, we're not related!) I have a couple of friends who have/had similarly close last names prior to marriage.

I speculate that if such people are married more frequently than they "ought" to be it is because of curiousity or something to talk about that gets people talking, and then one thing leads to another. This is obviously a way people could sort themselves, as it were, based on first names, too.

In our case, we were next to one another on an alphabetical list. In the case of high school, people might be seated near one another.