Mate Selection: A Study Summary

I wrote this in late March 2006. I decided to break this off from my section on Lent 2006 Musings because it got very long and contains several tables. If you jumped in from Google, I'm Adam, 31, unmarried and searching, Catholic, Ph. D. student in statistics, ... well, just click on About Me or the button at the bottom of the page if you want more information.

In March, I read the book "Why There are No Good Men Left" by Barbara Defoe Whitehead. It's not a quality book. First, the title is misleading; the book is almost exclusively about women, and the changes in mating patterns as women have become more financially and socially independent. There's lots of whining about the lack of tall, handsome, dominant males with six figure salaries. One Amazon reviewer puts it better than I: "Perhaps Ms. Whitehead could write a sequel: "There are Plenty of Good Men Left, but You're too Vain to Notice Them."

Of course, if one pursues college and career and travel instead of a husband, the change will be wrenching. Of course that's become more common and Title IX mattered. Of course cohabitation without serious discussion of its meaning will lead to confusion. Of course the market has fewer "good men" at the same age level, because the higher quality men in the market were already mostly taken. That's the advantage of entering the market early. Of course there's a Jane Austen referentital longing, this time to Henry Tilney. And I didn't appreciate the reminder on page 38, that older never-married men are suspect because they have never been married. In a pithy quote, "The odds are good, but the goods are odd."

It should surprise no one that I found the statistics and numbers most interesting. A summary of the 2000 Census, percentage of college-educated women unmarried by age, was interesting. The proportion declines rapidly between 23 and 27, decreases more slowly up to 32 or 33, then remains about constant as the number of divorces increases to approximately match the marriage count.

Since I'm interested in what females (and males, but more females) find attractive, I also tracked down a paper reference: "A Half Century of Mate Preferences: The Cultural Evolution of Values." It's by Buss, Shackelford, kirkpatrick, and Larsen, and appeared in the Journal of Marriage and Family 63 (May 2001): 491-503. The authors found that various college students had answered the same questionnaire in 1939, 1956, 1967, and 1977, then ran surveys in 1985 and 1996. This makes a reasonably consistent sample of undergraduates over time. In each survey, the heterosexual students were asked to rate 18 mate characteristics on a four-point scale. I've listed them in the table below. I suggest using the scale, then fleshing out the ranks 1 to 18, with rank 1 as most important. Here's the scale:

My Results

Well, I can't ask you to do a test without providing my results. Or can I? You can mail your results to adam@twelvefruits.com. Well, if you really want them, here are my ranking and scores (scores in parentheses). It might surprise you.

Modern and Old Results

There have been some shifts over time, between the pre-TV days of 1939 and my contemporaries. Also, there are differences in gender. The greatest shift occurred between 1967 and 1985; although I report the oldest and newest rankings, you can almost think of a cliff. For both genders, the table reports 1939 and 1996 results. I don't have the 1939 mean scores, but I put the 1996 means in parentheses. My comments on the table results appear below.

• Mutual attraction made major gains, almost entirely between the 1957 and 1977 surveys. A different study confirmed the results, with the question "If someone had all the qualities you desired, would you marry this person if you were not in love?" In 1967, 31% of men and 72% of women said yes. In 1984, only 11% of men and 13% of women said yes. Over just one generation, marriage gained a mandate for romantic love.
• Education and intelligence also made a big leap, as did sociability. This agrees with the concept of mutual attraction. With lives less separate, conversation and activities together become more common, and those skills become more valued.
• Physical attractiveness was the last big gainer, for both males and females. As one would expect, it's more important for guys than girls, yet both recorded substantial increases. The authors suggest that television, with its relentless beauty, has influenced concepts of attractiveness.
• I think the undergraduates underestimate the importance of religious and political matching. It's how people get along day-to-day, much more than similar educational background. People can take the same classes yet believe radically different things, differences that cause tension. Compared to the students, I also swapped social status and financial status, showing concerns about my poor Appalachian upbringing but relatively few worries about employment prospects.


• Chastity has plummeted in importance. Actually, for modern students, I wonder if "lack of chastity" and sexual experience would be more important than virginity.
• Refinement and neatness decreased. Society has gotten more casual, I suspect.
• Women are willing to accept less ambitious partners - though not too slothful, as their mean score is still 2.39.
• The importance of household tasks for men has decreased. Given my leanings towards female equality, I consider this not a bad thing.

The Short and Long Term

The Buss study is extremely useful, and he's done other work comparing preferences across cultures, but we should try to consider other perspectives. I tracked down another survey report, by Pamela Regan, Levin, Sprecher, Christopher, and Cate, in a 2000 issue of the Journal of Psychology and Human Sexuality. (Among other things, Dr. Regan has run a study on homosexual mate preferences. But that's off topic.) The Regan study asked for importance ratings on 23 characteristics, like "Honest and Trustworthy" and "Physically Attractive" and "Similar to Self on Background Characteristics". Rankings were in terms of relative percentage rank, from 90, "I would like my like my partner to be above about 90% of other same-sex individuals on this dimension", down to 40, "I would like my partner to be below average on this dimension compared to other same-sex individuals." This isn't the best design; after all, who wouldn't want a partner at the top in everything? Fortunately, the instructions overrode natural tendencies, and the respondents used the levels as relative ranks.

Half the college students answered the questions about a short-term partner, and half answered the questions about a long-term romantic relationship. The top three individual scores were the same for both time frames, "Honest and Trustworthy", "Healthy", and "Warm and Kind". However, most of the analysis was done on groups of characteristics, not single scores. Exploratory factor analysis with varimax rotation identified five distinguishable sets. The statistician in me was very happy. This is a very proper, well-done application of a statistical technique. I wish that there be more. Anyway, now that I've geeked out, let's go back to the analysis.

The five attribute dimensions are "Social Status", "Socially Appealing", "Similarity to Self", "Sexually Appealing", and "Expressive/Outgoing Personality". If you want to rank them yourself, the scale ranges from a minimum of 40 to a maximum of 90. The results appear in the table below, with the mean percentile rankings by gender. There are a few differences between short-term and long term, though not as many as I would have guessed. I describe those differences below.

Sexual desire, as one might expect, was more important in the short term than the long term, 77.7 to 75.1. This effect was primarily in females, though; males were roughly equal, while females were willing to accept less sexy males long term.
Social appeal was more important in long term relationships, 78.9 to 75.6. Actually, long term social appeal (intelligence, ambition, education, honesty, warmth and kindness) was higher than long term personality. Similarity was also more important in the long term, but still less than personality, sexuality, or social appeal.

Budget and Priorities

Well, we've got some survey results. Unfortunately, they're not completely tenable. First, all of the categories in the Regan study are over 50, which means that on average, people want a mate above average. This is not Lake Wobegon, so that's not possible for everyone. This means that at least some people get less than what they want. They either misvalue others' ratings, or settle. It doesn't take a brilliant researcher to discover the fact that people misvalue others, which makes that quite likely. It's also easy to want above-average mates when there are almost no constraints, as in the Regan study. Ranking, as with Buss, provides more structure, but the point scores can still be inflated. What happens when tradeoffs are mandatory? Social psychologists have been busy here, too. The Journal of Personality and Social Psychology contains a 2002 article by Norman Li, J Michael Bailey, Douglas Kendrick, and Joan Linsenmeier on this subject.

People were given budgets of "mate dollars" and characteristics to allocate the budget across. When constrained, women spent relatively more on resource acquisition, while men bought more physical attractiveness. When budgets expanded and choices were less constrained, there was less spending on these characteristics and more on others, such as creativity.

I adapted the following example from their studies. Say there are the following five budget items. It costs 1 dollar to ensure a mate has the 10th percentile in an item, $2 for the 20th, $5 for a median value, and $10 for a tip-top result. You have 20 dollars to allocate among the five items: Physical Attractiveness, Social Level, Creativity, Kindness, and Social Liveliness. What do you buy? Just quickly thinking, I purchased $4 of attractiveness, $2 of social level, $3 of creativity, $7 of kindness, and $4 of social liveliness. I put the mean results (out of $20) for men and women in the table below.

The entire article has a wealth of information, over multiple experiments, including populations other than college students. Both sexes consistantly prioritize three characteristics above others. Two are the same for women and men, intelligence and kindness. Women desire for social status and resources, while men want physical appearance.

The Summary

• Modern people value mutual attraction and social interaction more than pre-television folks, while chastity and refinement are less important.
• Both sexes say they greatly value intelligence and kind, pleasing personalities. This is near primary in the long term.
• For men, physical attractiveness is close to primary. For women it matters less, but is still in the middle ranks. Its value has increased as TV has become more ubitiquous.
• While actual social rank is not important, women generally require mates to have material resources, and value ambition and dominance. As the "providers", men worry less about female gathering skills.

Three Problems

1. Various studies do not use a standard set of traits. Furthermore, studies identify important things not picked up by other researchers, because they're not on the list. I'd like to settle on a common list. I'll propose a list of seven: Intelligence, Physical Attractiveness, Kindness/Pleasantness, Similarity to Self, Social Status, Resources/Financial Prospects, Outgoing/Social Personality. Running a budget experiment with $25 and $50 (out of $70) would be very interesting.
2. Many of the experiments are Lake Wobegon surveys. It's easy to say one wants a man with everything, and even when ranking, one can still rank everything very high. This is most obvious in the Regan study, but also apparent in the Buss three point scales. Li's budget idea removes one problem, but remains a survey. We really need to study actual selections, how people view their significant others versus their ideals.
3. In the studies, qualities are quantifiable and available. In real life, we have to make judgments on limited information, without calibrated scales. The outward characteristics - looks, money, and social status - are easiest to spot. Basic personalities and intelligence take some time, but not much. The internal qualities of similarity and kindness don't evaluate quickly. While females and males may say they desire internal traits, can those decisions be accurately made in early relationship stages? Like for a coffee date? Or do folks rely on the more precise external traits? We might design sequential studies Another idea might be to make some information more reliable than others, and compare choices to the full-information case.

Written March 2006. Thanks for reading - I hope you found it illuminating and interesting.