Signature Books has recently issued a press release about an article that I co-authored along with Craig Foster and Gregory Smith. In the article we “strongly suggest[ed] that … the age of Joseph’s wives was well within the norm for their time and place on the nineteenth-century American frontier.” The public relations employees of Signature Books are certainly entitled to disagree like Todd Compton did in a contrasting essay in the same book, The Persistence of Polygamy. To be clear, I agree with my critics that it is entirely inappropriate, not to mention illegal [thanks Last Lemming SteveDensley for correction], in today’s society to marry a 14 year old young woman.
At least one of Signature’s criticisms may have a degree of merit, perhaps my plots and graphs are too hard for the average reader to understand. That is one of the downsides of using methods found in academic demographics journals to reconstruct marital statistics in 1840 Illinois. I think Signature writers may have been a bit hasty in drawing conclusions from the data, but their loyalty to George D. Smith (whose opinion is critiqued in our article) is commendable.
The plots and tables in our chapter were meant to place the age distribution of Joseph Smith’s wives in as large of a context as possible. The statistics on marriage vary greatly across US regions and across decades. The average reader should be able to tell that Joseph Smith’s teenage bride’s ages compare reasonably well with frontier area dynamics, with the exception of the percentage of 14 year olds. The same reader could also focus on the data from recent times or from Northeastern states and argue that those should figure more prominently in their judgments against Joseph Smith.
For me the fairest test I could put Joseph Smith to was to 1) recover as best I can, the frequency distribution of first time brides who married in Illinois in 1840 and 2) to follow a common statistical method for evaluating a hypothesis, in this case evaluating whether it would be normal to find two fourteen year old brides in a set of 33 wives selected at random using statistics from the first step. In my test I utilized a worst case scenario for Joseph Smith wives. Using George D. Smith’s 1994 enumeration of 43 plural wives including 9 teens and 1 14 year old would have led to an even more favorable comparison for Joseph Smith. The numbers in the press release would fall somewhere in between.
So how does Joseph Smith’s teenage brides stack up against their 1840 Illinois peers? With the exception of 14 year olds they actually trail the cumulative frequency distribution: 14 (6.1% to 1.9% ), 15 (6.1% to 6.7%), 16 (15.2% to 15.6%), 17 (21.2 to 27.4% ), 18 (21.2% to 40.3%), and 19 (30.3% to 52.6%). If you were to select 33 brides at random many different times at these rates and track the number of 14 year olds, Joseph Smith’s brides would rank between the 86th and 98th percentile. Either edge falls below common 2 tailed-thresholds for testing hypotheses (the 2 standard deviation rule or the 5% rule.) In other words, it turns out that Joseph’s profile is not significantly different than that of his ambient society, hence one can safely conclude it was normal.
Now I realize that some readers might not be able to follow that analysis. The bar chart below might help conceptualize the experiment described above. Although the sample sizes vary widely, Craig Foster randomly selected some counties, mostly from the Illinois-Iowa region. I charted those in comparison to my Illinois Census reconstruction, of marriage cohorts in 1850 and 1880, and against some Mormon marriage cohorts. For good measure we added a county more out on the frontier (Hitchcock, Nebraska) that had about the same sample size as Joseph’s set. Not surprisingly, it had a higher frequency of 14 year old brides than those married to Joseph Smith.
While the Signature press release does make a number of good points, readers can find a fuller contextualization and discussion of those points within the original article itself. The Signature reaction failed to engage my statistical definition of marital age normalcy at all. My co-authors and I spent considerable effort collecting and analyzing data and surveying the relevant literature, so I would plead with the historical community to give us a fair hearing. I will have considered it time well spent if individual Latter Day Saints ^Latter-day Saints [hat tip Ardis] take a position on the issue that falls somewhere in between Todd Compton’s and ours. While Compton disagrees with us on how normal the 14 year old plural brides were compared to their monogamous peers, he has a brilliant way of explaining what he regards as an abnormality (Helen Mar Kimball’s marriage). We also seem to be in agreement that extremist commentary on the subject has been a disservice to the public.
I haven’t read your paper and have read only as much of the press release as is quoted over on Mormon Heretic. But I am not going to offer a general critique of your methods so that shouldn’t matter. I’m just trying to get my head around your claim that because the share of Joseph’s wives who were 14 falls between the 86th and 98th percentiles of the samples you drew from the 1840 Illinois population, that his profile qualifies as “normal.” It seems to me that you should be testing the hypothesis that Joseph’s profile was at the 50th percentile–what a typical person would regard as “normal.” In all of your samples, Joseph’s profile never came close to the 50th percentile–it was uniformly higher. You could draw samples until the millenium and you still would be very unlikely to find one in which Joseph’s profile was at the 50th percentile. So I don’t see how it can be characterized as “normal.” Asserting that the central 95 percent of a population is, by definition, normal is a misapplication of the 5% rule.
None of which says much about the larger question, which seems to cover all of Joseph’s wives and (at least in the eyes of some) the age of Joseph himself. But I think on this narrow question, your argument fails.
“To be clear, I agree with my critics that it is entirely inappropriate, not to mention illegal, in today’s society to marry a 14 year old young woman.”
Note that even today, it is not illegal in every state to marry at 14. Many states allow marriage at 14 with parental consent, including Utah up until 1999 when the age was changed to 16 with parental consent and 15 with consent of the court. (See U.C.A. section 30-1-2, and http://en.wikipedia.org/wiki/Marriageable_age.)
Of course, your point is to place the matter of Joseph’s marriages in historical context. I haven’t studied this issue much, but I was interested to find, through a quick internet search, that the age of consent in England was 12 until 1875, when it was raised to 13. (See page one of http://www.parliament.nsw.gov.au/prod/parlment/publications.nsf/0/165f479bcf457c38ca256ecf0009f629/$FILE/bf21-1997.pdf.)
Perhaps you should analyze how common it was for adult men to marry minors for whom they had been appointed legal guardian, since Joseph did so on more than one occasion. For that matter, you could do an analysis on how many minor females were married to adult men who alleged deity had directed them to so marry. Yes, I’m sure it’s all totally “normal,” so long as it’s Joseph Smith instead of David Koresh.
Thank you, Keller, for assuring us that readers who are not persuaded by your arguments are simply too stupid to follow the pretty pictures. That’s certainly the way to get someone to reread and reconsider your thoughts.
I don’t blame you for feeling the need to respond to Signature’s press release, but I hope this ends it, on both sides. It’s a petty squabble, a mountain out of a molehill, and it doesn’t reflect well on either side. (And yeah, I’ve already said as much in correspondence with Signature contacts.)
Has the Church made some stylistic change of which I am unaware? Both your post and Signature’s press release refer to “Latter Day Saints.”
Thank you, Keller, for assuring us that readers who are not persuaded by your arguments are simply too stupid to follow the pretty pictures. That’s certainly the way to get someone to reread and reconsider your thoughts.
As someone who usually tends to butt heads with Ardis, I have to stand alongside her on this one. I had the same reaction to that part of Keller’s essay. THANK YOU, Ardis, for speaking the truth plainly.
Put my scolding on hold, Keller; I’ll get back to you after rethinking. If Nick Literski agrees with me, there’s something wrong with my analysis.
LOL, Ardis!
I was watching an old movie/musical this weekend, a good old Gene Kelly toe-tapper called “An American in Paris” (1951, MGM) interestingly enough, the man who was intending to wed Lise Bouvier was her guardian Henri Baurel, who kindly releases her from her engagement when he learns of her love for Jerry Mulligan (Gene Kelly).
It seems that in 1951 the concept of marrying a guardian wasn’t considered all that taboo.
So, setting the statistics aside, check this one out:
http://marquee.blogs.cnn.com/2011/06/21/lost-actor-marries-16-year-old-girlfriend/?hpt=hp_t2
The commenters who are not denouncing the guy as a pedophile are insisting that the girl is actually older than 16 and that they are claiming she is younger as a publicity stunt. In other words, abnormal apparently has its advantages.
Thanks Ardis for your critique of how I come across to readers. I wish I had a fraction of your writing skills at three in the morning when I was trying to think of a way to segue in between points. You are right I should put more faith in my readers, many who are no doubt more intelligent than I. It is sometimes too easy for me to slip into my school teaching days thinking if I could just figure out how to explain things better (if manipulating equations don’t work, try drawing a picture, etc.) I could help others come up with the right solution.
But in reality, it is me getting schooled right now as everybody who wants to gets to critique my paper and tell me what I have done wrong. Thanks for helping me get the proper perspective and thanks for speaking up about recent exchanges.
Lemming,
First let me be the first to admit that someone more intelligent than I should be able to propose a better definition of normalcy and a more rigorous statistical test than the one I employ.
If my definition of being in the middle 95% seems to accommodating, hitting the 50% mark is much too narrow. At best only one person could claim to be normal if we were to parse up marital ages finely enough. That would make it normal to be abnormal. In order to make your proposal workable you have to give a range about the 50% mark.
I prefer to work with the 5% or 2 sigma rule, but let me try to explain why I think it is applicable.
Mormon Heretic brought up a good point about whether comparing a person to a region of the country might be an abuse of statistics. I gather though, that idea is precisely what drives statistical nuptiality studies. The timing of marriage for an individual is modeled as a random process. Some social scientists like Coale and Hernes pioneered studies that mixed empirical observations with mathematical derivations from theories regarding marriage to uncover probability distributions that have uncanny predictive power for individuals once some key parameters are found. Both used Census data to determine the parameters of their distribution. I went with Coale’s model in the paper as it is more widely used and easier to work with.
The theoretical underpinnings of Coale’s model is that marriage timing is the sum of three random variables each with their own distribution. The age of marriage equals the minimum age at which a young woman becomes viewed by society to become eligible for marriage + the time elapsed to meeting the future husband + courtship time. Coale associated the minimum age of eligibility with menarche and community standards/laws, but he first gave it a definition of the earliest age at which a statistically significant percentage of marriages occur. Later he teamed with McNeil and found the minimum age of eligibility to the mean – 1.73 times the standard deviation.
Joseph Smith never married below Coale’s minimum age of eligibility, so by Coale’s definition none of the ages of Joseph’s wives were socially unacceptable. I could have easily stuck to that benchmark but I wanted to press further
But it was important enough to make a point of including the minimum age of eligibility on some of the charts and graphs.
The paper uses state of the art methods for properly parameterizing the random distribution that could be said to govern the marriage timing of young women in Joseph’s society. I think once a trustworthy distribution is in place it is straightforward to ask whether a set like Joseph’s could have been randomly generated and hence completely explainable by randomness associated with the sample size.
I can’t conceive of a way to devise a similar statistical test that focuses on men aged 34-38 preferences that Nick suggests. That is way outside my area of expertise to compare two small sample sets that both don’t have a conventional expression for their distribution. The literature on age difference seem to use a different set of tools to analyze trends and underlying factors.
It has been my understanding that Todd Compton and other historians on Joseph Smith’s practice of plural marriage generally agree that there is no evidence for any sexual relationships between Joseph Smith and his teenage brides. Is that correct or am I misremembering? If so, then will somebody please remind me again what the fuss is about? It is one thing to say that Joseph Smith had some whacky religious beliefs about sealing power, eternal marriage, etc. But it is an entirely different thing to accuse him of being a lecherous pedophile, a la Hitchens, O’Donnell and Krakauer.
I don’t demand that Joseph’s profile be exactly at the median. If you had drawn all of your samples and found that Joseph’s profile fell at or below the 50th percentile in X percent of those samples, then you could state that there is an X percent chance that Joseph’s marriages to 14-year-olds was not abnormal. (To be clear, if Joseph’s profile were truly at the median–i.e., perfectly normal–X would be equal to 50, not 100. Thus, it would be unfair to exploit a value of X equal to, say, 70 by claiming that there is a 30 percent chance that Joseph’s marriages to 14-year-olds was abnormal. The only way to get to X=100 is to have no 14-year-old wives at all.) But your results show X equal to zero, which to me means that Joseph’s profile was unambiguously abnormal.
Last Leming I see what are saying. I have already provided the number you seem to be looking for, x is around 86.5%. It can be calculated exactly when the % of 14 year old brides in society is assumed known. That means I could go simulate at random, say 1000 sets of N= 33, and expect about 130 to 140 (around 1 in every 8 ) would be as least as extreme as Joseph Smith. Note the Hitchcock set is more extreme than Joseph Smiths (2/32) vs (2/33).
So Joseph Smith doesn’t fall in your range of 0-50%, but your test sounds like it is geared for making median/below median distinctions. Mine was in line with testing the hypothesis of whether Joseph Smith’s distribution (polygamy aside) could be considered a product of their society.
See for example:
http://en.wikipedia.org/wiki/Statistical_significance
No, that’s not what I had in mind but I don’t think we’re going to resolve this over the internet.
Also, the correction you made should be attributed to SteveDensleyJr, not me.
I have spent a great deal of time, in my genealogical research, combing over US census records from the 1800’s. Though I have not tabulated any exact statistical information, I have noted that it was not at all uncommon for females to be in the 15-16 yr old range at the time of their marriage. Indeed, I have 2 great-great grandmothers who married at the age of 15. Their husbands were ages 30 and 25, respectively. Both had many children and were married up until their husbands deaths. Neither family was Mormon, and their marriages occured in the mid 1850’s. People normally matured (mentally) at a much earlier age in those days and so many females were ready for marriage at these younger ages. My theory is that it was due to the much greater physical hardships that people had to suffer through in previous centuries compared to nowadays, which caused them to mature sooner.. to basically grow up sooner. They had to take on greater responsibilities and much younger ages than now. So what is looked on as inappropriate nowadays (rightly so), was more the norm in the 1800’s and earlier. I have also heard it said that in many places a woman was considered an “old maid” if she hadn’t married by the age of 20.
I didn’t take the essay as stating that those who disagreed were just those too stupid to understand why he was applying the 5% or 2 sigma rule.
Been a while since I’ve read a good statistical analysis.
Anyway, appreciate your taking the time to lay out the numbers and your math.
I think Stephen Smoot’s queery is the appropriate question and approach. Did Joseph have sexual relations with any 14 year olds or were these dynastic marriages of a completely different order than civil marriage? My research suggests that there is no evidence of sexual relations with any 14 years olds. Nor would we define his relations as “marriage” given the legal definition of agreement and contract we now use in Western jurisprudence.
Our own culture is in the process of redefining marriage (or defining it out of existence depending on how you see it). Joseph saw marriage very differently than a mere civil ceremony and in fact didn’t recognize the authority of government or other religions to define the marital relations. I suggest that his approach was entirely different than a civil marriage in his own day and time.
Thus, Nick’s suggestion that men who are legal guardians should not “marry” is appropriate given the legal relation that marriage now signifies — it is not inappropriate given the dynastic relations Joseph Smith was creating.
Congratulations, Keller, on you and your partners’ excellent work on your article, and on your backup comments. Having taken a number of statistics courses while pursing an MS in Finance, I could mostly understand (been many years ago when I studied) your arguments and assumptions. Yet, we should all know that comparing general monogamist society with a polygynist society is a hard sell, and combining that with religion certainly aggravates things further.
That said, I was somewhat surprised at many of the comments on Keller’s follow up – that there are so many that judge yesteryear’s cultural practices by today’s society’s mores. That because today’s youth in today’s modern society are incapable or generally irresponsible enough to not be able to take care of themselves let alone others, that the youth of yesteryear must also have been incapable as well. And, that young ladies of other cultures today are just as ignorant as western society’s girls. Thus, no one is or ever was mature/knowledgeable enough to get married at the age of 14. That nature is the problem.
Chalk another one up to political correctness?
Or, could it be that today’s modern society’s nurture is the problem vs. nature? That society’s/government’s “dumming down” march continues on in earnest?
“Hey, has anybody noticed it is getting pretty hot here in the kettle”?
Has the Church made some stylistic change of which I am unaware? Both your post and Signature’s press release refer to “Latter Day Saints.”
Since when are members of the Utah-based LDS Church the only members of the Latter Day Saint movement who have an interest in the history of Joseph Smith? The spelling convention “Latter-day Saints” and is specific to the Utah church only, while there are other interested parties in the Latter Day Saint movement at large. The original spelling was more appropriate and shouldn’t have been changed.