Unequivocally. A big chunk of our time was spent a.n.a.lyzing the bad-news studies with higher divorce rates so we could understand where the 50 percent came from in the first place. Looking at all these studies closely was how we first realized that 50 percent was only a projection, not an ironclad measurement. Our goal was never to prove any experts wrong, because some of the earlier projections from the late 1970s honestly seemed plausible at the time, given how much divorce was increasing during those years. We just kept digging to find the whole truth and the current reality about marriage as well as we could. And as you saw in chapter 4, our research led us to rigorously research and debunk a few commonly stated pieces of good news as well (such as those who pray together have an ultra-low divorce rate).
Someone much smarter than we are once said it is Truth that sets us free. Marriages would not be helped by our delivering inaccurate good news! We simply searched for the truth, and that is what we have tried to present to the best of our ability.
One Divorce for Every Two Marriages Each Year
How can you state the divorce rate isn"t 50 percent when there are half as many divorces as marriages annually?
The people who get married in any given year are not the same people getting divorced in that same year (reality television stars aside), so half of those marriages are not ending. Look at it this way: If a pastor marries twenty couples in one year, but then has ten couples in his church divorce in that same year, is the church"s divorce rate overall 50 percent? Of course not! (Unless there are only twenty couples in his church total.) Now, to keep the math simple, let"s say that church has six hundred married couples. So out of six hundred married couples, ten have divorced in one year. That means that church experienced a 1.6 percent rate of divorce for that year (10/600). Now, even if that pastor never marries another couple in that church ever again but ten more divorces occur every year, it would still take that church thirty years to reach a ratio of 50 percent divorced in their church.107 Nationally we have averaged about two marriages for every one divorce for decades.108 In our example, that pastor"s congregation may experience ten divorces every year, but the congregation would also probably continue to have twenty marriages added every year and thus would never get close to a 50 percent divorce rate. It works the same way in cities, counties, and the whole country.
What about the divorces and marriages that take place in my city each year? Is there a way to calculate the divorce rate from those numbers?
Let"s use an example from a pastor who asked Shaunti a similar question. This pastor lives in a county that experienced 14,327 marriages and 7,760 divorces the prior year. He argued that this means his county has a 54 percent divorce rate. But remember that we can"t look at the divorces that year as a percent of marriages that year (see above). Instead, you look at divorces that year as a percent of all the couples who could have gotten divorced that year, meaning all marriages in that county.
The correct way to calculate what percentage of couples experienced divorce for that year would be to take the number of divorces (7,760) and divide this number by the total number of marriages in that county. That county"s website demographics section says 46.9 percent of the 807,621 households in this county consisted of married couples, which equals 378,774 married households. So those 7,760 divorces equaled about a 2 percent divorce rate for that year. (A note of caution about all these types of numbers: the National Vital Statistics System uses two different populations when calculating marriage and divorce rates, and so might this county.)109 Other than divorces per year, the most likely way to find the prevalence of divorce in your area is if your local government publishes the "ever divorced" percentage from local surveys or vital statistics. And remember, the number of divorces will presumably include second, third, and fourth marriages, not just those of first marriages.
But with Everyone Living Together...
How relevant is your message about the good news in marriage when one of the main reasons divorce is dropping is because everyone is simply living together?
Even though the cohabitation numbers are alarming to sociologists and pastors alike, first keep the big picture in mind: according to Dana Rotz, as discussed in chapter 2, the primary reason why we"re seeing less and less divorce is because people are getting married at older ages.110 Second, although cohabiting has indeed increased, it actually has been very high for many years. Today, 74 percent of first unions consist of those who live together before or instead of getting married. But twenty years ago, it was already 58 percent. (The majority of the increase came from a very distinct group: those with high school diplomas rather than college degrees.)111 And third, most people still get married eventually, so the lower divorce rates are not just because people avoid marriage and live together instead. A Bureau of Labor Statistics (BLS) study shows over 85 percent of couples (90 percent of women) marry by age forty-six.112 The good news to remember about cohabitation is that at the end of the day, a vast percentage of people will tie the knot.
Validity of Current Studies (Example of a Big Study Needing a Big Adjustment)
How can you argue with a heavily doc.u.mented and highly used government report that says the divorce rate is still about 50 percent?
In chapter 2, we talked about the CDC report and how it presents the National Survey of Family Growth (NSFG).113 Both the executive summary within the CDC study itself and all the news reports about it reported that it found a 48 percent divorce rate. Even some of the experts we talked to accept the findings of the CDC/NSFG without question. But there is so much more to the story here and truly a lesson at looking at data closely.
When you compare the CDC/NSFG survey results to the results of one of the main census surveys-the Survey of Income and Program Partic.i.p.ation (SIPP)-for the same newlywed period, you see instantly that there is a problem: the CDC divorce rate is double that which the census found.
After Five Years of Marriage 2009 Census Report divorce rate 10% 2010 CDC/NSFG Report divorce rate 20% We realized the 2009 SIPP survey was more nationally representative for age at marriage and the CDC report was not (as you"ll see below). And keep in mind, the first five years of marriage show the highest rate of divorce anyway. So the CDC data was somehow taking a high divorce period in married life and making it look even worse. As with several other similar major discrepancies between surveys we found, we knew this one was worth investigating.
So why did one report show a divorce rate that was so much worse than the other?
The bottom line is that both were large surveys, but the CDC report had two major differences: (1) a very large percentage of survey takers were from a much higher-risk group (those married young), and (2) even where the lower-risk people were surveyed, the long-term average didn"t include them!
A few details: The CDC report"s "average" projected marriage survival rate at twenty years is 52 percent, which would then mean a divorce rate of 48 percent, and that looks awful. But once you look closely, you see that average only included women who married unusually young. Take a look at this data from the CDC/NSFG (we have put it in a table format that is easier to read, and we added a few words for clarity): Marriage Survival Rate of Women Ages 1544, at Each Anniversary, by Percentage, 20062010 In this table "N/A" means "not available." Source: Data from CDC/NSFG table 5 (page 16, abbreviated).
Take a look at the twenty-year-anniversary numbers. See that big "not available" under the lowest-risk group of those married at age twenty-five or older? That means that the 52 percent marriage survival average includes only the higher-risk groups of those who got married younger than twenty-five! Because this survey only surveyed women up to age forty-four, those who married over age twenty-five hadn"t yet had the opportunity to reach their twentieth anniversary.
Since the median age of marriage for women is almost twenty-seven, that supposed average is meaningless for a representative view of marriage and divorce. But that wasn"t the purpose of the study anyway; it was studying fertility, so the age forty-four cutoff makes sense for that purpose. Just not to estimate divorce.
But, really, how far off could these age-at-marriage numbers be from the actual population?
Pretty far off, actually. In fact, it is almost the opposite of the numbers in the actual population! As mentioned, the other reason we can"t take the CDC numbers at face value is because the CDC surveyed those married very young, at much higher rates than normal. We put together the table below to show just how skewed the original sample was. We used the Census Bureau 2009 SIPP data to find out what percent of people actually did get married at what ages. As you can see, in the real world (the census data), 76 percent of women get married at age twenty-five or later. But among those in the CDC study (which, again, was designed to study fertility!), 64 percent got married before the age of twenty-five!
Breakdown of Age at First Marriage by Percentage Women, Age at First Marriage Percent of Sample (CDC/NSFG) Actual Percent in Population (Census) Under 20 21 4.7 20-24 years 43 19.0 25 years & over 36 76.3 Total 100 100 Source: Data used from the 20062010 CDC/NSFG, table 5, and 2009 Census SIPP, table 6 (abbreviated). Total number rounded up from 65.6%.
Okay, so if this CDC survey had included more representative age groups, what would the divorce-rate numbers have been?
Here"s an adjustment that shows how different the results might have been if the age groups were more nationally representative.
Marriage Survival Rate of Women to Twentieth Anniversary, Weighted by Actual Population Source: Data used from 2006-2010 CDC/NSFG and 2009 Census SIPP (estimated). Total number rounded up from 65.6%.
Even with a very conservative estimate, using more characteristic numbers produces an overall marriage survival rate of 66 percent, meaning an estimated divorce rate of 34 percent.
Can you explain how you ended up with a 34 percent projected divorce rate?
Shaunti first looked at three possible ways of predicting what the twentieth anniversary survival rate would be for women twenty-five and older (the number that is "not available" on the CDC/NSFG table). All three methods of adjusting the numbers were based on other actual trends. She calculated these numbers using conservative estimates, and all delivered similar results of around 70 percent. We felt safe with that number (it could easily be a few points higher).114 For the second step, she estimated what the total national average would be if ages at marriage were more representative and with the new estimated percent of 70 percent of the over-twenty-five age group reaching their twentieth anniversary. We used the "ever divorced" out of "ever married" numbers from the 2009 SIPP (table 6 of their report) to determine the actual age-at-marriage percentages of the total married population. Then we weighted the twentieth-anniversary numbers (now including the 70 percent estimate) by what their percentage would have been if the CDC/NSFG numbers resembled the more nationally representative 2009 census numbers. (See table above.) This process gave us the average of a 66 percent marriage survival rate, thus a 34 percent divorce rate. This number resembles Barna"s 2008 survey (33 percent ever divorced) and is only a few points higher than the 2009 SIPP "ever divorced" number of 30.8 percent, so it is very much in the same ballpark. In other words ... it"s a lot more realistic than the 48 percent divorce rate that should never have been reported as a meaningful national average.
Would this type of rea.n.a.lysis actually hold up with the experts?
It depends on what you mean by "hold up." This type of informal adjustment calculation is common when trying to get a handle on how different certain numbers might be if certain factors were changed.115 This adjustment would not, as is, hold up to the kind of scrutiny professors have to go through to get a peer-reviewed paper published, proposing that this is the new right number. But as you know, our ultimate goal is not to nail down any one final number. And for our bigger-picture purposes, yes, experts have been very interested and supportive.
For example, we e-mailed the adjustments to the CDC/NSFG table to Francesca Adler-Baeder, a professor at Auburn University (and director of the National Stepfamily Resource Center), for feedback. As she put it, "I think you and Shaunti have some interesting questions and are clearly invested in getting the best estimates possible with the best data possible."116 Bottom line: always look closely at the actual data because, as the old saying goes, looks can be deceiving.
What Looks like a 50 Percent Rate.
But what about the census data that shows how few couples reach their fortieth or fiftieth anniversary?
Table 4 in the 2009 SIPP report that shows anniversaries reached for first marriages, and it looks really depressing at first glance. For women who married in the 196064 time frame, only 49.7 percent reached their fortieth wedding anniversary, meaning 50.3 percent didn"t! One might think, There"s the proof that half of marriages end in divorce. Back in the day, we used to a.s.sume that too, until we saw a completely different table from the same survey that showed how many in roughly the same age group had actually been divorced-and it was a much smaller number!
First, take a look at the depressing table on the next page (table 4). Notice the 49.7 percent in the far right column. Yikes!
Now, why is this not the divorce rate?
The first thing to remember is that one main reason for not reaching a golden-years anniversary is not just divorce but death of a spouse! Not surprisingly, since women live longer than men, 60.1 percent of men in that same age group had reached their fortieth wedding anniversary. (And today, this discrepancy in anniversaries becomes even more important in its relevance, since couples are marrying at older ages.) So how different is the picture when we look just at divorce? Another table from the same survey shows us. (See "Female" table 6, "Marital History,".) It is impossible to do a direct comparison between exactly the same people, unfortunately, since in this next table, the census folks grouped marital history by age, not by the year they got married. So we have to extrapolate a lot, but it gives us a very different ballpark estimate.
If someone got married in, say, 1964, let"s a.s.sume (to make it easy) that they were twenty-four years old at the time. That would mean they were born in 1940. So when this census survey was taken in 2009, they would have been sixty-nine years old. Now, look down the column in table 6 for women of that age, and you"ll see that only 34.5 percent have ever been divorced!117 Suddenly, the divorce rate doesn"t look like 50 percent at all.
click here to view a larger image.
click here to view a larger image.
click here to view a larger image.
What about college textbooks saying divorce remains at 50 percent?
This has come up in conversations Shaunti has had with many pastors, therapists, and others who work in the marriage arena who have been a.s.suming that textbooks are the most likely sources of good data. But after doing some research, we learned that while textbooks can indeed be helpful for a good overview of various factors, they are not necessarily the most up-to-date or thorough resource. In fact, we were taken aback at how often a 50 percent divorce rate is stated as fact in textbooks, without making clear that those were projections, not measurements-and without mentioning that the current "ever divorced" number is much, much better.
The late Norval Glenn, who was a highly revered marriage researcher and sociologist, penned a harsh critique about college textbooks on the subject of marriage and divorce. He stated in Closed Hearts, Closed Minds: The Textbook Story of Marriage, "Current textbooks convey a determinedly pessimistic view of marriage. Both by what they say and, sometimes even more importantly, by the information they omit, these books repeatedly suggest that marriage is more a problem than a solution." He also emphasized that "these books are typically riddled with glaring errors, distortions of research, omissions of important data, and misattributions of scholarship."118 However, we were delighted to find one highly respected expert who is actually a marriage champion (not a detractor) and is also currently in the process of revising his textbook. We spoke with Dr. David Olson who, besides directing his PREPARE/ENRICH organization, is an author of a college sociology textbook. Dr. Olson agreed that there was room to revise projections down from 50 percent and that projecting a "40 to 50 percent divorce rate is now more accurate."119 Of course, as you know, we think it is also important for people to be aware that the current prevalence of divorce has not reached that range.
But aren"t there a lot of divorced baby boomers who could drive the divorce rate up?
In short: yes, the baby boomers have the greatest divorce risk, but it looks like the risk is more relevant for second marriages than first ones. Unfortunately, the press isn"t really making that distinction clear in all the articles about it. "The Gray Divorce Revolution" is one newer and highly publicized study about divorce among those over fifty.120 At first glance, this study looks like a bad-news study that might refute the idea that divorce will continue to decline, because it shows an increase in baby boomer divorces between 1990 and 2010. But, once we looked closer, the numbers actually revealed better news than ever expected. Nearly 70 percent of those fifty and over are still married and in their first marriage ... very similar to what the Census 2009 SIPP data found for everyone else!
Plus, some believe the survey used here, the American Community Survey (ACS), might overestimate divorce and should not (as this study does) be compared to a completely different data set-in this case, actual vital statistics.121 Yes, the findings on divorce in remarriages were not as positive, as the "Gray Divorce" study indicated a higher number of remarriages (and some short-lived first and second marriages) that end in divorce. However, as stated in chapter 5, the most recent study on remarriages ending by divorce, by the Bureau of Labor Statistics (BLS), found the youngest baby boomers experienced redivorce at a rate of only 36 percent.122
Projections of a 50 Percent Rate.
Haven"t many experts said divorce would eventually reach 50 percent?
Yes, many experts over the years have projected the divorce rate will hit 50 percent or higher. But again, the key word here is projection. Except for some of the very highest-risk groups (such as those married as a teenager), the actual "ever divorced" divorce rate has not hit that, even after many decades of marriage.
Dr. Larry b.u.mpa.s.s, though retired, is still considered one of the preeminent experts in this field. But his projections used older data, largely during the years when divorce was still especially high. For instance, using the 1985 Current Population Survey (CPS), Dr. b.u.mpa.s.s projected a 56 percent divorce rate by forty years of marriage and a possible lifetime estimate of a two-thirds disruption rate.123 That would mean only about 33 percent of first marriages would actually survive! But in terms of what has actually happened, as you"ll remember, the Census 2009 SIPP shows that 72 percent of first marriages are still intact, and the numbers are similar even among some of the highest-risk baby boomer groups.
Dr. Robert Schoen is another authority in the demography arena, and he used different data in the past, including vital statistics and the 1995 CPS to reach an estimate of an approximately 43 percent marriage-ending rate.124 And yet, as another expert-Dr. Amato-reminded us (see chapter 2), Schoen"s data did not distinguish whether a marriage was a first marriage or not, or, more importantly, if it ended in death or divorce.125 But are the experts today, such as Dr. Amato and others, still projecting a 40 to 50 percent rate?
Yes, they are. Consistently.126 And as we said, they are the experts and time may prove them right. But there"s an important behind-the-scenes observation worth noting. Among nearly all the elite researchers we have spoken with, there is clearly a reluctance to go too far "out on a limb" (as several put it), in projecting something lower than that 40 to 50 percent rate. Not because their calculations couldn"t deliver a different number (remember, projections are based on a.s.sumptions), but because that is the consensus range of the research community. Two of these researchers cautioned Shaunti that speculating about a projected divorce rate lower than that barrier of 40 percent would risk the credibility of the book because it would fall outside the perceived norm.
However, we are not alone in wondering about these projections. Glenn Stanton at Focus on the Family is another a.n.a.lyst who has been working to bring the good news about marriage into the public awareness. He has also questioned whether the experts" projections sufficiently reflect the positive trends. As he told Shaunti, b.u.mpa.s.s has done amazing work, even though it is older now. The problem is, to some degree we all tend to a.s.sume that the things that get cited by good people are timeless and not changing. But circ.u.mstances do change, and our understanding changes. We hopefully are learning new things all the time. So we probably should be willing to take a look at it. We should absolutely be willing to ask, "What can we know and not know from the 1995 data b.u.mpa.s.s used, and where should we not have relied on that, when the Census Bureau has recent stuff?" Some of the experts have told me, "That"s the best data we have," and the question we need to ask is, "How reliable is the best data we have?" There"s a lot of work to do on clearing up misconceptions, and we need good data to do it.127 The weight of evidence suggests that it might be reasonable to adjust the a.s.sumptions for some of the more positive trends and see if there are alternative projections worth discussing. That was one reason we were so grateful for Dr. Amato"s willingness to speculate out loud that the divorce rate could in fact be lower than 40 percent, perhaps 35 percent or some other number.
Remarriage and Redivorce.
If all these leaders say second marriages have a 60 percent divorce rate and third marriages have a 73 percent chance of divorce, then does it matter if an actual study can"t be found?
It matters a lot, precisely because those same experts could not point us to an actual study. As we said in chapter 5, the Census Bureau is often cited as the source of these percentages, but they have no record of these numbers. We contacted several experts who have these numbers on their websites, and they could not point us to any actual study that substantiates these percentages.
We recognize that Dr. b.u.mpa.s.s, along with Kelly Raley, Dr. Andrew Cherlin, and Dr. Mavis Hetherington, projected higher divorce rates in remarriages in the past. For instance, Andrew Cherlin said in one of his books that he estimates 37 percent of remarriages will end in divorce within ten years of marriage, compared to 30 percent of first marriages.128 He based his estimate on data by Larry b.u.mpa.s.s, which stemmed from surveys in the 1980s. (Incidentally, Dr. Cherlin"s book is already, as of this writing, twenty-two years old.) In light of the most recent BLS study indicating only 36 percent of second marriages of high-risk baby boomers have ended and the fact that no one so far has reproduced the actual quoted census study (or any others) that show an actual 60 percent and 73 percent divorce rate for higher-order marriages, those high numbers seem highly unlikely to be accurate.
Because Dr. Cherlin was on sabbatical in 2013 while we were writing this book, he explained that he was unable to be interviewed due to the timing. We heard his name most often about this research, and he seems like the best candidate to put his hands on these numbers, if the study is out there.
I looked in the census tables and I don"t see a table for "marriages still intact" for first marriages or remarriages. Can you explain how you got your numbers?
The numbers have to be calculated from the Census 2009 SIPP. Officially, the report is called "Number, Timing, and Duration of Marriages and Divorces: 2009." Table 6 of the SIPP report (the "Marital History" table on this page) details all the marriage information and data. The table below summarizes how you do the calculations for men and women individually: *Includes those married three or more times.
Isn"t that third-marriage dissolution rate for women still pretty high?
Actually, for both women and men in third marriages, the odds are highly probable the dissolution results from death, not divorce. They are presumably significantly older at the beginning of their marital unions. Since women tend to live longer than men, one would expect a difference in all their percentages if the main reason is widowhood, which is exactly what we see. Moreover, table 6 of the SIPP report lists widowhood; currently widowed men are at 2.6 percent, compared to women at 8.9 percent currently widowed.
Your calculations show a higher number for men still in first marriage and a slightly lower one for women. How did you get an overall 72 percent first-marriage rate, then?
There"s a SIPP table (table 10 on the this page) that actually shows the overall number for couples, not just men or women.
Both spouses married only one time and still married is 71.5 percent, which the Census Bureau rounds up to 72 percent. This table also shows the breakdown by percentage of second and third or more marriages.
click here to view a larger image.
More Studies About the Impact of Faith.
You mentioned other studies about the impact of religion/faith on marriage. Can you provide some examples?
Yes, there are so many studies we could not give them all their due credit. Below we list a few that we found relevant and significant to our research.
a. The CDC/NSFG 2002 report found the importance of religion to lower divorce rates. As we already pointed out in this section, that survey is skewed on divorce-rate projections overall. However, within the survey, the difference between religious and non-religious survey takers is still helpful. Those who said faith was "very important" to them had a 28 percent lower divorce rate within fifteen years compared to those who said it was "not important."129 b. Dr. W. Bradley Wilc.o.x, mentioned several times in the book already, is well known in the academic world for his marriage research. Besides being a sociology professor at the University of Virginia and director of the National Marriage Project, Brad is also princ.i.p.al investigator (along with collaborator Jeffrey Dew) for the Survey of Marital Generosity. This survey, among other things, shows a roughly 70 percent drop in how likely the couples think they are to divorce, among couples where both partners attended church regularly, compared to those where both partners didn"t.130 c. Dr. Bradley Wright of the University of Connecticut studied the General Social Survey (GSS) for his 2010 book Christians Are Hate-Filled Hypocrites ... and Other Lies You"ve Been Told. He found that evangelicals who attended church weekly decreased their rate of divorce by 22 percent over those evangelicals who never attended. The GSS is conducted regularly by the National Opinion Research Center at the University of Chicago and goes beyond attendance and affiliation, by asking questions about religious activities, prayer, and if the survey takers "feel guided by G.o.d in daily activities."131 d. Dr. Annette Mahoney of Bowling Green State University leads the way in academic research looking at the impact of religion on families. In one key project, she published a review of 184 peer-reviewed studies and summarized that, overall, "the findings imply that higher general religiousness helps form (e.g., marital unions) and maintain (e.g., lowers divorce risk) traditional family bonds." In another study, Mahoney calculated and found a 13 percent decrease in the divorce rate in people with a specific religious identification compared to those with none. When it came to religious attendance, she calculated a 16 percent lower divorce rate with those who went to church regularly compared to those who didn"t attend on a regular basis.132 e. The FamilyLife Family Needs Survey, mentioned before, found that those who prayed or read the Bible together in any kind of a regular way, even just a few times a month, cut their marital danger signals in half.133 FamilyLife uses a sophisticated "Marital/Relational Health a.s.sessment" to a.n.a.lyze whether a relationship seems quite healthy, moderate with some concern, or high risk, and uses the traffic light a.n.a.logy of green, yellow, and red to indicate each. Among those who prayed together sometimes, only 7 percent had a lot of "red" danger signals, compared to 14 percent among those who rarely or never prayed together. They also found that 69 percent of those who prayed together sometimes were in the healthy "green" category, compared to 56 percent among those who rarely or never prayed together. There was, however, very little reduction in the various rates of past divorce based on the prayer and Bible reading categories.
Other Ways to Get Data.
How can I access some of these numbers and information for my talks, sermons, school papers, or other uses?