Declining Divorce in Jacksonville: Did the Culture of Freedom Initiative Make a Difference?

Executive Summary

In 2016, the Culture of Freedom Initiative (COFI), a project of The Philanthropy Roundtable launched an effort to strengthen marriage in Jacksonville, Florida. Working with Live the Life, a Florida nonprofit, and a range of churches and other nonprofits, COFI sought to improve marriage and family life in Duval County, the home county for Jacksonville. It sponsored marriage and relationship education programs, public events, and a sustained public campaign on behalf of this mission, reaching more than 11,000 adults per year from 2016-2018, for a total of approximately 50,000 adults in a three-year period. COFI and its partners also sponsored more than 28 million digital impressions, advertising its services and programs, and promoting a marriage-friendly message in Duval County during this time period. It did all of this with a budget of $1,703,415 per year from 2016-2018.

COFI coincided with a decline of 24% in the divorce rate per 1,000 persons in Duval County from 2015- 2018. However, because comparison data is only available for other counties and the nation as a whole for the 2015-2017 period, we focus much of this report on trends from 2015-2017. In that time period, the divorce rate in Duval County fell 27%, a markedly larger drop than the divorce rate decline in the U.S. over this same time, which was about 6%; it was also larger than the 10% decline in Florida. Moreover, in this time frame, the divorce decline in Duval County was larger than similar declines in divorce in counties and cities with naval bases across the U.S., as well as counties in Florida with more than 800,000 residents.

Our statistical analyses comparing divorce trends in Duval County and 29 other demographically comparable counties from 2015-2017, which controlled for a range of socioeconomic factors that might have influenced the divorce rate, indicate that the decline in divorce in Jacksonville was statistically significantly larger than the decline in similar counties. Specifically, our statistical analyses show that from 2015-2017, divorce fell about 21% more in Duval County than it did in comparable counties across the U.S..

It is important to note that family stability was already improving in Duval County prior to the Culture of Freedom Initiative. However, the 24% decline in the Duval County divorce rate from 2015- 2018 represented a substantial acceleration of downward divorce trends in the county. The divorce rate (per thousand inhabitants) had fallen from 6.2 in 2000 to 4.9 in 2010 to 4.1 in 2015. The Culture of Freedom Initiative coincided with a steeper decline. From 2000-2010, the rate fell 21% over a decade, and from 2010-2015, the rate fell 16% over five years. By contrast, from 2015-2018, the divorce rate declined by 24% over three years.

There was also a steep decline in divorce in Duval County when measured by the “refined divorce rate”—the number of divorces per 1,000 married people, as opposed to the number of divorces per 1,000 inhabitants. The refined divorce rate fell from 12.00 in 2015 to 8.41 in 2017, a 30% decline. We were able to locate administrative data for 29 other large, comparable counties. In this group, only one county—Erie County, NY—had a larger divorce decline. We also found divorce survey data from the American Community Survey (ACS), which is less reliable than the administrative data, for 61 other counties with 800,000 or more inhabitants. Only two of these counties, New York, New York (Manhattan) and Prince George’s, Maryland, had a steeper decline in their refined divorce rate during 2015-2017. This means the divorce drop for Duval County in this time frame was larger than it was in 97% of the other counties in these two different data sets.

Is the Culture of Freedom Initiative responsible for the dramatic decline in divorce in Duval County from 2015-2018? This report from the Institute for Family Studies suggests that the answer could be yes. We cannot answer the question definitively, as there may have been other factors driving divorce down in Duval County that we were not able to measure. What we can say is that the increase in family stability in Jacksonville during the years of the Culture of Freedom Initiative was larger than the increase in family stability witnessed in the vast majority of other large, comparable counties across the U.S..

The distinctive contribution of COFI in Jacksonville seems to have been its combination of microtargeted digital marketing with a broad network of religious congregations committed to strengthening marriage. By combining a digital air campaign with an in-person ground campaign, COFI was able to reach a substantial minority of the approximately 331,000 people who are married in Duval County.¹ The digital messaging COFI sponsored also conveyed the message that marriage matters to a large number of people in Jacksonville.

The leadership and extensiveness of the on-the- ground network may well account for the much more dramatic decline in Jacksonville than in two other counties—Montgomery County, OH (whose county seat is Dayton) and Maricopa County, AZ (Phoenix)—where COFI also tried to strengthen marriage and family life. In these two counties, COFI was not able to build an extensive countywide, on-the-ground network, and, hence, there is no evidence the efforts had an impact on family stability in Montgomery and Maricopa counties.

Introduction

Strong and stable families play an integral role in advancing the welfare of children, men, women, and communities. When it comes to children, as Princeton sociologist Sara McLanahan and Brookings economist Isabel Sawhill have observed, “most scholars now agree that children raised by two biological parents in a stable marriage do better than children in other family forms.”² The research tells us that children from intact, married homes are more likely to flourish in school, to graduate from college, to be gainfully employed, and to steer clear of detours, such as incarceration and teen pregnancy, that can derail their lives.³ Men and women who get and remain stably married are much more likely to enjoy higher incomes, greater assets, and better physical and mental health.⁴ And, finally, communities across the U.S. with more married parents have markedly higher levels of economic mobility and median family income, as well as significantly lower levels of child poverty and crime.⁵ Strong and stable families pay large dividends for children, adults, and communities.

There’s a growing social scientific consensus, then, that such families are ideal. But this consensus coexists with two awkward social and political facts in America:

Over the last half century, the U.S. has witnessed a dramatic retreat from marriage, and
The nation has not yet identified public policies that make a dramatic difference in renewing marriage and family stability.

Since 1970, the marriage rate (per 1,000 people) dropped by about 30%, the divorce rate (per 1,000 people) rose by more than 30% (peaking around 1980), nonmarital childbearing increased from 11% to 39%, and most importantly the share of children living with two married parents declined from 85% to 65%.⁶ The end result is that marriage is less likely to anchor the lives of men and women in this country and to provide children with the stability and security they need to thrive in today’s world.

What’s more, public initiatives to revive marriage and strengthen family stability have met with mixed success. Since 2006, the federal government has provided grants to community organizations to provide marriage and relationship education (MRE) services targeted primarily to disadvantaged, at-risk individuals and couples. Evaluations of these efforts have not proved consistently positive, however, and effect sizes, where they have been found, have not been large. In a study of programs in eight cities serving unmarried couples having a child together, only one site—Oklahoma City—showed that participants had greater family stability over the course of three years than a control group of similar couples, although the 20% increase in stability for couples at that one site was an important outcome.⁷ According to another rigorous study, programs targeted to low-income married couples usually yield only minor improvements in marriage and family outcomes.⁸ A third study across all 50 states found evidence that investments in these types of programs had “small changes in the percentage of married adults in the population and children living with two parents.”⁹ One state in particular—Oklahoma—made headway in building a statewide initiative to deliver MRE services to hundreds of thousands of citizens to strengthen marriages and couple relationships across the state; nevertheless, the state only witnessed modest increases in family stability.¹⁰ Another state, Texas, witnessed a 1.5% reduction in divorce rates after it passed a number of premarital education policies.¹¹ Moreover, a new study of the Parents and Children Together (PACT) program found improvements in warmth, declines in domestic violence, and increases in family stability for couples in the program, compared to a control group of couples not in the program.¹² But another evaluation found no evidence that community-wide MRE initiatives had a significant impact.¹³ Overall, then, publicly-funded efforts to support marriage and families have had, at best, modest success thus far.

In recent decades, however, a number of private, philanthropic efforts have launched to strengthen marriage and family life in cities, towns, and counties across the U.S.¹⁴ One such initiative is the Culture of Freedom Initiative (COFI), which was designed to strengthen marriage and family life in three counties across America—Duval County, FL, Maricopa County, AZ (county seat Phoenix), and Montgomery County, OH (county seat Dayton). COFI was founded and funded by The Philanthropy Roundtable from 2015-2018 with the intention of being spun off as an independent organization. Indeed, in late 2018, COFI went independent, has since been renamed Communio, and is now an independently operating nonprofit.

The Roundtable asked the Institute for Family Studies to study the impact of COFI.¹⁵ We focus in this report on the impact that COFI had on Duval County (2017 population: 937,934), the home of Jacksonville, Florida, both because COFI devoted the most money per capita and programmatic resources there and because divorce fell dramatically in the wake of COFI’s intervention in the area. (Note: The vast majority of Duval County is constituted by Jacksonville [2017 population: 892,062]; however, the four small independent municipalities of Atlantic Beach, Baldwin, Jacksonville Beach, and Neptune Beach also are included in the county.) Furthermore, COFI withdrew its support from Montgomery County, OH, in 2017 because it was dissatisfied with the programmatic activity of its nonprofit partners in the county, and there is no evidence that COFI had a marked impact on marriage and family trends in Maricopa County, AZ. In both counties, COFI was not able to build a large and deep network of nonprofits working to strengthen marriage and family life.

Since 2016, COFI, in partnership with Live the Life, a Florida nonprofit, and three other mobilizing organizations, worked with 93 churches and other local nonprofits to serve Duval County with a range of marriage and relationship education programs, public events, and a sustained public campaign on behalf of strong marriages and families in the Jacksonville area. In fact, COFI directly served more than 11,000 adults per year from 2016-2018, with a budget of about $1,750,000 per year for its Jacksonville area efforts. From 2016-2018, the initiative and its partners also sponsored more than 28 million digital impressions advertising its services and programs and promoting a marriage-friendly message in Duval County. And from 2015-2017, the divorce rate per 1,000 people in Duval County fell 27%. (Note:The divorce rate fell 24% from 2015-2018, as divorce ticked up slightly in 2018).

Accordingly, we seek to answer three related quantitative and qualitative sets of questions in this Institute for Family Studies report:

What does a qualitative portrait of the COFI suggest about the initiative’s character and potential impact, if any, on marriage and family life in Jacksonville—especially the divorce rate?
Did the decline in the Duval County divorce rate in the wake of the introduction of COFI exceed that of the decline in the U.S., in Florida, in other roughly comparable counties with a population of about 300,000 or more nationwide, and in counties/cities with large naval bases (like Jacksonville)?¹⁶ Are there comparable divorce declines in other large American counties from 2015-2017? In other words, does the descriptive evidence suggest that the divorce decline in Duval County was unusual?
Using data from the American Community Survey and administrative data from a number of counties and cities across the U.S., does the evidence suggest the divorce decline from 2015-2017 was statistically significantly larger in Duval than declines in other counties, after controlling for a wide range of sociodemographic factors? In other words, does Duval County’s divorce decline look particularly large compared to other counties, even after controlling for factors such as the race, education, and income of their citizens?

Section One

A Qualitative Portrait of the Culture of Freedom Initiative’s Work in Jacksonville ¹⁷

When Dr. Richard Marks, a counselor working with Live the Life in Jacksonville, asked new client Tommy Davis to rate his commitment level to his marriage, Marks got the kind of response one might expect to hear from a man on the brink of divorce.

“I’m 1% in and 99% out,” Davis said.

To which Marks replied, “I want you to give 100% of that 1% for two days, and after that, you can do whatever you want.”

So, Tommy and his wife, Sondra, agreed to attend a HOPE weekend retreat for troubled marriages, a faith-based program that Marks regularly led in conjunction with Live the Life in its efforts to strengthen family life in the Jacksonville area.

The retreat proved to be transformative. “That HOPE Weekend didn’t just save our marriage,” Sondra said. “It changed our lives.”

HOPE and other programs like it sponsored by COFI from 2016-2018 may well have changed the lives of many other Jacksonville couples. In this three-year period, Duval County experienced a 24% decline in divorce¹⁸, a decline that is considerably larger than that found in other parts of the state during this same time period (see below for more details).

This significant drop may well be thanks, in large part, to COFI’s innovative community-wide initiative designed to strengthen marital stability in Jacksonville. COFI brought together scores of ministers, marriage counselors, and nonprofit leaders in Duval County, and connected them with some key national leaders who provided start-up funding and programming expertise.

At the center of it all stood Dennis Stoica, the chairman of the Florida-based marriage enrichment non-profit, Live the Life (LTL), which served as the primary “on-the-ground” coordinator of this ambitious endeavor. Stoica had moved to Florida after leading California’s Healthy Marriage Initiative, a federally funded program that offered couples an array of skills-based marriage education classes on effective communication, conflict resolution, and other topics.

During his time with the California project, Stoica saw the limits of an exclusively skills- based approach, especially one cut off from a larger ecosystem of supportive relationships.

“It was easy for us to get permission to hold marriage education classes at a local community center like the YMCA,” Stoica said. “And the couples who came to our classes clearly benefitted from the instruction in relational skills that we offered them.” Nevertheless, after the classes were over, there was no organic network of relationships—no ongoing system of support—available should couples need additional guidance and encouragement.

From his own background as a Catholic layman, Stoica knew that churches typically offer their members an intricate web of enduring cross- generational relationships—and that interactions within this ecosystem often prove highly beneficial to married couples. Yet, Stoica also knew that in order for churches to partner with his taxpayer-funded initiative, they would have to avoid church-state entanglement issues by refraining from any instruction that presented marriage in a larger religious context.

Some California churches agreed to do just that. But this ended up neutering the faith-based programs of much of their spiritual power, according to Stoica and other observers. As J.P. De Gance, the then-executive vice president of The Philanthropy Roundtable who directed COFI, noted, “If you’re going to gut and remove faith from the equation, you’re going to have a hard time changing family behavior.”

Convinced of the need for a different approach to strengthening marital stability—a privately- funded initiative that relied heavily on church- based leadership—De Gance decided to pull in some philanthropists associated with his national organization to offer Stoica and his LTL team in Florida the kind of financial assistance they would need to jump-start an intensive community- wide effort to reduce divorce in Jacksonville.

56% of Jacksonville’s residents say they are affiliated with a church.

Why Jacksonville?

Jacksonville easily qualified as a logical place to launch an initiative of this kind. It had a comparatively high divorce rate. It had a number of religious leaders who were familiar with, and intrigued by, the long-running success of a marriage-strengthening project LTL had initiated in Tallahassee (several hours away) in 1998.¹⁹And the ringleader of that Tallahassee effort, LTL founder and president Richard Albertson, had been laying the groundwork for a major organizational expansion to Jacksonville for the previous three years.

Jacksonville was also hosting another effort to strengthen marriage and family life around this time. SMART Couples, a federally funded initiative across the state of Florida, started in 2015 and ended up offering relationship education to about 1,000 men and women in Jacksonville alone between its founding and the present day.²⁰ This initiative, then, was also seeking to improve the family climate in the city about the same time that LTL was looking to expand into Jacksonville.

In addition, Jacksonville boasts a native population that is, in many ways, more Southern (and more culturally Christian) than most other places in Florida. Indeed, Floridians like to joke that in the Sunshine State, “the farther north you go, the more Southern it gets.” And one of the best ways to illustrate this fact is by tracing the northward progression of hometowns for some of Florida’s most iconic music legends: Pitbull (Miami), Jim Morrison (Clearwater), Tom Petty (Gainesville), and Lynyrd Skynyrd (Jacksonville).

Now, the fact that Jacksonville has both high levels of divorce and high levels of religiosity may surprise those familiar with the Biblical passage where God says, “I hate divorce” (Malachi 2:16). But this cultural contradiction is actually quite common in Southern life. In fact, six of the 10 states with the highest rates of divorce are found in the South.²¹ And fans of the South’s most popular homegrown music—country—are accustomed to singing along to Lee Greenwood’s “God Bless the USA” or Carrie Underwood’s “Jesus Take The Wheel,” right after tapping their toes to Hank Williams “Your Cheatin’ Heart” or George Strait’s “All My Ex’s Live in Texas.”

In seeking to make sense of the South’s strange confluence of religiosity and divorce, some scholars have found it useful to make a distinction between religious practice and religious identification. For example, research indicates a strong correlation between weekly church attendance and marital stability, but not between conservative Protestant religious identification and marital stability.²²

In other words, research affirms the old adage that “the family that prays together stays together.” But it also suggests that a city with a sizeable number of nominal Christians and occasional churchgoers can have an unusually high divorce rate, particularly if that city has other characteristics commonly associated with high divorce. For example, sociologist Norval Glenn has shown that one of the most significant factors affecting regional differences in marital stability is “geographic rootedness”–or the lack thereof, with regions marked by high levels of mobility also facing higher divorce rates.²³ This helps to explain why high divorce rates are often found in steadily-growing, transient Sunbelt cities like Jacksonville.

In sum, then, Jacksonville is the closest thing that Florida has to a major city in the cultural Bible Belt. Jacksonville has lots of churches, big and small, and lots of churchgoers. (In fact, 56% of Jacksonville’s residents say they are affiliated with a church.²⁴) But local leaders report that Jacksonville has many people who identify as Christian, yet rarely go to church— which, for the purposes of this initiative, is quite significant. That’s because when nominal Christians and occasional churchgoers find themselves in need THE PHILANTHROPY ROUNDTABLE of some sort of marital counseling or assistance, they often have fewer apprehensions about participating in a church-based marriage program—especially one sponsored by a church with which they are familiar—than non-attenders who have no cultural or historical ties to any local congregation.

These factors, then, combined to make Jacksonville an excellent location for COFI’s innovative campaign.

COFI’s Theory of Change and Strategic Plan

Jacksonville’s Culture of Freedom Initiative can be best thought of as a privately-funded civil society project that relied heavily on church-based leadership to strengthen marital stability in Duval County, Florida. While the project’s programs taught many of the same relationship skills that Stoica had emphasized in the secular California Healthy Marriage Initiative, COFI’s theory of change was largely built around harnessing and developing the unique marriage- strengthening capabilities of local congregations.

“Faith and family tend to be mutually reinforcing,” observed J.P. De Gance, COFI’s director. “Married couples are more likely to attend church; and churchgoers are more likely to form and maintain healthy marriages.”

Similarly, churches are uniquely positioned to provide all three components that Stoica believes are needed to promote marital stability:

A vision for marriage—which churches often offer through their various teachings about the covenantal or sacramental character of marriage, and norms—like fidelity and the Golden Rule—that foster stronger marriages
Support—which churches and other religious institutions typically facilitate through the relationships built in small group ministries, Bible study classes, and various social gatherings;
Skills—which are often taught (and modeled) in both formal instruction and informal interactions.

Moreover, Stoica says that some of the benefits to marriage that come from religious participation arise indirectly from general teachings that can be applied to one’s family life. For example, a church sermon on “how to forgive others just as God has forgiven you” may not be targeted primarily to married couples–or even contain a specific reference to family life–yet can still be transformative in the life of a marriage, if the husband and wife apply these teachings to their relationship.

Still, Live the Life’s programming is built around direct instruction at face-to-face gatherings where couples at every life stage and at every point on the marital stability continuum can gain both insight and inspiration to strengthen their relationships. Some of these programs are targeted narrowly to particular types of couples (such as the newly engaged or those on the brink of divorce), but others offer general principles that are broadly applicable to wide swaths of couples.

For example, one of LTL’s signature programs, Adventures in Marriage (AIM), is a weekend retreat for married couples to get away and focus anew on their relationship. Albertson often compares this time to the 30,000-mile check-up that car owners periodically schedule for their automobiles. Yet, AIM weekends sometimes attract couples whose relationships need far more than just routine maintenance.

One such couple, Eddie and Stella Villanueva of Jacksonville, found an exercise at the AIM weekend particularly helpful in addressing a number of unsettled issues that had accumulated over their 26 years of marriage. “After we did that night of ‘taking out the trash,” Stella said, “for the first time, I felt like there was no heavy weight on me.”

While LTL brought to Jacksonville the full array of programs that it had successfully developed in Tallahassee, Albertson made a special point of signaling to prospective partners that LTL viewed its role as that of a catalyst and an aggregator, and not just that of a content provider.

“We wanted to be sure that church leaders in Jacksonville recognized that we weren’t there to simply peddle our own programs, but that we were looking to assemble a local ‘tool kit’ that would offer couples in Jacksonville a full menu of options for improving their marriage,” Albertson said. “I think this helped to build trust among our local partners who had marriage programs of their own, while also easing the fears of those who’ve seen parachurch organizations come in and try to ‘build something on their back.”

To encourage participation in the various marriage education classes and programs, COFI created a web site (jaxmarriage.org) that provided listings of all the different marriage education program options. It also relied on microtargeted outreach online to steer Jacksonville residents with a potential interest in marriage or marital assistance to its activities. And it provided seed money for Facebook ads, radio spots, and outdoor billboards that directed people to the website. All in all, COFI registered more than 28 million digital impressions (see Table 1) in the Jacksonville area from 2016-2018 in its efforts to get the word out about its local offerings and to promote a family-friendly message in the area.

Interestingly, the messaging for these ads drew heavily from some research COFI commissioned from Right Brain People, a brand strategy firm that works with many Fortune 500 corporations. According to De Gance, this research concluded that messaging around marriage needs to be tailored for three different groups: 1) romantics, who tend to have unrealistic expectations about marriage; 2) pessimists, who want lifelong marriage but tend to question its likelihood; and 3) independents, who tend to invest less in marriage and family because they regard other life pursuits as more important.

Table 1

The number of digital impressions in Duval County, by year

Source: Culture of Freedom Initiative

Using data on the different emotional needs of each of these groups, COFI then worked with communication experts to develop ads, videos, and other messaging designed to address the unique concerns and barriers to marital success of each psychographic group. Generally, this meant developing messages that helped romantics develop more realistic expectations for marriage and that helped worrywarts gain greater confidence in their ability to experience marital success. (Not surprisingly, De Gance said, these two marriage-hopeful groups were more likely to respond to messaging about marriage than the independents.)

“Lives Are Influenced by Relationships,” De Gance Said. “Those Authentic Personal Relationships Are the Bread and Butter of Civil Society.”

Whatever the case, De Gance believes COFI’s sizeable investment in this front-end research not only proved useful in Jacksonville, but that it should also prove useful in future cities that may want to replicate, or improve upon, the Jacksonville project.

COFI’s extensive use of big data and cutting- edge technology clearly distinguished it from many other faith-based initiatives, which have not operated at this level of sophistication.

“Microtargeted marketing has long existed in the commercial world. It’s existed in the political world. It’s even used in the intelligence world,” De Gance noted. “But in a lot of ways, the family and faith sectors are still living, technologically, in the 1990s. This project is bringing it forward.”

Still, De Gance and other COFI leaders did not want the technological sophistication of their plan to detract from a simple “low-tech” truth central to their strategy. “Lives are influenced by relationships,” DeGance said. “Those authentic personal relationships are the bread and butter of civil society.” Thus, even in its use of technology, COFI sought to direct married couples to programs and events that would allow them to build and deepen relational ties with others in Jacksonville who could guide, encourage, and support them.

In this important regard, then, Jacksonville’s highly- relational, high-tech project bore more than just a passing resemblance to the Tallahassee initiative that Albertson and LTL had launched in 1998—and to more than 150 other highly-relational, low-tech community marriage projects adopted around that same time through the leadership of Marriage Savers, a Maryland-based national nonprofit organization founded by Mike and Harriet McManus.

Flooding the Market

When Live the Life launched the Jacksonville project in 2016, Stoica and Albertson vowed to flood the market by: 1) increasing the supply of marriage education programs, 2) increasing the demand for such programs, and 3) conditioning the landscape for this effort.

“We wanted to ‘normalize’ the idea that people should invest in their marriage at any and every stage of their relationship” Stoica said. “So, we continually told couples that it’s ‘never too early—and never too late’ to take part in a marriage enrichment program.”

Over the course of the three-year initiative, Jacksonville program participants completed approximately 50,000 COFI-sponsored courses involving at least four hours of education related to marriage or family life. (A larger number of Jacksonville residents had some kind of contact with events and programs offered through COFI.) Most of this COFI-sponsored programming focused directly on marriage enrichment and/or helping couples in crisis, but some of this programming addressed other topics that significantly affect marital quality such as parenting, faith formation, financial management, and work-related issues.

Table 2

People participating in COFI activities in Duval County, by year

Source: Culture of Freedom Initiative

Nearly 45% of these participants took part in a program developed by Live the Life, while the rest participated in a program developed by a local church or some other local organization. For example, one of the most popular programs for couples facing money management issues was a faith-based course developed by Dave Ramsey called “Financial Peace University.”

Table 2 indicates that more than 11,000 people per year were touched by the efforts of COFI from 2016 to 2018-for a total of more than 58,000 people in this three-year period. Note, however, that some of the people counted here may not be unique participants. In other words, the number of unique members of the Jacksonville community who were reached by COFI-related activities is undoubtedly less than 58,000.

Multi-week courses like Ramsey’s were often held in church facilities, but some of COFI’s weekend or one-day programs were held in meeting rooms and conference centers not affiliated with a local church. For example, COFI partnered with Family Life. a national parachurch ministry based in Little Rock, Arkansas, to bring its highly-acclaimed “Weekend to Remember” marriage enrichment program to a Jacksonville-area conference hotel.

Four out of five COFI program participants had at least some connection to a Jacksonville church. One in 10 attended at the encouragement of a church member who invited them. And one in 10 took part even though they had no prior connection to the church or any of its members.

Almost all of those in the “no prior connection” category (and some in the other categories) learned of the marriage education classes via COFI’s digital marketing campaign, which generated more than 28 million media impressions over the course of the project. This media messaging started with a bang, with eight million impressions, to get people’s attention.

But program participation numbers in 2016 (almost 12,000) were comparatively lower than any other year. Some of this no doubt stemmed from the nature of any start-up enterprise, where one must first build awareness before expecting action. But COFI organizers determined that another factor was hindering program participation.

In year one, the campaign relied heavily on high- intensity/high-commitment offerings (multi-week classes, multi-day weekend retreats, etc.). While this represented an ideal level of instruction (at least from the organizers’ perspective), COFI officials perceived that some prospective participants—perhaps especially those not in the habit of going to church regularly-found it difficult to commit to a series of weekly meetings or to “give up an entire weekend” to participate in a marriage education program.

So, the campaign decided to offer more low- intensity/low-commitment introductory events designed to “whet the appetite” of couples who might want or need more substantial marriage education. These introductory events were often billed as “date night” programs for couples and featured everything from popular national speakers on marriage to a clean comedy act that poked fun at the everyday challenges of married life.

Interestingly, these special community-wide LTL programs fostered greater-than-usual cooperation from different denominational groups. Indeed, when the local Catholic diocese and the Jacksonville Baptist Association co-sponsored one of these special events, it represented the first time in the city’s history that these two groups had ever joined together to host such a program.

That these two groups came together to promote marital stability was no accident. From the outset, Stoica and Albertson had targeted Baptist and Catholic congregations in particular, since more religiously- affiliated Jacksonville residents identify as Baptist (35%) or Catholic (25%) than anything else. (Almost all of the rest are divided between various Protestant groups; Jews, Muslims, and Hindus collectively comprise only 2% of the Jacksonville population.)

Live the Life organizers found that they had to employ very different strategies to enlist the participation of Catholic and Baptist leaders. Some of this had to do with differences in governance. Since Catholics have a hierarchal governing structure, LTL’s Catholic outreach coordinator Lori Gramer found it easy to get cooperation from parish priests and lay leaders once she had the endorsement of the diocese. Conversely, since each Baptist church is governed at the congregation level, getting buy-in from Baptists was “a lot like herding cats,” as one associational leader put it.

Moreover, Baptists and Catholics often view ministry quite differently. For example, when it comes to church growth, Catholics tend to think in terms of growing from within. So, they put considerable emphasis on encouraging large families and on welcoming back those who were baptized in the Catholic church as children but no longer practice the faith as adults. Conversely, when it comes to church growth, Baptists tend to give considerable emphasis to growing from outside—to winning new converts to the faith from among the unchurched.

Both of these factors sometimes made it more challenging for LTL to enlist the involvement of Baptist (and other evangelical) church leaders, who often wanted to know how this marriage initiative would complement and enhance their existing outreach efforts. Indeed, of the 20 most-active congregations involved in the project, more than half were Catholic.

Nevertheless, LTL succeeded in getting enthusiastic support from some of Jacksonville’s most prominent evangelical congregations, including Chets Creek, a multi-site Baptist church with four Duval County locations, and Celebration Church, the largest evangelical “megachurch” in the greater Jacksonville area. Celebration’s involvement in the COFI project proved particularly notable for several reasons. Celebration is far more racially diverse than most Jacksonville congregations. Accordingly, it represented one of the Jacksonville project’s greatest success reaching African-American couples. (Even though about 20% of the married population in Jacksonville is African American, COFI was only able to establish relationships with two historically- black churches, Potter’s House and Hopewell Baptist Church; however, LTL relied on a number of African American mentors in its programmatic activity.)

In addition, Celebration experienced considerable growth during the three-year span of the COFI project—going from just under 7,000 attendees on a typical Sunday in 2016 to more than 10,000 attendees on a typical Sunday two years later. One of Celebration’s executive pastors, Wayne Lanier, believes this church growth is directly related to a strategic shift that took place at Celebration thanks in large part to the COFI project. Prior to the Jacksonville initiative, Celebration largely viewed itself as an “attractional” (or seeker-sensitive) congregation; but over the course of the COFI campaign, the church restructured its ministerial focus around serving marriages and families at every stage of the life cycle.

This is precisely the kind of transformation De Gance hoped would take place in Jacksonville, as the project sought to (re)build a culture around faith and family—and the meaning and purpose that these institutions foster in everyday life.

In all, approximately 50 Protestant and Catholic churches, as well as more than 40 nonprofits, took part in the initiative in some way. Several of these congregations had Spanish language programs to serve Hispanic churchgoers interested in content delivered in their native tongue, but it should be noted that the size of Jacksonville’s Hispanic population is considerably smaller than in most other major Florida cities.

LTL fostered these collaborations through initial one- on-one meetings with senior ministers, subsequent training sessions with lay leaders and mentor couples, regular participation in every-other-month associational meetings, and an annual presentation of “Marriage Champion” awards to churches and individuals that showed exemplary leadership in working to strengthen marital stability in Jacksonville.

Moreover, the Jacksonville project sought to address the most-pressing needs within each congregation–even if the pro-marriage aspects of this work might be seen more in the long run than in the near term. For example, one of COFI’s non-profit collaborators–Flourish Now—partnered with several African-American congregations to hold job fairs at urban churches as a way of addressing the underemployment of black men (since stable employment is a key factor affecting the “marriageability” of men).

Cultivating close ties with these churches reflected LTL’s emphasis on helping to “establish, develop, and grow sustainable marriage ministries in local churches,” according to Meiko Paige (who, along with her husband Christopher, helped train marriage coaches in various congregations). Moreover, cultivating close ties with church leaders facilitated the project’s shift, over time, from a heavy reliance on advertising and digital marketing to a less-costly reliance on in-house church platforms (notices in church bulletins and e-newsletters, announcements during church services, and the like) as well as word of mouth advertising from lay couples recommending marital enrichment programs to others.

Year-to-Year Changes and Possible Factors Affecting Change

While Duval County saw a 24% decline in divorce over the duration of the three-year project (measured from 2015-2018), it’s important to note that this decline did not occur in equal increments each year. Indeed, the official county statistics show substantial drops in divorce in 2016 (down 20%) and 2017 (down 9%) followed by a slight rise in 2018. But these numbers are somewhat misleading because they fail to account for the effects of Hurricane Irma, which apparently pushed some late 2017 divorces into 2018.

Hurricanes not only cause lots of closings, disruptions, and displacements, but they also frequently complicate divorce proceedings—particularly when property damage is involved (and the valuation of various possessions is affected). While the process for getting a divorce in Florida typically takes around six months (from filing to completion), major hurricanes often lengthen this process considerably—as Floridians saw with Hurricane Irma in 2017 and again with Hurricane Michael in 2018 (which hit areas several hours west of Jacksonville). So, one reason that divorce ticked up slightly in 2018 could be that there was a “Hurricane Irma effect” that led couples to postpone divorce in 2017 and then turn to it again in 2018.

There are a number of other possible explanations for this curious three-year pattern in Duval County divorce rates. To the extent that the COFI project played a role, three possible explanations would appear to merit particular consideration:

A Better Mix of ‘Thick and Thin’ Programming in 2016? In any kind of endeavor like the COFI Jacksonville project, there is an inherent tension between breadth and depth in content. Should one seek to reach the widest possible audience, even if that means offering a less-than-ideal amount of instruction? Or should one seek to offer the ideal level of instruction, even if that means reaching a less sizeable audience? Obviously, there is a sweet spot where the optimal mix of “thick and thin” programming is found. And it’s possible that the COFI organizers came closer to hitting that sweet spot in 2016 than in subsequent years (and that the corresponding year-by-year divorce rates reflect this).
The Positive Effects of Digital Messaging Apart from Boosting Program Participation? COFI viewed digital marketing and advertising as a key to driving traffic to its “jaxmarriage” web site where couples could learn about, and sign up for, various marriage education classes. But the web site also contained a wealth of other resources for interested couples, including articles, books, and links to numerous marriage ministries, counselors, and programs. So, it could be that some Jacksonville couples who never took part in a face- to-face marriage education class were still positively affected by COFI’s digital messages and by the other resources available at the web site—and took steps accordingly to improve their marriage. It may be the digital messaging in the initial stages of the initiative was perceived as more novel by Jacksonville couples in its first year, 2016, which might help to explain why divorce rates dropped more in 2016 than in subsequent years.
Did the divorce decline reverse? In the nation as a whole, divorce declined during the Great Recession and then went up a bit in 2010 until falling again in 2012. The recession may have prevented some divorces, while only delaying other divorces that were finalized from 2010-2011 as the economy improved. Likewise, one possibility in Jacksonville is that COFI prevented some divorces while only delaying other divorces. That is, the COFI initiative may have initially encouraged many couples to reconsider divorce as the best option for them; but some of these couples may have decided to go ahead with a divorce in 2018 because they were not able to resolve the problems in their marriage. This kind of pattern would also help explain the reverse J-Curve in the divorce rate in Duval County from 2015-2018.

Of course, COFI-related factors were not the only variables affecting Jacksonville divorce rates. A 2017 study by researchers at Florida State University identified a number of longer-term trends contributing to an improved climate for marital stability in Duval County, including a stronger economy (since the Great Recession) and a reduction in the number of long-term military deployments out of Jacksonville’s Naval Station Mayport (since the height of the wars in Iraq and Afghanistan).²⁵ We take up these explanations in greater detail below.

To be sure, changes in military life significantly affect the Jacksonville area. Duval County is home to four major military installations that collectively employ roughly 10,000 active duty personnel and another 12,000 civilian personnel. When family members—and the 50,000 military veterans living in or near Jacksonville—are added to the equation, the total number of people with some tie to one of Jacksonville’s four bases swells to at least 100,000 of Duval County’s nearly 1 million residents.

According to Bill Dougherty of the Navy’s southeast regional public affairs office, there have been some modest year-to-year shifts in the base population of different Jacksonville-area military installations since 2010. For example, Naval Station Mayport’s active duty population grew by around 2,000 personnel in 2016 when the base became the home port for three new ships.

Overall, however, the size of the Jacksonville military population has remained quite stable in recent years, Dougherty said, with slight upticks in one installation offset by slight decreases in others. Similarly, the length and frequency of deployments from Jacksonville bases have not changed dramatically during the 2010s.

“The operations tempo has been very steady in recent years,” Dougherty says. “We haven’t seen any major changes since the time when the wars in Iraq and Afghanistan were much more intense.”

In the wider Jacksonville community, family law experts also report no major changes in the legal climate surrounding divorce in recent years. In early 2017, the Chief Judge of Florida’s Fourth Judicial Circuit (which includes Duval County) tightened a longstanding requirement that parents going through a divorce must attend a family stabilization class in person rather than online. But family law experts and marital stability advocates do not believe this modest change had any significant effect on divorce numbers, since family stabilization classes like “Two Parents, Two Homes” (offered by Jacksonville’s Hope Haven Children’s Clinic and Family Center) are designed to help couples minimize the adverse effects of divorce on children and rarely lead to reconciled marriages.

“In all my years of working with troubled marriages, I’ve only had one couple ever referred to me by someone in the legal system,” reports Marks, the Jacksonville marriage counselor who heads the HOPE Weekend program for couples in crisis.

Given the absence of legal changes surrounding mediation and divorce, county court officials in Duval were reluctant to speculate about other possible causes for the reduction in divorce cases. For example, a spokeswoman for Family Law Judge Lance Day said, “I couldn’t even begin to guess what is responsible for the [decline in] divorce numbers.”This reluctance to speculate probably isn’t all that surprising given the premium placed on proof in the court system; but it does serve to underscore the fact that those in the legal system do not perceive Jacksonville’s divorce decline to be due to court-related reasons.

While Jacksonville saw no major changes in the legal proceedings surrounding divorce during the three- year span of the COFI project, it did see an important demographic change beginning in 2013 that may have had an effect on divorce rates in 2016-2018. Indeed, one of the reasons why divorce rate reductions may have slowed in 2017 and disappeared entirely in 2018 is because of a significant uptick in the number of Duval County marriages beginning in 2013. Some of this increase in the mid-2010s marriage rate is no doubt due to the U.S. Supreme Court’s Obergefell decision (legalizing same-sex marriage) in 2015. But it should be noted that Duval’s increase in marriage actually began two years prior to this decision and appears to be primarily attributable to the unusually large millennial generation reaching full-fledged adulthood (and prime marital age) in the mid-2010s.

As with everything else, one must be careful not to overstate the significance of this increase in marriage. And while the shift that occurred in 2013 was no doubt seen as a welcome development by LTL and other Jacksonville-area marriage advocates, it nevertheless may have had the ironic effect of making their work (of driving down divorce) harder since a number of research studies show that the marriages most at risk of breaking up are those that are still in their first seven to 10 years.

Proof of Concept

The year-to-year shifts in Jacksonville’s divorce rates are certainly interesting, but the larger story of what happened to divorce rates over the three- year duration of the Culture of Freedom Initiative is far more interesting and far more important.

The essence of the Jacksonville initiative closely resembles the Community Marriage Policy (CMP) strategy that Richard Albertson and his Tallahassee team successfully implemented in the late 1990s. In 2004, a research team led by Stan Weed found statistically-significant but modest declines in divorce rates among the more than 150 CMP initiatives that it studied. Writing in the academic journal, Family Relations, Weed and his team observed:

Our findings indicate that divorce rates appear to decline more rapidly following the signing of a [Community Marriage Policy] than would be predicted by the passage of time alone… The results reported here are important not because of their magnitude (which was modest), but because they are present. In reality, finding a significant program effect is surprising when the context of the program implementation is considered: volunteers implement the program, there is high turnover among those doing so, there is wide variation in the intensity of the program implementation, there is often a low proportion of signed congregations in the context of the larger county population, and this largely city-level intervention is only testable using the county statistics in which their results are embedded.²⁶

To be sure, COFI’s Jacksonville project represented an attempt to improve upon the CMP model. It was savvier, more sophisticated, and made better use of technology. Yet, the leaders of the Jacksonville initiative also had the wisdom to understand that the power of local religious communities lies in the vertical and horizontal relationships that are built around them. It is in and through these relationships that lives—and marriages—are transformed.

Christopher and Lacresha Hannah of Jacksonville know first-hand about marital transformation. Several years ago, their marriage faced enormous problems: infidelity; a spouse preoccupied with working and providing; a lot of hurt and bitterness and feelings of betrayal—a colossal mess in nearly every respect. But the Hannahs learned of the marital resources newly available in their city through Live the Life. And they began attending classes at their church, where they found healing and forgiveness, restoration and renewal.

“Without the church support, these classes, and a great group of friends and relatives, the success of our marriage wouldn’t be possible”

When Christopher looks back now at how far he and Lacresha have come, he marvels. “Without the church support, these classes, and a great group of friends and relatives, the success of our marriage wouldn’t be possible,” he says.

Thanks to COFI initiative, the Hannahs’ story is a familiar one in Jacksonville. And as COFI is now taking its work to other communities under a new name (Communio), many marriage and family advocates are hoping that more and more couples around the country will see their marriages transformed—just like the Hannahs, the Villanuevas, the Davises, and the many other Jacksonville couples whose lives were changed as part of this initiative.

Section Two

Is Duval’s Divorce Decline Exceptional? — A Quantitative Assessement

Divorce Is Down

Even though many Americans think divorce is rising and has been rising since the divorce revolution took off in the 1960s, the truth is that divorce has been falling since about 1980. In fact, over the past two decades, the overall U.S. divorce rate per 1,000 people fell by about 28%. (What is called the refined divorce rate—the divorce rate per 1,000 married people—is also down, by 20%, since 2000). In the country at large, then, divorce is down.

In Duval County, the divorce rate per 1,000 people fell almost as much—24%—in just three years, from 2015-2018, as the national rate fell over nearly two decades.²⁷ This drop is quite remarkable.

In this section, we will explore how rare Duval County’s divorce decline really has been. How does it compare to trends in the U.S., the state of Florida, comparable counties in Florida, comparable counties in the U.S., and counties and cities with a substantial naval presence much like Duval County? And is the drop larger in Duval County than in comparable counties after controlling for sociodemographic factors like education, race, and income? These are the questions we take up in this section.

Data and Methods

To answer these questions, we draw on data from the U.S. Centers for Disease Control, the U.S. Census, the American Community Survey (ACS, the Florida Department of Health, and administrative data collected from courts in a number of counties across the U.S. to map trends in divorce rates in Duval County, which includes Jacksonville, Florida), a number of other counties across the U.S., the state of Florida, and the U.S.. This effort also included counties and cities with other large naval bases, including San Diego, CA, Norfolk, VA, and Kitsap, WA. Using 2010 decennial census data, we then selected comparable counties in terms of population size, race-ethnicity, income, and/or marriage/divorce rates. These counties include Cameron, TX, Escambia, FL, Guilford, NC, Hillsborough, NC, Jefferson, KY, King, WA, Maricopa, AZ, and Orange, FL.²⁸ In our statistical analyses, we also include data from Jefferson, AL; Pima, AZ; New Castle, DE; Brevard, Lee, Palm Beach, Pinellas, and Polk, FL; Anne Arundel and Baltimore, MD; Wake, NC; Erie, NY; Clark, NV; Davidson, TN; and, Collin, Denton, and Travis, TX. In total, divorce trends in Duval County are compared to divorce trends in 61 other counties across the U.S. for our descriptive analyses using the ACS, and to 29 other counties for multivariate analyses of trends in divorce using administrative data (See Appendix Table 1 and Appendix Table 2 for more details).

While the comparisons between Duval County and these other counties are not perfect, this methodology allows us to see if divorce trends are similar across counties. We intentionally selected counties that match Duval’s demographic characteristics and then accounted for many other factors in our statistical models.

After charting trends in the divorce rates for the U.S., Florida, Duval County, and a number of other counties across the nation, especially in the South, we conducted statistical analyses to test whether the trends evident in Duval County differed to a statistically significant extent from the trends observed in the comparison counties and cities.²⁹ For our regression analyses, we focus on divorce trends based on administrative data, which allows us to accurately examine trends in divorce rates, since the administrative data tell us exactly how many divorces occurred in a given county or city in a given year. For the independent variables, however, in our regression analyses, we draw upon ACS data to measure age, ethnicity, education, and income of the counties and cities in our analyses.

Background: Divorce Trends in the Past Two Decades

Fueled by increases in expressive individualism, the sexual revolution, the growing economic independence of women, and the rise of no-fault divorce, divorce rates in the U.S. rose sharply in the 1960s and 1970s. But from 1980 on, the U.S. divorce rate has headed downwards. For instance, from 2000 to 2017, the divorce rate at the national level dropped by 28%, measured by the number of divorces per 1,000 people. This decline in divorce has been driven by the increasingly selective character of marriage—that is, Americans who marry today are comparatively more educated, older, affluent, and religious than the population at large—and by a growing appreciation of the importance of family stability for children.

The overall divorce rate in the state of Florida is higher than the national average, but the trend in Florida mirrors the national trend. During the same period, from 2000-2017, the Florida divorce rate fell from 5.1 divorces per 1,000 people to 3.6 divorces per 1,000 people, a drop of 29% (the refined Florida divorce rate fell 19% over this time period).

Duval County, where Jacksonville is located, has experienced a more dramatic decline in divorce rates since 2000. The county’s divorce rate dropped more than 50% between 2000-2017. And during the 2015-2017 period alone, the rate dropped by 27%, which is rare to see. (The refined Duval County divorce rate fell 30% in this period).

Between 2015-2017, the average divorce rate fell 6% in the U.S. and 10% in Florida. But in Duval County, over this time frame when COFI launched its effort to strengthen marriage, the drop was 27%. This is clearly larger than the decline of the divorce rate at the national and state levels.

To be fair, a sharp divorce decline in Duval County happened more than once in the past two decades. Between 2011-2012, there was a 19% drop in Duval’s divorce rate, and between 2007 and 2008, the drop was 14%. However, these declines were single-year drops, with the rate rising or staying the same in the successive year. In the case of the recent decline in divorce, the single-year drop between 2015-2016 was 20%, the highest in recent Duval history and it followed with a continuing rate drop in 2017. This has been unprecedented in Duval’s history in the past two decades. Of course, there was a slight uptick in the divorce rate from 2017-2018, which means the three-year decline between 2015-2018 is 24%.

Figure 1: Divorce Trends 2000-2017

Number of divorces per 1000 population

Note: Only reporting states and D.C. are included in the national divorce rate, which covers divorces in about 90% of U.S. married population. Source: U.S. and Florida data from CDC/NCHS National Vital Statistics System. Duval country data from Florida department of health.

Trends in the Past Decade: County Level Comparisons

Given the smaller population sizes in counties, the divorce rate change at a county level is often more volatile than it is at the national and state level. Therefore, it is important to compare Duval County’s divorce rate with rates in other counties that are comparable in sociodemographic factors such as population size, regional culture, and geographic location.

To better understand the significance of Duval’s decline in divorce rates, we take the following four steps: First, we look at several other counties within the state of Florida, given that these counties share the similar population sizes and, often, a similar regional culture. Second, we search for counties in other states that are similar in population size, racial-ethnic breakdown, poverty rate, and marriage/ divorce patterns³⁰ to Duval and located throughout the U.S. Third, because Jacksonville has one the largest naval bases in the U.S., we check other areas that have large naval bases (e.g., Kitsap, WA, Norfolk, VA, and San Diego, CA) in an effort to determine if divorce trends in Kitsap County, Washington, San Diego County, California, and Norfolk, Virginia parallel the divorce trends in Jacksonville, Florida. After all, there could be distinctive patterns of naval deployments or policy that would have affected the marital stability of enlisted members and officers in the navy that would impact divorce trends in Duval. Fourth and finally, we use data from the American Community Survey to determine if any other counties in the U.S. saw declines in their refined divorce rate from 2015-2017 that were equal to, or larger than, the decline of 30% in Duval County’s refined divorce rate over this period.

Duval vs. Other Florida Counties

Because the Florida Department of Health updates divorce statistics more quickly than other jurisdictions, we present divorce trends in Florida counties up to 2018 here. As shown in the next figure, the divorce rate in Duval County ticked slightly up in 2018 to 3.1 divorces per 1000 people. This brings Duval’s decline in its divorce rate since 2015 to 24%.

In contrast, Orange County (2017 population: 1.349 million), another large county in Florida, saw an increase in its divorce rate since 2015. Divorce rates in Hillsborough County (2017 population: 1.409 million) and Escambia County (2017 population: 313,512) in Florida also fell since 2015, but the rate of decline was much softer. For example, the divorce rate in Hillsborough County went from 4.0 in 2015 to 3.7 in 2018, a drop of 7.5 percent.

In additional to Orange, Hillsborough, and Escambia, we looked at all other counties in the state of Florida. Population sizes in Florida counties vary dramatically. Counties with smaller population sizes tend to have much more volatile divorce rates. For example, Franklin County experienced a 33% drop in its divorce rate between 2015-2017. However, the population of Franklin was 11,675 in 2017, only 1.3% of Duval County’s population (937,934 in 2017). Among the nine counties that are comparable in population size in Florida (with about 600,000 people or more), the drop in Duval County’s divorce rate was the largest since 2015.
Duval vs. Non-Florida counties

A look at comparison counties outside of Florida in the U.S. also suggests that Duval’s divorce rate drop was exceptional. We present 10 other counties across the country that were large and had administrative data available for divorce in recent years: Cameron County, Texas (2017 population: 423,725), Guilford County, North Carolina (2017 population: 526,953), King County, Washington (2017 population: 2.189 million), Jefferson County, Alabama (2017 population: 659,460), and Maricopa County, Arizona (2017 population: 4.307 million in 2017).

In this section, because divorce data is only available up to 2017 in these jurisdictions, we focus on trends between 2015-2017 in Figure 3a. Between 2015-2017, Guilford County in North Carolina experienced a slight increase in its divorce rate, from 2.6 to 2.9 divorces per 1000 population. The divorce rate in Maricopa County, Arizona was unchanged during the same period, and so was Davidson County’s in Tennessee. In contrast, Cameron County in Texas had a recent decline in divorce, but the drop was only 3% between 2015-2017. Meanwhile, Jefferson County in Alabama saw a slight decline in its divorce rate between 2015-2016, but this single- year change was less dramatic than what Duval County experienced during the same period.

In addition to the southern counties noted above, we looked at other regions in the country in Figure 3b. We were able to get the administrative data for King County, Washington, New Castle County, Delaware, Baltimore County, Maryland and a few other counties as a comparison. The figures below suggest Duval County’s divorce rate decline was larger than these counties between 2015-2017.

The administrative data was not available in all comparable counties we searched, and some county records we collected were incomplete. Nevertheless, Duval’s recent decline in the divorce rate looks larger in comparison to the vast majority of the larger counties in the U.S. for which we were able to locate administrative data on divorce. The one exception was Erie County, NY.
Duval vs. Counties and Cities with Naval Bases

With a population of close to 1 million, Jacksonville is the most populous city in Florida. It is also home to one of the largest Naval bases in the U.S.. The military presence in Jacksonville and Northeast Florida region includes about 75,000 active duty, reserve, and civilian personnel, which is about 14% of the region’s workforce.³¹

As noted above, the large number of military personnel in Duval County may have an impact on the area’s divorce rates, given the transient nature of military jobs and the unique stresses facing members of the armed services. To test the possibility that Duval County divorce trends were driven by a unique naval effect, we look at a few other areas in the U.S. that have a similar naval presence. We searched for divorce records in Kitsap, Washington, San Diego, California, and Norfolk, Virginia Beach, Chesapeake, and Newport News—all in Virginia.
Due to incomplete administrative records, here we only present the data for which there was divorce data: Kitsap County, San Diego County, Norfolk city, and the city of Newport News.

As shown in the figure below, divorce rates in Duval County, San Diego County, and the city of Newport News were identical in 2015. While Duval County’s rate dropped sharply in the following two years, the rate was unchanged in Newport News and dropped mildly (12%) in San Diego County. By 2017, Duval County had the lowest divorce rate among the four areas hosting, or in close proximity to, large naval bases.

We also examined trends in Norfolk’s divorce rate, which is exceptionally high when compared to rates in other areas we’ve looked at. Part of the reason is that the Norfolk Circuit Court does not have restrictions on residency for divorce filings. The records we obtained from the court include divorces filed by people who did not live in the city of Norfolk. That said, it is reasonable to assume that the vast majority of people filing in Norfolk Circuit Court for divorce live in or near Norfolk, which is the largest naval station in the U.S..

Divorces in Norfolk rose between 2015-2016 and then dropped sharply in the following year. However, the overall drop of Norfolk’s divorce rate from 2015-2017 was smaller than that of Duval County (19% vs. 27%).
Duval vs. 61 Other Counties with 800,000 or more residents

Finally, we analyzed data from the American Community Survey (ACS) to determine if any other large counties in the U.S. witnessed a similarly large decline in divorce from 2015- 2017. Specifically, we used the ACS to determine if any counties in the U.S. with 800,000 or more residents for which we were not able to get administrative data on divorce witnessed a decline in their refined divorce rate of 30% or larger in this time period which was the decline witnessed by Duval County over this period. Note, however, that because the ACS is a survey, its estimate of divorce trends is not as accurate as data derived from the courts that administer divorce in counties across the U.S.³²

In our ACS analyses of 61 counties with 800,000 or more residents for which we did not have administrative data, we find that only two counties, New York, New York and Prince George’s County, Maryland, saw divorce declines that surpassed the 30% decline witnessed in Duval County from 2015-2017.This suggests that 97% of large counties in the U.S. did not witness as large a decline as did Duval County over this period, at least according to the ACS data. Moreover, given that New York and Prince George’s County have a markedly different demographic makeup than Duval, the drivers behind such a comparatively large decline in the refined divorce rate may be different than those for Duval County. Importantly, however, it also indicates that Duval County’s divorce decline is not completely exceptional.

Figure 2: Duval County in Comparison With Other Florida Counties 2010-2018

Number of divorces per 1000 population

Source: U.S. and Florida data from CDC/NCHS National Vital Statistics System. Duval country data from Florida department of health.

Figure 3a: Duval county in comparison with non-Florida counties

Number of divorces per 1000 population

Figure 3b: Duval County in Comparison With Non-Florida Counties 2010-2017

Source: Florida department of health, Delaware Health and Social Services, North Carolina State Center for Health Statistics, Tennessee Department of Health, Maryland Department of Health.

Figure 4a: Duval County in Comparison With Other Naval Bases 2010-2017

Number of divorces per 1000 population

Source: Florida department of health, San Diego Superior Court, Newport News Circuit Court, Washington State department of Health.

Figure 4b: Duval County in Comparison With Other Naval Bases 2010-2017

Number of divorces per 1000 population

Source: Florida department of health, Norfolk Circuit Court.

Statistical Analyses

Finally, we also examined the association between these different counties and divorce in a more sophisticated, statistically rigorous way, using a combination of data from the ACS and administrative data on divorce from courts in the counties where it was available. To ensure the analysis was as rigorous as possible, we focus on 2015-2017 to ensure that our analyses are a) as removed from the effects of the Great Recession, which had its own unique effects on divorce, as possible, and b) focused explicitly on the time frame paralleling when COFI was in operation and administrative data are available. From here on, each test is based on ordinary least-squares (linear) regression, commonly referred to as the “bread and butter” of social science research, and the key statistical model on which nearly all other statistical models are based. The outcome variables in each case represent the percent drop (relative to the latest period) in the divorce rate between the years.

We examined two time periods. The first, 2015-2017, allowed us to examine long-run trends that allow for the natural variation in divorce rates that year-to- year trends are subject to, and the second, 2015-2016, allowed us to examine whether Duval’s divorce rate drop was unique in its first year of operation and when its presence was most innovative for the county.

All models control for the year the data were collected and a battery of county characteristics. These included population size, average age, median income, and the percent of the population that was female, married, white, college educated, employed, and a veteran (the latter two were dropped from administrative analyses due to statistical limitations).³³ Duval County is compared to counties that are roughly matched to Duval’s characteristics at the past (2010) Census for racial-ethnic distribution, poverty, population size, and marriage/divorce rates.

In these analyses, we make three particularly important methodological refinements. First, we only selected counties that were good statistical matches to Duval County’s demographics in the 2010 Census by using a sophisticated statistical matching analysis. Factors used in the matching analysis included population size, racial-ethnic distribution, and the poverty, marriage, and divorce rates of the counties. Thus, we did not merely select counties at random; each comparison county was selected based on a set of characteristics that made it similar to Duval in some important way, either possessing a naval base or a set of overall demographics similar to Duval’s. This is crucial to identifying whether the decline in the refined divorce rate in Duval was different from other counties’ declines, as selecting similar counties to act as statistical controls reduces the number of possible alternative explanations for the findings, meaning we have greater confidence in the results.

Second, we use administrative data obtained from the counties themselves about the number of divorces that occurred in that county in that year. The American Community Survey, despite its considerable benefits (it covers nearly all large counties in the U.S.), yields only estimates of the number of expected divorces, estimates that are subject to error. Here, we use the exact number of divorces that occurred in these counties to examine trends in divorce between 2015- 2017. Because we obtained these data through the courts themselves, we know precisely how many divorces occurred in each county each year. Third, for this analysis, we use what demographers call the “refined” divorce rate, which uses the number of married individuals in the county, instead of the “crude” divorce rate, which uses the population size as the denominator. This is important as the crude divorce rate fails to adjust for different counties’ risks of exposure to divorce.The more people who are at risk of getting divorce (i.e., married), the greater the divorce rate will be. If, on the other hand, very few people are married, the divorce rate may appear smaller, even if a very large proportion of marriages (what few there are) end in divorce.

Using statistically selected counties, administrative data, and the refined divorce rate provides the most accurate, sensitive, and sophisticated approach to comparing trends in Duval’s divorce rate to those in other counties.

Administrative Data Analyses: Duval vs. Comparison Counties Using Administrative Data

Table 3 compares Duval County’s decline in the divorce rate (again, expressed as a negative percent change, meaning divorce in Duval County went down in absolute terms). Here we compared Duval County to all comparable counties with administrative data, including Jefferson, AL; Maricopa, AZ; San Diego, CA; New Castle, DE; Brevard, Escambia, Hillsborough, Lee, Orange, Palm Beach, Pinellas, and Polk, FL; Anne Arundel and Baltimore, MD; Guilford and Wake, NC; Montgomery, PA; Davidson, TN; Cameron, Collin, Denton, and Travis, TX; and King and Kitsap, WA counties. In total, we had sufficient data for 29 counties besides Duval County for this multivariate statistical analysis (see Appendix Table 1 for a complete list of these other counties).

Between 2015-2016, the results showed that Duval’s decline in divorce rate was steeper than that observed in almost all the other counties (the one exception was Erie County, NY), and the difference was significant. The pattern held for the longer period of time between 2015 and 2017 as well. Specifically, the divorce decline was, on average, 21% greater in Duval than in these comparable counties from 2015-2017 after controlling for potential confounding factors like ethnicity, age, and income. Again, the decline in the divorce rate in Duval County was steeper than the decline in these counties, even after controlling for a range of other sociodemographic factors that could influence the divorce rate. This is particularly striking in light of the fact that we used the most sensitive tests possible, namely the exact number of divorces in each county, and calculated the refined divorce rate to examine differences between Duval and the other counties solely among the married, the only people at risk of divorce. Note, however, that ancillary analyses also indicate that the decline in divorce was significant using the crude divorce rate, as well for both the 2015-2016 and 2015-2017 periods.

The descriptive evidence and our statistical analyses, taken together, suggest that the decline in Duval County’s divorce rate was uniquely large compared to other counties. That is, descriptive statistics suggest that Duval’s decline was unique among counties comparable in population. Linear regression models reinforce this notion, especially models using the more accurate administrative data for divorce rates.

Table 3

OLS Regression Predicting Refined Divorce Rate Changes in Duval County with All Other Counties with Administrative Data on # of Divorces

Observations	90	90
	CHANGE IN DIVORCE RATE 2015-2016	CHANGE IN DIVORCE RATE 2015-2017
Duval County	0.18*** (0.02)	0.21*** (0.04)
Year	-0.00 (0.01)	0.00 (0.01)
Population Size	0.00 (0.00)	0.00 (0.00)
% Female	-4.23* (1.68)	-2.56 (2.28)
Average Age	0.01 (0.01)	0.01 (0.01)
% Married	-0.50 (0.61)	-1.05 (0.91)
% White	0.04 (0.08)	0.12 (0.23)
Median Income	0.00 (0.00)	-0.00 (0.00)
% College Degree	-0.37 (0.32)	-0.11 (0.37)
Constant	11.51 (10.93)	-0.31 (15.06)

Standard errors in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001

Limitations and Next Steps

Thus, we can say with confidence that divorce declined more in Duval County than in comparable counties across the U.S. The Culture of Freedom Initiative may have been responsible for this achievement, but the data limitations mentioned earlier and the constraints of the evaluation design preclude more definitive assertions. We cannot rule out the possibility that unmeasured factors—for instance, the role of other groups in Jacksonville, such as SMART Florida—may have also played a role in driving divorce down here. This is a problem common to many philanthropically and publicly funded efforts. What to do?

Thus, We Can Say With Confidence That Divorce Declined More in Duval County Than in Comparable Counties Across the U.S.

Understanding where this project falls within the three stages of evaluative research may help. Here is a formulation by the U.S. Department of Education’s Institute of Education Sciences and the National Science Foundation:

Efficacy research seeks “to determine whether an intervention or strategy can improve outcomes under what are sometimes called ‘ideal’ conditions. For example, these conditions may include more implementation support or more highly-trained personnel than would be expected under routine practice, or in contexts that include a more homogeneous sample of students, teachers, schools, and/or districts than is typical.

“Efficacy studies may involve the developer in the implementation of the intervention or strategy; however, the study should include reasonable safeguards for ensuring the objectivity and integrity of the study. Sometimes Efficacy studies are used to replicate previous evaluations of an intervention, but under different conditions (e.g., with a different population or using a variant of the intervention or strategy).”

Effectiveness research seeks “to estimate the impacts of an intervention or strategy when implemented under conditions of routine practice. To this end, implementation should be similar to what would occur if a study were not being conducted.

“An Effectiveness study should be carried out with no more developer involvement than what would be expected under typical implementation.”

Scale-up research seeks “to estimate the impacts of an intervention or strategy under conditions of routine practice and across a broad spectrum of populations and settings. That is, Scale-Up studies should be conducted in settings and with population groups that are sufficiently diverse to broadly generalize findings.

“An Effectiveness study should be carried out with no more developer involvement than what would be expected under typical implementation.”

Under this framework, this Jacksonville study would be considered a reasonably successful efficacy study that justifies going to the next stage, an effectiveness study. The problem is that effectiveness research is many times more expensive than efficacy research—because it involves a much more elaborate research design, multiple sites, and extensive data collection.

This is a general issue for philanthropy: Rigorous methods may identify what seems to be a promising program, but the resources are absent to go to the next stage—determining effectiveness. One option, of course, is to turn to the government for support, but that rarely succeeds and, when it does, it often comes with conditions that undermine the very nature of the program—remember the Culture of Freedom Initiative’s strong religious orientation.

Hence, a more promising avenue to explore may be to build multi-donor strategies for stage two, just as this study had support from multiple sources for its efficacy stage. Accordingly, while the approach that COFI took to strengthen marriage—partnering with a range of local non-profits and religious institutions while also relying on digital outreach— seems promising, it will take other successful efforts to achieve similar results in other locations to give us greater confidence that this strategy is replicable.

Conclusion

The impact of community initiatives like the Culture of Freedom Initiative is difficult to evaluate. It’s difficult, perhaps impossible, to select a set of well-matched communities and then randomly assign them to provide extensive marriage and relationship education services or not, which would be the ideal way to evaluate the impact of this type of initiative. Because this was not possible, we compared demographically and geographically similar communities with Duval County on trends in divorce over the last four years. From a descriptive perspective, the decline in the Duval County divorce rate from 2015-2017 was distinctive, surpassing trends in the U.S. as a whole, the state of Florida, and other large Florida counties. The 30% decline in the refined divorce rate was also a larger decline than that found in 97% of other large counties around the U.S. for which we were able to locate data.

Our multivariate analyses also show that the decline in Duval County was statistically significantly larger that the decline in comparable counties across the U.S. for which there is accurate administrative data. In other words, Duval County did in fact experience a decline in divorce that is unusually large compared to other populous counties in the U.S. Of this, we are very confident. Given the magnitude and character of the work of the Culture of Freedom Initiative and its partners, and the timing of Duval County’s divorce decline, it is highly plausible that some share of the large divorce decline in Jacksonville may be attributed to this initiative. Nevertheless, we cannot say for sure that Duval County’s unusually high decline in divorce flows directly from the Culture of Freedom Initiative, given the possibility that unmeasured factors may account for Duval County’s distinctive decline.

The findings detailed in this report are noteworthy both because of the known benefits to children of being born into married families and because of the public costs associated with marital breakdown.³⁴ ³⁵ They are also noteworthy because most public and private efforts to strengthen marriage and family life have met—at best—with only modest success.

Community-level strategies that succeed in giving more children the benefit of being born into and raised by a two-parent, married family will yield significant benefits for children, make communities better places to live, and reinvigorate the health of the American Dream in communities across the nation. Accordingly, further assessments of the ongoing work of COFI, now renamed Communio, on behalf of strong families—as well as assessments of other local initiatives to strengthen marriage and family life—are merited.

Appendix

Appendix Table 1

Refined Divorce Rate Change in 30 Counties Between 2015 and 2017: Actual (Administrative data) vs. Estimated (ACS)

State	County	Actual change (%)	Estimated change (%)
Alabama	Jefferson	-5.2	31.7
Arizona	Maricopa	-1.4	4.2
Arizona	Pima	-6.5	-30.5
California	Contra Costa	-9.3	-31.3
California	San Diego	-13.1	-13.9
Delaware	New Castle	0.4	2.7
Florida	Brevard	-19.1	-54.8
Florida	Duval	-29.9	-10.3
Florida	Escambia	-19.1	13.9
Florida	Hillsborough	-11.2	22.8
Florida	Lee	-1.3	114
Florida	Orange	-11.1	58.4
Florida	Palm Beach	-13.2	-13
Florida	Pinellas	-17.2	3.4
Florida	Polk	-12	-36.8
Florida	St Lucie	-9.8	-33.1
Maryland	Anne Arundel	-5.6	11.7
Maryland	Baltimore	-0.1	-0.9
Nevada	Clark	-7.9	-39.5
New York	Erie	-43	-40.6
North Carolina	Guilford	14.1	39.5
North Carolina	Wake	-3.6	-20.6
Pennsylvania	Montgomery	19.7	-47.4
Tennessee	Davidson	-7.3	-6.7
Texas	Cameron	0.5	24.7
Texas	Collin	-1.4	5
Texas	Denton	-12.5	-16.8
Texas	Travis	-7.3	-15.4
Washington	King	-4.8	-22.3
Washington	Kitsap	4.4	-15.4

Source: Authors’ calculation based on administrative data at the county level and American Community Survey.

Appendix Table 2

Divorce Rate Change in 61 counties Between 2015 and 2017, ACS estimates

State	County	2015 (Refined)	2017 (Refined)	Rate change (Refined)	2015 (Crude)	2017 (Crude)	Rate change (Crude)
California	Alameda	12.9	10.3	-20.3%	5.3	4.2	-20.9%
California	Fresno	14.2	18.8	32.6%	4.9	6.6	36.1%
California	Kern	14.5	16.3	12.4%	5.4	6.0	9.7%
California	Los Angeles	13.2	13.2	0.2%	4.8	4.9	0.8%
California	Orange	11.2	13.2	18.2%	4.7	5.6	18.0%
California	Riverside	16.3	13.2	-18.9%	6.5	5.3	-18.6%
California	Sacramento	18.0	16.3	-9.5%	6.8	6.3	-7.6%
California	San Bernardino	12.1	16.1	32.6%	4.4	6.1	37.4%
California	San Diego	19.4	16.7	-13.9%	7.6	6.7	-12.0%
California	San Francisco	13.0	13.1	0.6%	4.7	4.8	1.0%
California	Santa Clara	11.4	8.9	-21.9%	5.1	3.9	-23.3%
California	Ventura	11.4	12.1	6.2%	4.8	4.9	2.9%
Connecticut	Fairfield	16.0	11.3	-29.7%	6.8	4.8	-29.6%
Connecticut	Hartford	13.5	17.6	30.4%	5.4	6.8	26.8%
Connecticut	New Haven	13.0	17.2	32.6%	4.7	6.4	36.5%
Florida	Broward	27.1	21.4	-20.8%	10.1	8.0	-20.1%
Florida	Hillsborough	17.2	21.1	22.8%	6.5	8.1	25.4%
Florida	Orange	11.4	18.1	58.4%	4.1	6.7	63.3%
Florida	Palm Beach	16.5	14.4	-13.0%	6.9	5.8	-15.4%
Florida	Pinellas	16.6	17.2	3.4%	6.6	6.8	2.8%
Georgia	Gwinnett	12.1	16.5	36.6%	4.9	6.9	38.8%
Hawaii	Honolulu	13.0	13.6	4.3%	5.4	5.7	4.5%
Illinois	Cook	14.4	10.8	-25.2%	5.1	3.8	-24.8%
Illinois	Du Page	9.9	10.8	9.0%	4.5	4.9	10.4%
Indiana	Marion	23.6	17.4	-26.5%	7.0	5.6	-20.7%
Maryland	Baltimore	16.5	16.3	-0.9%	6.7	6.5	-2.0%
Maryland	Montgomery	13.1	13.0	-0.6%	5.8	5.8	-0.1%
Maryland	Prince Georges	21.7	14.8	-31.8%	7.5	5.2	-31.2%
Michigan	Macomb	15.1	18.1	19.4%	6.1	7.3	19.3%
Michigan	Oakland	12.7	13.8	8.5%	5.7	6.1	8.0%
Michigan	Wayne	21.6	17.8	-17.7%	6.7	5.8	-14.3%
Minnesota	Hennepin	11.0	17.6	60.2%	4.4	7.0	59.5%
Missouri	St Louis	14.2	16.2	13.9%	5.8	6.7	15.3%
New Jersey	Bergen	10.8	12.6	16.7%	4.9	5.8	18.2%
New Jersey	Middlesex	8.4	7.1	-14.9%	3.7	3.2	-13.2%
New York	Bronx	12.4	20.4	64.4%	3.7	5.9	60.6%
New York	Kings	13.2	13.0	-1.5%	4.5	4.5	-0.9%
New York	Nassau	6.9	7.4	7.5%	3.1	3.3	8.9%
New York	New York	20.4	9.8	-52.2%	6.3	3.2	-48.5%
New York	Queens	14.4	12.2	-15.3%	5.8	4.7	-17.9%
New York	Suffolk	12.8	11.2	-12.6%	5.4	4.8	-12.2%
North Carolina	Mecklenburg	16.8	17.4	3.5%	6.3	6.6	5.7%
North Carolina	Wake	13.7	10.8	-20.6%	5.9	4.7	-19.9%
Ohio	Cuyahoga	17.6	17.5	-0.3%	6.1	6.1	-0.3%
Ohio	Franklin	15.9	19.0	19.6%	5.5	6.8	22.1%
Ohio	Hamilton	15.2	19.3	26.9%	5.5	6.9	25.4%
Pennsylvania	Allegheny	8.2	12.4	51.5%	3.2	5.0	56.0%
Pennsylvania	Philadelphia	17.1	13.2	-22.4%	4.6	3.4	-26.8%
Tennessee	Shelby	18.2	17.0	-6.5%	6.1	5.7	-5.9%
Texas	Bexar	19.2	16.4	-14.7%	7.0	5.9	-15.0%
Texas	Collin	14.1	14.8	5.0%	6.6	6.9	4.3%
Texas	Dallas	17.2	17.3	0.5%	6.4	6.6	2.4%
Texas	El Paso	19.2	17.7	-7.7%	7.4	6.6	-11.3%
Texas	Harris	17.6	17.1	-3.1%	6.8	6.6	-3.1%
Texas	Hidalgo	8.9	9.3	5.2%	3.4	3.5	2.6%
Texas	Tarrant	21.1	17.2	-18.9%	8.2	6.9	-15.7%
Texas	Travis	21.9	18.5	-15.4%	8.1	7.0	-13.0%
Utah	Salt Lake	17.0	15.1	-10.9%	6.8	6.2	-8.4%
Washington	King	16.2	12.6	-22.3%	6.9	5.3	-22.5%
Washington	Pierce	20.5	17.0	-17.0%	8.4	7.0	-16.4%
Wisconsin	Milwaukee	12.1	16.6	37.3%	3.7	5.0	34.1%

Source: Authors’ calculation based on American Community Survey 2015 and 2017, IPUMS.

Declining Divorce in Jacksonville: Did the Culture of Freedom Initiative Make a Difference?

Executive Summary

Introduction

Section One

A Qualitative Portrait of the Culture of Freedom Initiative’s Work in Jacksonville 17William Mattox, a Florida writer affiliated with the James Madison Institute in Tallahassee, Florida, provided assistance with this section of the report.

Why Jacksonville?

COFI’s Theory of Change and Strategic Plan

Table 1

The number of digital impressions in Duval County, by year

Flooding the Market

Table 2

People participating in COFI activities in Duval County, by year

Year-to-Year Changes and Possible Factors Affecting Change

Proof of Concept

Section Two

Is Duval’s Divorce Decline Exceptional? — A Quantitative Assessement

Divorce Is Down

Data and Methods

Background: Divorce Trends in the Past Two Decades

Figure 1: Divorce Trends 2000-2017

Number of divorces per 1000 population

Trends in the Past Decade: County Level Comparisons

Figure 2: Duval County in Comparison With Other Florida Counties 2010-2018

Number of divorces per 1000 population

Figure 3a: Duval county in comparison with non-Florida counties

Number of divorces per 1000 population

Figure 3b: Duval County in Comparison With Non-Florida Counties 2010-2017

Figure 4a: Duval County in Comparison With Other Naval Bases 2010-2017

Number of divorces per 1000 population

Figure 4b: Duval County in Comparison With Other Naval Bases 2010-2017

Number of divorces per 1000 population

Statistical Analyses

Administrative Data Analyses: Duval vs. Comparison Counties Using Administrative Data

Table 3

OLS Regression Predicting Refined Divorce Rate Changes in Duval County with All Other Counties with Administrative Data on # of Divorces

Limitations and Next Steps

Conclusion

Appendix

Appendix Table 1

Refined Divorce Rate Change in 30 Counties Between 2015 and 2017: Actual (Administrative data) vs. Estimated (ACS)

Appendix Table 2

Divorce Rate Change in 61 counties Between 2015 and 2017, ACS estimates

Mentioned on this page

Declining Divorce in Jacksonville: Did the Culture of Freedom Initiative Make a Difference?

A Qualitative Portrait of the Culture of Freedom Initiative’s Work in Jacksonville ¹⁷