Background
In 1960, cohabitation was practically nonexistent in the United States, yet research suggests that cohabitation has skyrocketed since then. In 2012, the U.S. Department of Health and Human Services estimated that over two thirds of marriages begin with cohabitation. According to a Pew Research Center study, “most Americans (69%) say cohabitation is acceptable even if a couple doesn’t plan to get married.” In addition, around half of U.S. adults who were asked about the impact of cohabitation on marriage success believed couples had a better chance of marriage success if they cohabitate first. Furthermore, around four-in-ten cohabiting adults state that convenience and finances are the major contributing factors for their decision to cohabit.
There is a plethora of research discussing cohabitation trends in the United States. However, there is a lack of evidence indicating whether the increase in cohabitation is consistent across same-sex and opposite-sex partnerships. This study aims to visualize the increase in cohabitation overall, as indicated by research, and determine whether this trend is consistent across same-sex and opposite-sex partnerships.
Materials & Methods
This study first used Social Explorer to visualize and interact with choropleth maps of the American Community Survey’s 1-Year Estimates on Unmarried Partner Households overall as well as Unmarried-Partner Households by Sex of Partners (same-sex and opposite-sex partners). The cutpoints were customized to use a quantile classification method with 5 classes. This classification method was used for the purpose of distributing the data as 5 equal-sized segments of the dataset, as seen in the bottom right legends of the choropleth maps (Figures 1-3).
The American Community Survey (ACS) was inaugurated in 2005 as a demographics survey program conducted by the U.S. Census Bureau. Since the ACS only introduced the measure for Unmarried-Partner Households (UPHs) in 2006, this was the chosen starting point for data collection. 1-Year Estimates were available for every year between 2006-2021 except for 2020. To maintain cohesion in the analysis, 2019 was chosen as the cutoff for data collection.
The data was collected by state through Social Explorer, then exported individually for each year between 2006 to 2019. The data from all 14 collected datasets was cleaned, sorted, extracted, and analyzed using Microsoft Excel.
The total number of UPHs overall as well as UPHs by Sex of Partners was calculated from each dataset across all states. This data was then extracted from each year’s dataset to create a new table including these totals for their respective years. Within each of these categories, the percent change was calculated between 2006-2019 overall in addition to the percent change between each individual year across the United States as a whole. The data was then segmented by state to calculate the percent change of UPHs by Sex of Partners between 2006-2019. The state-by-state data was sorted by decreasing percent change in opposite-sex UPHs to more easily visualize any potential relationships between same-sex and opposite-sex UPHs by state. Only Wyoming was missing data for 2019, and so data collected from Wyoming was between 2006-2018. Thus, the percent change for Wyoming, as shown in Figure 6, is between this time frame.
The extracted data was imported into Datawrapper to create three visualizations: a scatterplot to visualize the increase in UPHs in the U.S. overall, an area chart to visualize the difference in trends between same-sex and opposite-sex UPHs, and a bar chart to visualize the percent changes in same-sex and opposite-sex UPHs in each state. Since the ACS collects data separately for the District of Columbia, it is visualized as a “state” in Figure 6.
Results
Figure 4 depicts the growth in cohabitation over time across the United States as a whole, as indicated by prior research. The total number of Unmarried-Partner Households (including both same-sex and opposite-sex partners) has increased by 33.9% from 2006 to 2019. The coefficient of determination is 0.959, so there is only slight variation in this linear trend as the values are very close to the regression line.
The visualization in Figure 5 demonstrates how the increasing cohabitation trends communicated across research are generalized across Unmarried-Partner Households overall. In fact, same-sex UPHs decreased by 47.1% from 2006 to 2019, compared to a 46.0% increase in opposite-sex UPHs.
After visualizing the sudden overall drop in UPHs between 2015-2016 (Figure 4), it was suspected to be caused by the legalization of same-sex marriage in 2015. However, after comparing the trends between same-sex and opposite-sex UPHs over time (Figure 5), it was clear that this observation was primarily caused by the drop in the total number of opposite-sex UPHs. In addition, the total number of same-sex UPHs decreased at a significantly higher rate between 2007-2008 (-25.1%) and 2012-2013 (-25.7%) compared to the rate between 2015-2016 (-7.63%). This demonstrates that there are other underlying factors contributing to the decrease in same-sex UPHs, and also that it is unlikely there is a correlation between the legalization of same-sex marriage and the sudden decrease in UPHs overall between 2015-2016.
Figure 6 breaks down the data further to visualize whether there is any relationship between the trends in same-sex and opposite-sex Unmarried-Partner Households state-by-state. As you can see, the general decrease in same-sex UPHs and increase in opposite-sex UPHs between 2006-2019 (as visualized in Figure 5) is consistent across all 50 states. Aside from this observation, there seems to be no obvious relationship between the percent change of same-sex UPHs and opposite-sex UPHs, as the magnitude of these changes vary greatly state-by-state. For example, North Dakota and Washington both have the same percent increase of opposite-sex UPHs between 2006-2019 (65.3%), but considerably different percent decreases for same-sex UPHs (76.5% and 47.6%, respectively). There are several possible factors contributing to the large variation in trends state-by-state, including cultural differences, changes in each state’s population size and/or demographics, underreporting, etc.
Discussion
Although the American Community Survey collects more descriptive and detailed demographic data than the U.S. Census Bureau, there are still limitations, many of which are the result of sample size. This particularly affects the availability and accuracy of the data in rural areas and states with low populations. In addition, cohabitors may not use the term “unmarried partner” when describing their relationship, which can potentially lead to couples being missed by this measure. This data is also restricted due to its measurement by sex of partners. If the ACS expands the measurement of cohabitation to include gender identity (beyond their current binary measurement of gender), this could lead to substantial changes in the cohabitation data collected in the U.S.
Change in social norms around marriage and cohabitation are likely contributors to the shift in cohabitation patterns. Other factors such as age, sex, gender, education, socio-economic status, religious affiliation, and the legal ramifications of divorce (which are not a factor when splitting from a cohabiting partner) may also contribute to the overall rise of cohabitation of unmarried-partners.
It is currently unknown as to why there is such a drastic change between the cohabitation trends in same-sex and opposite-sex partners. One study suspected that there was less stability in same-sex partner cohabitation, but found that same-sex and opposite-sex cohabitating partners have similar relationship stability levels. Thus, stability is an unlikely cause for this substantial difference in trends.
It would be interesting to compare data on Unmarried-Partner Households by Sex of Partners to Married-Couple Households by Sex of Partners to see if there is a correlation between cohabitation and marriage.