Religious Activities, Christian Media Consumption and Marital Quality among Protestants

Abstract

Although associations between religiosity and marital quality have been demonstrated in previous research, mechanisms still remain unclear. Three 3-step hierarchical multiple regression analyses were conducted to determine whether 10 individual, dyadic or family religious activities or uses of 7 forms of Christian media predicted positive relationship quality, negative interaction and intimate partner violence in a sample of North American Protestants. Joint spousal and family religious activities predicted higher levels of relationship quality. Individual activities, such as reading the Bible, and parent-child activities, such as praying with children and discussing Christians values with children, predicted lower levels of relationship quality. Listening to Christian talk radio and viewing Christian websites or blogs predicted lower levels of relationship quality. The authors inferred that individuals in low-quality relationships use activities such as reading the Bible, listening to Christian talk radio, and viewing Christian websites and blogs to seek information to improve relationships or promote healthy adaptation. Similarly, the authors speculated that praying with children and discussing spiritual values with them were seen as interventionary measures to protect children when parents were in low-quality relationships.

1. Introduction

The relationship between religion and marital quality is complex (Atkins and Kessel 2008; Eggebeen and Dew 2009; Vaaler et al. 2009), but generally religious behavior and values have been associated with higher marital quality (Fincham et al. 2011; Mahoney 2010; Mahoney et al. 2001). The couples that practice their religion in every aspect of life tend to improve their relationship not only with God but also their spouse (Chaney et al. 2016). However, findings sometimes are inconsistent, and questions remain concerning the mechanisms that explain this relationship. This study expands the existing research by examining the relationship between religious activities and marital quality among North American Protestants. Distinctive contributions to the literature include the exploration of how the consumption of Christian media is related to marital quality and how religious activities and Christian media consumption are related to Intimate Partner Violence (IPV).

More than two decades ago, Fincham and Linfield (1997) observed that marital quality includes both positive and negative dimensions that are distinct and should be measured separately, and Fincham and colleagues have created validated scales that measure the constructs separately (Rogge et al. 2017). A negative interaction in relationships has been associated with lower satisfaction and higher rates of relationship dissolution (see Gottman and Notarius 2000 for review). Our study explores the association between negative interaction and religious activities, including the use of Christian media.

Religious media have received increased attention (Campbell 2017; Cheong 2017). According to the Christian Bookseller’s Association (2017), sales of books, Bibles, music and other products exceed $5 billion. An increased use of social media and the Internet by churches also is reported (Newman and Benchener 2008). However, there is little or no existing research about how the consumption of these media—often even empirically-based content—is related to marital quality. For example, Doss et al. (2009) noted they were not aware of research addressing the effectiveness of relationship-themed self-help books (secular or sectarian) designed to enhance marriages or to help with specific relationship problems. Our study helps to address the gap of the knowledge related to Christian media and marital quality.

Some research is related to this topic incidentally rather than explicitly. A few studies have looked at the effects of various media on relationships, but we could find none that specifically explored the Christian media variables in our study. For example, some studies examine the relationship between television viewing and marital quality. Robinson and Martin (2008) found that happier people report more social activity and more religious involvement and newspaper reading but that more television viewing was associated with lower happiness, controlling for marital status. Various streams of research look at the relationship between listening to music and a variety of well-being measures (e.g., quality of sleep, (Urponen et al. 1988) and anxiety among dementia patients (Sung et al. 2010)), but we could find no studies that examined marital quality and listening to Christian music.

Based on their recent study, DeAngelis et al. (2018) observed that “readers approach scripture with preconceived biases conditioned by life experiences and other social constraints” (p. 12). Thus, people will turn to scripture for a variety of reasons and with varied perspectives based, in part, on their own needs. For example, they found that respondents with poor physical health were more likely to read scripture to find insight into attaining health. Their findings suggest that reading scripture, and perhaps using other Christian media, could be a help-seeking behavior that also could be relevant for individuals in distressed marriages looking for insight on improving marital quality.

As part of a larger longitudinal study, Doss et al. (2009) examined help-seeking behaviors among 213 couples that had been recruited by religious organizations to receive one of two forms of a relationship education program. Although the participants were recruited by religious organizations, the resources they sought out were not necessarily Christian in orientation. The researchers framed their study in terms of where couples sought help when they faced relationship challenges. Among the categories of help-seeking, books (cf., marital therapy and workshops/retreats) were the most commonly used (n = 49). The authors found that reading self-help books with relationship themes reflected help-seeking for individuals in distressed marriages. These findings would suggest that the consumption of Christian media, specifically books, is a help-seeking behavior for individuals in low-quality relationships.

There has been increased research regarding the effect of technology use on relationships (e.g., Roberts and Meredith 2016; Shen 2015; Padilla-Walker et al. 2012). Kerkhof and colleagues (2011) found that compulsive Internet use predicts lower levels of marital quality, though not vice versa. Several studies examine the effectiveness of formal Internet-based relationship education programs. Duncan et al. (2009) compared the effectiveness of web-based marriage and relationship education with traditional face-to-face formats. They found that both approaches produced positive changes compared to the control group and that there was no difference in the amount of change between the intervention approaches. Several studies have examined the effectiveness of a computer-based preventive intervention (ePREP) on IPV (Braithwaite and Fincham 2007; Braithwaite and Fincham 2009; Braithwaite and Fincham 2011; Braithwaite and Fincham 2014). In each case, the authors determined that an online intervention could be an effective means of reducing the incidence or severity of IPV. Although this research associates Internet use with a higher marital quality, the programs in these studies differ from our variables in that they are (1) structured, couple focused and empirically based and (2) not explicitly Christian.

IPV is a serious problem in the United States and internationally, but research regarding IPV in Protestant churches is limited. According to a 2015 data brief from the Centers for Disease Control and Prevention (Smith et al. 2018), more than 30 percent of both women and men have experienced physical violence in the context of intimate relationships. The criteria for IPV are not universal and can include sexual violence, physical violence, stalking and psychological violence. The data for our study are based on responses to this question: “Have there been instances of pushing, grabbing, shoving, hitting and/or slapping within your relationship in the last twelve months?”

There has been little research related to religion and IPV. Ellison et al. (1999) found that regular attendance at religious services was associated with lower rates of IPV, though dissimilarity in religious views and practice were associated with higher rates of IPV. More recently, Ellison et al. (2007) found that church attendance protected against IPV, especially for African-American and Hispanic men. Brinkerhoff et al. (1992) determined that religion had little or no relation to spousal violence. Todhunter and Deaton (2010) explored whether nine religiosity variables from Wave III of the National Longitudinal Study of Adolescent Health (Add Health) influenced male-perpetrated IPV. They found no predictive model for Christian male-perpetrated IPV and questioned the merit of focusing on religious or spiritual values in faith-based interventions aimed at male batterers. Our study examines some of the same variables but also looks at the relationship of Christian media consumption among Protestants.

More than 40 percent of the U.S. population identifies with Protestant churches (Pew Research Center 2015), making research on this population of continued importance. For this particular study, which measures traditional religious practices such as church attendance, Bible reading and prayer, this population is important because of the emphasis on these behaviors among Protestants. For example, compared to Catholics, Protestants are 25 percent more likely to attend services at least weekly and are 153 percent more likely to participate in prayer or scripture study groups (Pew Research Center). The use of Christian media also is more common among Protestants: Compared to Catholics, Protestants are 52 percent more likely to watch religious television, 87 percent more likely to listen to religious talk radio, 145 percent more likely to listen to Christian rock music and 80 percent more likely to share their own faith online (Pew Research Center 2014).

We suggest that the Relational Spirituality Framework (Mahoney 2010) is an appropriate lens through which to view our study. This framework “discriminates three stages of family relationships over time: (a) formation, or the creation and structure of family relationships; (b) maintenance, or processes to conserve family relationships; and (c) transformation, or fundamental changes in the structure or processes of distressed family relationships” (p. 807). Depending on the relationship and individual histories, participant experiences potentially could fall within any of these three states. For example, for engaged or newly married couples, religious behaviors and media consumption could be part of the process of building a solid foundation for a healthy, lasting marriage. For more than a fifth of our participants, those that have been married 10–15 years, these behaviors could be approached as a way to maintain or promote marital satisfaction and stability. An individual experiencing IPV may utilize religious or media resources to leave a distressed marriage or to seek positive transformation from distressed to satisfying. In this vein, the same behavior could be enacted from entirely different motivations by different people in different circumstances. For one participant, discussing the Bible with a spouse could be an eagerly anticipated activity that reinforces marital satisfaction, whereas for someone else the discussion could be an ongoing platform for conflict or even violence. Similarly, one individual could use social media to express fond memories or delightful experiences from a satisfying marriage, while others might seek out advice on Christian websites about how to extricate themselves from a dangerous relationship.

Based on previous research and considering the Relational Spirituality Framework, we anticipate the following general findings from the analysis of our data collected from North American Protestants:

Hypothesis 1.

Higher levels of participation in religious-related activities will predict a higher level of positive marital quality, a lower level of negative interaction and a lower level of IPV.

Hypothesis 2.

Higher levels of Christian media consumption will predict a lower level of positive marital quality, higher levels of negative interaction and higher levels of IPV.

2. Materials and Methods

This study utilizes secondary data from the Family Needs Survey administered during 2012 and 2013 by FamilyLife as a service to Protestant churches to assess the health of families in individual congregations (Gritzon 2013). Using this relatively large but nonrandom sample (N = 6613), we examine the relationship of individual and dyadic spiritual activities to marital quality. To protect the confidentiality of the local churches, data were provided to the researchers without identifying information. Thus, the location, size and specific denomination of the congregations are not known. FamilyLife reported to the researchers that the data include diversity in denominations, location, geography and size and represent a variety of Protestant denominations from throughout North America. Use of the data was secured through a license with FamilyLife. Since these were existing data, the authors had no input into the formulation of items included in the survey. However, the dataset includes multiple variables of possible interest to scholars of religion and families, including some widely used and psychometrically validated scales. Institutional Review Board approval was granted by the Office for Research Compliance at Mississippi State University (IRB-16-443).

The majority of the participants were female (n = 4001; 60.5%), and the modal age category was 50–59 years (n = 1448; 21.9%). The number of years married was more widely dispersed, with the most common ranges being 10–19 years (n = 1213; 22.2%) and more than 40 years (n = 1321; 24.2%). The most common relationship status was in a first marriage (n = 4274; 64.6%), with other statuses including engaged (n = 88; 1.3%), cohabiting (n = 51; 0.8%), widowed (n = 372; 5.6%), separated (n = 32; 0.5%), divorced (n = 360; 5.4%) and remarried (n = 988; 14.9%). Some commonly used demographic variables such as ethnicity/race, education and income were not included in the data.

See

Table 1. A summary of the interval variables.

Variable	N	M	SD	Range	SEM
Positive marital quality	5407	33.78	5.65	32	0.077
Negative interaction	5359	8.30	2.60	10	0.035
Spiritual development	6465	2.95	0.02	8	0.018
Read the Bible	6408	3.81	1.40	5	0.017
Concentrated prayer	6370	4.56	1.43	5	0.018
Family devotion	6163	2.65	1.30	5	0.016
Pray with your spouse	6086	2.56	1.32	5	0.017
Read or discuss the Bible with spouse	6068	2.78	1.22	5	0.016
Talk about spiritual values with child	5904	3.26	1.45	5	0.019
Pray for children	6158	4.69	1.69	5	0.021
Pray with children	5836	2.98	1.70	5	0.022
Share Christ with others	6191	3.17	1.07	5	0.014
Watch Christian television/video	6164	2.66	1.02	5	0.013
Listen to Christian talk radio	6190	2.80	1.34	5	0.017
Listen to Christian music	6277	3.58	1.50	5	0.019
Read Christian book or magazine	6240	3.11	1.17	5	0.015
Use online social networking tools	6217	3.61	1.88	5	0.024
Listen to sermon/teaching audio/podcast	6190	2.73	1.15	5	0.015
View Christian web sites/blogs	6005	2.49	1.05	5	0.014

and

Table 2. A summary of the categorical and ordinal variables.

Variable	Categories	n	%
Intimate partner violence in last 12 months
	Never	5337	a 95.8
	Once	139	2.5
	Twice	54	1.0
	3–5 times	30	0.5
	6–10 times	4	0.1
	More than 10 times	6	0.1
	Total	5570	100.0
Gender
	Male	2612	39.5
	Female	4001	60.5
	Total	6613	100.0
Age
	19 and under	79	1.2
	20–29	500	7.6
	30–39	934	14.1
	40–49	1300	19.7
	50–59	1448	21.9
	60–69	1310	19.8
	70 or over	1042	15.8
	Total	6613	100.0
Years married
	0–4	378	6.9
	5–9	543	9.9
	10–19	1213	22.2
	20–29	1076	16.3
	30–39	937	17.1
	40+	1321	24.2
	Total	5468	100.0
How long have you been a Christian?
	I am not a Christian	16	0.2
	0–5 years	82	1.3
	6–10 years	97	1.5
	10+ years	6232	97.0
	Total	6427	100.0
How often do you attend worship services?
	Less than once a month	254	3.9
	Once a month	182	2.8
	Twice a month	496	7.6
	Three times a month	1318	20.2
	Four or more times a month	4261	65.4
	Total	6511	100.0

^a The percentages are calculated based on the number of valid responses.

for summaries of all the variables. A correlation table including all variables was too cumbersome for inclusion here but is available from the corresponding author.

2.1. Criterion Variables

Three criterion (dependent) variables were selected for this study: positive marital quality, negative interaction and IPV. To test the validity of our decision to use negative interaction as a discrete construct separate from positive marital quality, we conducted a principal component analysis entering all 14 items making up the three criterion variables. Three components emerged after six iterations, based on an Eigenvalue greater than 1. The items that loaded on the second component were the five items in our Negative Interaction Scale and the single item measuring IPV, with the IPV item having a visibly smaller loading. The items with the highest loadings on the first and third components were the eight items in our Positive Marital Quality Scale. The analysis was consistent with our use of the Negative Interaction Scale as a separate criterion variable.

2.1.1. Positive Marital Quality Scale

The variable Positive Marital Quality was measured by combining and modifying two existing scales: Marital Satisfaction Scale (Schumm et al. 1985) and Positive Bonding Scale (Allen et al. 2010). The modified Marital Satisfaction Scale (α = 0.898) includes the following questions, each measured using a 5-point Likert-type scale ranging from 1 = very dissatisfied to 5 = very satisfied:

• How satisfied are you with the way you connect with your spouse/fiancé/significant other?
• How satisfied are you with how your relationship functions day-to-day?
• How satisfied are you with your spouse/fiancé/significant other’s contributions to your relationship?
• How satisfied are you with your own contributions to your relationship?

The Positive Bonding Scale included the following items, each measured using a 5-point Likert-type scale with possible responses ranging from 1 = strongly disagree to 5 = strongly agree:

• We regularly have great conversations where we just talk as good friends.
• I want this relationship to stay strong no matter what difficult times we may encounter.
• I believe we can handle whatever conflicts that may arise in the future.
• My relationship with my spouse/fiancé/significant other is more important to me than almost anything else in my life.

2.1.2. Negative Interaction Scale

The second criterion variable was adapted from the Danger Signs Scale (Stanley et al. 2002). The scale (α = 0.832) includes the five items below, which were measured with a 3-point Likert-type scale which had the following possible responses: 1 = almost never, 2 = once in a while and 3 = frequently.

• Little arguments escalate into ugly fights with accusations, criticisms, name calling and/or bringing up past hurts.
• My partner criticizes or belittles my opinions, feelings or desires.
• My spouse/fiancé/significant other seems to view my words or actions more negatively than I intended for them to be.
• When we have a problem to solve, how often does it feel like we are on opposite teams?
• When we argue, one of us withdraws, that is, doesn’t want to talk about it anymore or leaves the scene.

2.1.3. Intimate Partner Violence

The third criterion variable is a single-item measure of IPV taken from the Danger Signs Scale that asks, “Have there been any instances of pushing, grabbing, shoving, hitting and/or slapping within your relationship in the last twelve months?” Possible responses were 1 = never, 2 = once, 3 = twice, 4 = 3–5 times, 5 = 6–10 times and 6 = more than 10 times.

2.2. Predictor Variables

Three categories of predictor (independent) variables were included in the analysis: control variables, religious activities and Christian media consumption.

2.2.1. Control Variables

Based on previous research that associates them with marital quality, the following were included as control variables:

• gender

  ○

  1 = male

  ○

  2 = female

• age

  ○

  1 = 19 and under

  ○

  2 = 20–29

  ○

  3 = 30–39

  ○

  4 = 40–49

  ○

  5 = 50–59

  ○

  6 = 60–69

  ○

  7 = 70 or over

• years married

  ○

  1 = 0–4 years

  ○

  2 = 5–9 years

  ○

  3 = 10–19 years

  ○

  4 = 20–29 years

  ○

  5 = 30–39 years

  ○

  6 = 40+ years

• “How long have you been a Christian?”

  ○

  System missing = Don’t know

  ○

  0 = I’m not a Christian

  ○

  1 = 0–5 years

  ○

  2 = 6–10 years

  ○

  3 = 10+ years

• Spiritual development

  ○

  0 = No spiritual development

  ○

  1 = Very low

  ○

  2 = Low

  ○

  3 = Low average

  ○

  4 = Medium

  ○

  5 = Slightly above average

  ○

  6 = Above average

  ○

  7 = High

  ○

  8 = Very high

  ○

  9 = Maximum spiritual development

2.2.2. Religious Activities

• Worship attendance

  ○

  1 = Less than once a month

  ○

  2 = Once a month

  ○

  3 = Twice a month

  ○

  4 = Three times a month

  ○

  5 = Four or more times a month

The following religious activity variables all are measured with the same potential responses, which are listed after the list of variables:

• Bible reading
• Prayer
• Having a family devotion
• Praying with a spouse
• Reading/discussing the Bible with a spouse
• Talking about spiritual values with the children
• Praying for the children
• Praying with the children
• Sharing Christ with others

  ○

  1 = does not apply

  ○

  2 = rarely/never

  ○

  3 = occasionally

  ○

  4 = several times a month

  ○

  5 = several times a week

  ○

  6 = almost every day

2.2.3. Christian Media Consumption

All Christian media variables used the same possible responses, which are listed following the variables below.

• Watch Christian television/video
• Listen to Christian talk radio
• Listen to Christian music
• Read a Christian book and/or magazine
• Use online social networking tools (i.e., Facebook, Twitter, YouTube, etc.)
• Listen to sermon/teaching audio/podcast
• View Christian websites/blogs

  ○

  1 = does not apply

  ○

  2 = rarely/never

  ○

  3 = occasionally

  ○

  4 = several times a month

  ○

  5 = several times a week

  ○

  6 = almost every day

2.3. Analytical Strategy

Three 3-step hierarchical multiple regression analyses were conducted to determine whether religious activities (including the use of Christian media) demonstrated surplus explanatory power relative to more general types of predictors in relation to positive marital quality, negative interaction and IPV. For all three analyses, the same predictor variables were entered in the same order. Missing cases were excluded listwise.

3. Results

Three 3-step hierarchical multiple regression analyses were conducted to determine whether religious activities predicted positive marital quality, negative interaction and IPV. The significance level was set at 0.05. See

Table 3. A summary of the hierarchical regression analysis for variables predicting a positive marital quality (N = 6613).

Variable	Step 1			Step 2			Step 3
	B	SEB	β	B	SEB	β	B	SEB	β
Gender	−0.554	0.180	−0.049 **	−0.374	0.180	−0.033 *	−0.321	0.185	−0.028
Age	−0.140	0.109	−0.034	−0.455	0.113	−0.111 ***	−0.478	0.116	−0.117 ***
Years married	0.119	0.099	0.032	0.194	0.097	−0.052 *	0.180	0.096	0.048 *
How long a Christian	−0.301	0.324	−0.015	−0.159	0.314	−0.008	−0.113	0.312	−0.006
Spiritual development	0.719	0.064	0.182 ***	0.523	0.072	0.132 ***	0.526	0.071	0.133 ***
Attend services				−0.113	0.098	−0.020	−0.043	0.099	−0.007
Bible reading				−0.447	0.082	−0.110 ***	−0.311	0.086	−0.076 ***
Prayer				0.085	0.074	0.021	0.102	0.074	0.026
Family devotion				0.120	0.084	0.026	0.133	0.084	0.029
Pray with spouse				0.472	0.086	0.105 ***	0.488	0.085	0.108 ***
Read/discuss Bible with spouse				1.108	0.105	0.218 ***	1.157	0.105	0.228 ***
Talk to children about spiritual values				−0.397	0.090	−0.098 ***	−0.359	0.090	−0.088 **
Pray for children				0.287	0.077	0.076 ***	0.254	0.077	0.067 **
Pray with children				−0.162	0.075	−0.049 **	−0.135	0.074	−0.041
Share Christ							0.101	0.094	0.019
Christian TV/video							−0.027	0.105	−0.005
Christian talk radio							−0.283	0.080	−0.068 ***
Christian music							−0.080	0.075	−0.022
Christian book or magazine							−0.051	0.096	−0.011
Social media							0.061	0.054	0.018
Sermon/teaching audio							−0.089	0.083	−0.018
Christian websites							−0.262	0.098	−0.049 **
R2change		0.037 ***			0.068 ***			0.012 ***
F	26.691 ***				30.468 ***			23.304 ***

* p < 0.05. ** p < 0.01. *** p < 0.001.

,

Table 4. A summary of the hierarchical regression analysis for variables predicting a negative interaction (N = 6613).

Variable	Step 1			Step 2			Step 3
	B	SEB	β	B	SEB	β	B	SEB	β
Gender	0.031	0.084	0.006 **	−0.066	0.086	−0.013	−0.077	0.089	−0.015
Age	−0.018	0.051	−0.010	−0.089	0.054	−0.047	0.093	0.056	−0.049 **
Years married	−0.027	0.046	−0.016	−0.052	0.046	−0.030	−0.049	0.046	−0.028
How long a Christian	−0.152	0.152	−0.016	−0.212	0.150	−0.022	−0.220	0.150	−0.023
Spiritual development	−0.240	0.030	−0.131 ***	−0.170	0.034	−0.093 ***	−0.172	0.034	−0.094 ***
Attend services				−0.031	0.047	−0.012	−0.047	0.047	−0.018
Bible reading				0.137	0.039	−0.072 ***	0.079	0.041	−0.041
Prayer				−0.024	0.035	−0.013	−0.031	0.035	−0.017
Family devotion				−0.106	0.040	−0.050	−0.112	0.040	−0.053
Pray with spouse				−0.163	0.041	−0.078 ***	−0.170	0.041	−0.081 **
Read/discuss Bible with spouse				−0.328	0.050	−0.139 ***	−0.336	0.050	−0.143 ***
Talk to children about spiritual values				0.090	0.043	−0.048 ***	0.077	0.043	0.041
Pray for children				−0.046	0.037	−0.026	−0.034	0.037	−0.019 **
Pray with children				0.099	0.036	0.064 **	0.090	0.035	0.069 *
Share Christ				0.023	0.044	0.009	0.003	0.045	0.001
Christian TV/video							−0.032	0.050	−0.012
Christian talk radio							0.181	0.038	0.094 ***
Christian music							−0.033	0.036	−0.019
Christian book or magazine							0.060	0.046	0.026
Social media							−0.021	0.026	−0.015
Sermon/teaching audio							−0.037	0.040	−0.016
Christian websites							0.125	0.047	−0.050 **
R2change		0.020 ***			0.035 ***			0.010 ***
F	16.115 ***				15.217 ***			12.409 ***

* p < 0.05. ** p < 0.01. *** p < 0.001.

and

Table 5. A summary of the hierarchical regression analysis for the variables predicting intimate partner violence (IPV) (N = 6613).

Variable	Step 1			Step 2			Step 3
	B	SEB	β	B	SEB	β	B	SEB	β
Gender	−0.004	0.012	−0.006	−0.008	0.013	−0.011	−0.005	0.013	−0.006
Age	0.004	0.007	−0.016	−0.010	0.008	−0.036	0.008	0.008	0.030
Years married	−0.015	0.007	−0.060 *	−0.015	0.007	−0.061 *	−0.016	0.007	−0.063
How long a Christian	−0.030	0.022	−0.022	−0.031	0.022	−0.023	−0.030	0.022	−0.021
Spiritual development	−0.006	0.004	−0.023	−0.010	0.005	−0.036	−0.009	0.005	−0.032
Attend services				−0.001	0.007	−0.003	−0.002	0.007	−0.005
Bible reading				0.010	0.006	−0.038	0.009	0.006	0.033
Prayer				—	0.005	0.000	0.000	0.005	−0.001
Family devotion				−0.006	0.006	−0.020	−0.006	0.006	−0.021
Pray with spouse				0.001	0.006	0.003	0.000	0.006	0.001
Read/discuss Bible with spouse				−0.010	0.007	−0.029	−0.010	0.007	−0.030
Talk to children about spiritual values				0.000	0.006	−0.002	−0.001	0.006	−0.005
Pray for children				−0.006	0.005	−0.022	−0.005	0.005	−0.020
Pray with children				0.012	0.005	0.053 *	0.011	0.005	0.049 *
Share Christ				0.010	0.006	0.027	0.008	0.007	0.049
Christian TV/video							0.010	0.007	0.025
Christian talk radio							0.018	0.006	0.062 **
Christian music							−0.002	0.005	−0.007
Christian book or magazine							−0.007	0.007	−0.021
Social media							−0.005	0.004	−0.023
Sermon/teaching audio							−0.010	0.006	−0.029
Christian websites							0.005	0.007	0.014
R2change		0.004 **			0.003 ***			0.005 ***
F	3.305 ***				2.027 ***			2.268 ***

* p < 0.05. ** p < 0.01. *** p < 0.001.

for the regression summaries.

3.1. Positive Marital Quality

In step 1 of the regression analysis where Positive Marital Quality was the criterion variable, the control variables added were gender, age, years married, length of time as a Christian and spiritual development. Gender (β = −0.049; p = 0.002) and spiritual development (β = 0.182; t = 11.158; p < 0.001) were significant predictors of positive marital quality. Step 1 was statistically significant (∆R² = 0.04, F(5,3899) = 29.69, p < 0.001) and explained 3.5% of the variance. Tests to see if the data met the assumption of collinearity indicated that multicollinearity was not a concern (Gender: Tolerance = 0.97, VIF = 1.03; Age: Tolerance = 0.35, VIF = 2.86; years married: Tolerance = 0.35, VIF = 2.83; years a Christian: Tolerance = 0.98, VIF = 1.03; and spiritual development: Tolerance = 0.93, VIF = 1.08).

In step 2, individual and family-centered religious activities were added to the model. Four of the control variables were significant predictors: gender (p = 0.038), age (p < 0.001), years married (p = 0.038) and spiritual development (p < 0.001). Among the variables measuring religious activities, reading the Bible (β = −0.110; t = −5.474; p < 0.001), praying with a spouse (β = 0.105; t = 5.499; p < 0.001), reading/discussing the Bible with a spouse (β = 0.218; t = 10.560; p < 0.001), talking about spiritual values with the children (β = −0.098; t = −4.393; p < 0.001), praying for the children (β = 0.076; t = 3.713; p < 0.001) and praying with the children (β = −0.162; t = −2.166; p = 0.030) were significant predictors. Step 2 was statistically significant (∆R² = 0.07, F(10,3904) = 30.47, p < 0.001) and explained 6.8% of the variance. Tests to see if the data met the assumption of collinearity indicated that multicollinearity was not a concern (attend worship: Tolerance = 0.79, VIF = 1.26; read the Bible: Tolerance = 0.57, VIF = 1.74; individual prayer: Tolerance = 0.66, VIF = 1.53; family devotion: Tolerance = 0.69, VIF = 1.47; pray with a partner: Tolerance = 0.64, VIF = 1.58; read/discuss the Bible with a partner: Tolerance = 0.54, VIF = 1.86; discuss spiritual values with the children: Tolerance = 0.46, VIF = 2.15; pray for the children: Tolerance = 0.56, VIF = 1.80; pray with the children: Tolerance = 0.45, VIF = 2.24; and share Christ: Tolerance = 0.76, VIF = 1.31).

Media-related religious activities were added in step 3. Among the control variables, age (p = 0.038) and spiritual development (p < 0.001) were significant predictors of positive marital quality. Among the religious activities, reading the Bible (p < 0.001), praying with a spouse (p < 0.001), reading/discussing the Bible with a spouse (p < 0.001), talking about spiritual values with the children (p < 0.001) and praying for the children (p = 0.001) were significant predictors. Among the media variables, only listening to Christian talk radio (β = −0.068; t = −3.522; p < 0.001) and viewing Christian websites/blogs (β = −0.049; t = −2.680; p = 0.007) were significant predictors of positive marital quality. Step 3 was statistically significant (∆R² = 0.01, F(22,3904) = 23.30, p < 0.001) and explained 1.2% of the variance over and above other variables. The final model explained 11.7% of the variance. Tests to see if the data met the assumption of collinearity indicated that multicollinearity was not a concern (Christian TV: Tolerance = 0.69, VIF = 1.44; Christian talk radio: Tolerance = 0.62, VIF = 1.62; Christian music: Tolerance = 0.55, VIF = 1.80; Christian book or magazine: Tolerance = 0.59, VIF = 1.70; social networking: Tolerance = 0.71, VIF = 1.42; sermon recordings: Tolerance = 0.81, VIF = 1.24; and Christian websites: Tolerance = 0.68, VIF = 1.48).

Consistent with previous research, joint spousal religious activities were positive predictors of positive marital quality, as was praying for the children. However, individual activities (Bible reading) and parent-child dyadic activities (talking about spiritual values and prayer) were associated negatively with marital satisfaction. The only two media-related activities that were significant (Christian talk radio and Christian websites/blogs) also had a negative relationship with positive marital quality.

3.2. Negative Interaction

In step 1 of the regression analysis with negative interaction as the criterion variable, only spiritual development among the control variables was a statistically significant predictor (β = −0.093; t = −4.988; p < 0.001). Step 1 was statistically significant (∆R² = 0.02, F(5,3897) = 16.12, p < 0.001) and explained 2.0% of the variance. Tests to see if the data met the assumption of collinearity indicated that multicollinearity was not a concern (Gender: Tolerance = 0.98, VIF = 1.03; Age: Tolerance = 0.35, VIF = 2.89; years married: Tolerance = 0.35, VIF = 2.87; years a Christian: Tolerance = 0.98, VIF = 1.03; and spiritual development: Tolerance = 0.93, VIF = 1.08).

In step 2, individual and family-centered religious activities were added to the model. Spiritual development (p < 0.001) was the only control variable that was a statistically significant predictor. Among the variables measuring religious activities, reading the Bible (β = 0.072; t = 3.514; p < 0.001), having a family devotion (β = −0.050; t = −2.642; p = 0.008), praying with a spouse (β = −0.078; t = −3.979; p < 0.001), reading/discussing the Bible with a spouse (β = −0.139; t = −6.582; p < 0.001), talking about spiritual values with the children (β = 0.048; t = 2.099; p = 0.036) and praying with the children (p = 0.006) were significant predictors. Step 2 was statistically significant (∆R² = 0.04, F(15,3887) = 15.217, p < 0.001) and explained 3.5% of the variance. Tests to see if the data met the assumption of collinearity indicated that multicollinearity was not a concern (attend worship; Tolerance = 0.79, VIF = 1.27; read the Bible: Tolerance = 0.57, VIF = 1.74; individual prayer: Tolerance = 0.65, VIF = 1.53; family devotion: Tolerance = 0.68, VIF = 1.47; pray with a partner: Tolerance = 0.64, VIF = 1.57; read/discuss the Bible with a partner: Tolerance = 0.54, VIF = 1.84; discuss spiritual values with the children: Tolerance = 0.47, VIF = 2.14; pray for the children: Tolerance = 0.55, VIF = 1.80; pray with the children: Tolerance = 0.45, VIF = 2.22; and share Christ: Tolerance = 0.77, VIF = 1.31).

Media-related religious activities were added in step 3. Spiritual development (p < 0.001) was the only control variable that predicted a negative interaction in this step. Among religious activities, predictors included family devotions (p = 0.005), praying with a spouse (p < 0.001), reading/discussing the Bible with a spouse (p < 0.001) and praying with the children (p = 0.011). The only Christian media variables that were statistically significant predictors of negative interaction were listening to Christian talk radio (β = 0.094; t = 4.723; p < 0.001) and viewing Christian websites/blogs (β = 0.050; t = 2.670; p = 0.008). Step 3 was statistically significant (∆R² = 0.01, F(7,3880) = 12.41, p < 0.001) and explained 1.0% of the variance over and above other variables. Tests to see if the data met the assumption of collinearity indicated that multicollinearity was not a concern (Christian TV: Tolerance = 0.69, VIF = 1.45; Christian talk radio: Tolerance = 0.61, VIF = 1.63; Christian music: Tolerance = 0.55, VIF = 1.81; Christian book or magazine: Tolerance = 0.59, VIF = 1.70; social networking: Tolerance = 0.70, VIF = 1.43; sermon recordings: Tolerance = 0.80, VIF = 1.24; and Christian websites: Tolerance = 0.68, VIF = 1.48).

Confirming Hypothesis 2, joint spousal or family religious activities were negative predictors of negative interaction. However, disconfirming our hypothesis, praying with the children was a positive predictor of negative interaction, and consistent with our hypothesis, the only two media-related activities that were significant predictors of negative interaction (Christian talk radio and Christian websites/blogs) both had a positive relationship with negative interaction.

3.3. Intimate Partner Violence

In step 1 of the regression analysis where IPV was the criterion variable, years married (β = −0.060; t = −2.265; p = 0.024) was the only control variable that predicted IPV. Step 1 was statistically significant (∆R² = 0.004, F(5,3969) = 3.31, p = 0.006) and explained 0.3% of the variance. Tests to see if the data met the assumption of collinearity indicated that multicollinearity was not a concern (Gender: Tolerance = 0.98, VIF = 1.03; Age: Tolerance = 0.35, VIF = 2.83; years married: Tolerance = 0.36, VIF = 2.81; years a Christian: Tolerance = 0.98, VIF = 1.02; and spiritual development: Tolerance = 0.93, VIF = 1.07).

In step 2, individual and family-centered religious activities were added to the model. Years married (p = 0.024) remained the only control variable that predicted IPV, with couples married fewer years more likely to experience violence. Among the religious activity variables, praying with the children (β = 0.053; t = 2.250; p = 0.009) was the only significant predictor, with those that prayed with their children being more likely to experience IPV. Step 2 was statistically significant (∆R² = 0.004, F(10,3959) = 2.03, p = 0.011). Tests to see if the data met the assumption of collinearity indicated that multicollinearity was not a concern (attend worship: Tolerance = 0.79, VIF = 1.26; read the Bible: Tolerance = 0.58, VIF = 1.74; individual prayer: Tolerance = 0.66, VIF = 1.52; family devotion: Tolerance = 0.69, VIF = 1.48; pray with a partner: Tolerance = 0.64, VIF = 1.56; read/discuss the Bible with a partner: Tolerance = 0.54, VIF = 1.84; discuss spiritual values with the children: Tolerance = 0.47, VIF = 2.14; pray for the children: Tolerance = 0.56, VIF = 1.80; pray with the children: Tolerance = 0.45, VIF = 2.24; share Christ: Tolerance = 0.77, VIF = 1.30).

Christian media-related religious activities were added in step 3. Years married (p = 0.019) was the only control variable that predicted IPV. The only statistically significant religious activity variable that predicted the criterion variable was praying with the children (p = 0.037). Among the Christian media variables, the only significant predictor of IPV was listening to Christian talk radio (β = 0.062; t = 3.110; p = 0.002). Step 3 was statistically significant (∆R² = 0.01, F(7,3952) = 2.27, p < 0.001) and explained 0.7% of the variance over and above other variables. Tests to see if the data met the assumption of collinearity indicated that multicollinearity was not a concern (Christian TV: Tolerance = 0.69, VIF = 1.44; Christian talk radio: Tolerance = 0.62, VIF = 1.62; Christian music: Tolerance = 0.55, VIF = 1.81; Christian book or magazine: Tolerance = 0.59, VIF = 1.70; social networking: Tolerance = 0.70, VIF = 1.42; sermon recordings: Tolerance = 0.81, VIF = 1.24; and Christian websites: Tolerance = 0.68, VIF = 1.48).

Few of the variables of interest predicted IPV. Consistent with previous research that links IPV to younger ages, those that were married longer tended to be less likely to experience violence in the relationship. However, consistent with the findings of Todhunter and Deaton (2010), none of the religious activity variables appearing in previous research predicted IPV. However, praying with the children was a positive predictor of IPV. Among Christian media, only listening to Christian talk radio was significant, predicting a greater likelihood of IPV.

4. Discussion

This study investigated the relationship between marital quality and seventeen religion-related activities in a sample of Protestant Christians from North America. We hypothesized that higher levels of participation in individual, dyadic and family religious activities would predict higher levels of positive marital quality, lower levels of negative interaction and lower levels of IPV, whereas the consumption of Christian media would predict lower levels of positive marital quality, higher levels of negative interaction and higher levels of IPV.

4.1. Control Variables

Before discussing the findings related to our hypotheses, it would be beneficial to examine briefly the outcomes related to the control variables. Based on previous research, we anticipated that being a male, being a Christian for a longer time and having higher levels of spiritual development would predict higher levels of positive marital quality and lower levels of negative interaction as well as IPV and that greater age and longer marriages would predict lower levels of marital quality. Sometimes, these expectations were confirmed. However, often, there was no relationship between the control variable and the criterion variable, and, sometimes, the relationship was in an unexpected direction. These inconsistencies seem to be related to findings by Wilmoth et al. (2015) that demonstrated the interaction effects of age and negative interaction with religiosity on marital satisfaction.

Although the single-item variable Spiritual Development has not been used widely in prior research, it was included as a control variable in these analyses because of the findings of Wilmoth et al. (2018) that Spiritual Development was a consistently salient predictor of individual well-being. As expected, in these analyses, Spiritual Development was one of the most consistent and powerful predictors of marital quality.

4.2. Hypothesis 1

Although attendance at religious services has a complex relationship with marital quality (Booth et al. 1995; Vaaler et al. 2009), generally, attendance has been found to be associated positively with marital quality in a number of studies (e.g., Lichter and Carmalt 2009; Mahoney et al. 1999; Strawbridge et al. 2001) and, in some studies, has served as a single-item proxy for religiosity (e.g., Lim 2015; Wen 2010; Wilmoth et al. 2014; Young 2011). Generally, it is included in scales that measure religiosity (e.g., Koenig and Büssing 2010). However, in this study, attendance did not emerge as a predictor of marital quality in any of the analyses. At least two possible reasons for this surprising non-finding are (1) the religious homogeneity of this sample: All participants were associated with a Protestant church, whose members typically attend more frequently than some other Christian groups (Pew Research Center 2015); (2) the likely effect of multicollinearity, with the control variables explaining the variance otherwise explained by attendance. For example, although test results for multicollinearity were within acceptable limits, Spiritual Development was correlated with every other predictor variable at a level of p < 0.001.

Both prayer and Bible reading/study generally are included in scales measuring religiosity (e.g., Koenig and Büssing 2010), and both have been associated positively with marital quality (Fincham and Beach 2014; Marks 2005; Phillips et al. forthcoming; Robinson 1994). However, individual prayer never emerged as a predictor of marital quality in this study, possibly for the same reasons suggested above for attendance at religious services. In addition to religious homogeneity as a possible explanation, we speculate that a pious individual in a lower-quality relationship is likely to perceive prayer as a means of improving the quality of the relationship. This conjecture is consistent with the findings of Booth et al. (1995) that an increase in marital happiness increases the extent to which religion influences church service attendance as well as the conclusion of Doss et al. (2009) that reading self-help books with relationship themes reflected help-seeking for individuals in distressed marriages.

Although observing religious rituals has been associated with multiple benefits for marriages, having family devotions did not predict positive marital quality in our study. This non-finding is consistent with the expectation that family devotions can be a bonding experience for couples and families but a source of conflict and resentment for other families.

Joint religious activities such as praying together or reading the Bible together have been associated with positive marital quality (Ellison et al. 2010; Lichter and Carmalt 2009). For example, Lichter and Carmalt found that, for both husbands and wives, high rates of joint participation (e.g., praying together) was positively associated with higher scores in each level of marital quality. Similarly, Mahoney and colleagues (1999) found that participating in joint religious activities, including praying together and discussing spiritual issues, was related to higher levels of marital adjustment and perceived benefits from marriage. As expected, these variables were positively correlated with positive marital quality in our study.

The communication of parents with their children about values (e.g., sexual, Suleiman et al. 2016) generally is associated with positive results, and discussing faith values with children is considered important for the intergenerational transmission of faith. Longitudinal research has found that transmitting religious values promoted emotional closeness between parents and children (Bengtson et al. 2013). Thus, we anticipated that talking to children about spiritual values would predict a higher couple relationship quality. However, this variable had a significant negative relationship with positive marital quality. We suspect that a religious parent would discuss spiritual values with children to guard against deleterious effects of a negative relationship between parents.

The only Christian media variables that predicted positive marital quality were listening to Christian talk radio and viewing Christian websites or blogs. In both cases, the association was in the expected direction: Individuals with lower levels of positive marital quality were more likely to listen to Christian talk radio or to view Christian websites or blogs. Although we could not find research related specifically to Christian talk radio, Rubin (2000) found that talk radio listeners often perceive a host to be a relational partner and credible source of information. We speculate that individuals in troubled relationships could seek out Christian talk radio programs, particularly those focused on family-centered topics (e.g., FamilyLife and Focus on the Family) to find solutions to relationship problems. Similarly, an individual in an unhappy relationship could seek out relationship advice on Christian websites or blogs.

4.3. Negative Interaction

Most research investigating the relationship between religious behavior and marital quality does not distinguish between positive marital quality and negative interaction. Thus, most of the discussion in the previous section is relevant for our comments regarding negative interaction. However, there are some specific findings that merit additional comments.

Mahoney’s review (Mahoney 2005) noted that greater levels of religiousness have not been associated with an increase in maladaptive communication in marriage. Mahoney et al. (1999) used several measures of marital quality, including the frequency of conflict. Though conflict is not synonymous with a negative interaction, the concepts are closely related. Mahoney et al. found that joint religious activities such as praying together were associated with lower levels of conflict. In contrast, Booth et al. (1995) found that increases in religiosity were not related to decreases in marital conflict.

Although having family devotions did not have a significant relationship with positive marital quality or IPV, it had a negative association with negative interaction in relationships (p = 0.007). We suspect this may reflect the finding of Booth et al. (1995) that increases in marital happiness predicted greater religious involvement.

As expected, joint religious activities such as praying together or reading the Bible together negatively predicted negative interaction. This is in line with the findings of Mahoney and colleagues (1999), who found that praying together and other religious activities predicted lower levels of conflict.

The communication of parents with their children about spiritual values had a positive relationship with negative interaction. As mentioned in the previous section, we suspect that a religious parent would discuss spiritual values with their children when concerned about the potential negative effects of a distressed relationship between parents. Praying with children also had a positive relationship with negative interaction, and we speculate that the same principle would be at work here.

Confirming Hypothesis 2, individuals that listened more frequently to Christian talk radio or that viewed Christian websites more often had higher levels of negative interaction. As discussed earlier, we suspect this behavior is driven by the desire to find answers for unhappy relationships.

4.4. Intimate Partner Violence

Some additional observations about IPV are relevant. The number of years married predicted lower levels of IPV. One explanation for this association is that couples that experienced IPV would be likely to dissolve their relationships earlier. This is consistent with observations that young age may be a risk factor for IPV among adults (Johnson et al. 2014).

The only individual, dyadic or family religious activity that predicted IPV was praying with the children. As discussed earlier, this association likely indicates an attempt by a victim of IPV to find helpful information or spiritual support. The failure of religious activities to predict IPV was consistent with the earlier findings of Todhunter and Deaton (2010) but differed from the findings of Ellison et al. (2007).

Listening to Christian talk radio is the only media-related variable that predicted IPV. This finding is consistent with Hypothesis 2. Although no previous research has examined the relationship between Christian talk radio and IPV—or marital quality in general—this finding is expected in light of the research by Doss et al. (2009) that couples facing relationship challenges sought help by reading books and of the study by DeAngelis et al. (2018) that showed individuals use the reading of scripture as a help-seeking behavior. Listening to talk radio should be an even more likely target of help-seeking in light of the findings by Rubin (2000) that talk radio listeners often consider a host to be not only a relational partner but also a credible source of information.

Although there is no way to know the severity of violence in any of these relationships, the definition of IPV in this study used less violent behaviors than many other measures (cf., Smith et al. 2018). Also, fewer than 5 percent of the respondents reported incidents of IPV, compared to more than 30 percent in the U.S. population (Smith et al. 2018).

5. Conclusions

This research has several limitations, some of which are related to the sample. Because the participants all were associated with Protestant Christian churches, the findings cannot be generalized to other religious groups and, considering the diversity within Protestantism, cannot be generalized to specific Protestant denominations. Although the sample was not selected at random, the expected large percentage of responses within each congregation suggests responses likely are representative of each participating congregation; however, the congregations were not selected randomly and possibly are not representative of Protestant congregations in general. The use of denominational affiliation would have been a useful control variable, but that information was not available to the authors.

Also, there are limitations related to the instrument. For example, the questionnaire did not include traditional demographic variables such as ethnicity, education or income. Also, the questions related to Christian media were exploratory in nature and should be refined to answer further questions more accurately and thoroughly. DeAngelis et al. (2018) suggested that more nuanced measures should be used to study why individuals read scriptures, and the same principle would be relevant for research into why individuals use various Christian media. For example, it would have been helpful to address the motivations for using these media: Were participants seeking help for distressed marriages, or were they utilizing the media for personal enrichment or pleasure? Another limitation of the data is that only individual data were collected, whereas couple data would be helpful in future research. Despite these limitations, we believe our analyses provide new information that is beneficial for both academics and practitioners.

This study was conducted with the expectation that religious activities and Christian media-related activities would predict higher relationship quality while the use of Christian media would predict lower relationship quality. Generally, our expectations were correct, although certain religious activities (e.g., praying with the children) were associated with lower levels of positive marital quality, higher levels of negative interaction and a greater likelihood of IPV. We suspect that any causation in these relationships, particularly between media use and relationship quality, might be bidirectional or flow from the relationship quality to the religious activities or media consumption. Perhaps individuals in poor-quality relationships seek out media (and other religious resources) looking for ways to improve their marriages. In contrast, couples with stable and satisfying marriages would be more likely to read the Bible or to pray together. Further, we suspect that some religious activities are perceived to provide preventive or remedial benefits by spouses in distressed relationships. It also should be noted that rates of IPV were much lower in this sample than in the general population. Clergy and family practitioners in religious settings can use this information to identify possible problematic relationships, and purveyors of Christian media can use the information to craft content that would address relationship issues.