Something to talk about: Topics of conversation between romantic partners during military deployments

Abstract

Long-distance communication has been frequently identified as essential to military couples trying to maintain their relationship during a military deployment. Little quantitative research, however, has assessed the types of topics discussed during such communication and how those topics relate to overall relationship satisfaction. The current study draws on a sample of 56 Army couples who provided data through online surveys while the service member was actively deployed. These couples provided information on current marital satisfaction, topics discussed during deployment (problem talk, friendship talk, love talk), and how they communicated via synchronous media (e.g., phone calls, video calls) and letters during deployment. Non-parametric Friedman tests followed by paired t-tests revealed that synchronous communication was primarily utilized for friendship talk, whereas letters included friendship talk and love talk in similar amounts. Both synchronous communication and letters included less problem talk than other topics. In mixed level modeling, only topics of communication for synchronous media (not for letters) were related to relationship satisfaction. Love talk via synchronous media was related to higher relationship satisfaction, whereas problem talk via synchronous media was related to less relationship satisfaction. The current study offers the first quantitative assessment of topics within deployment communication media and associations with relationship satisfaction.

U.S. service members (SMs) have experienced high rates of deployment in recent years, with over 2 million SMs deployed to combat zones since 2001 and regular deployments likely to continue (Ryan & Raghavan, 2016; Starr, 2016; Tan, 2009; Tilghman, 2016). Consistent with the general population, approximately half of SMs are married (Department of Defense, 2015). SMs and their romantic partners have often reported that deployments are stressful and can have negative impacts on their relationships (Andres, 2014; Karney & Trail, 2016; Greene, Buckman, Dandeker, & Greenberg, 2010; Newby et al., 2005; Turner & Chessor, 2015). However, deployments have also been reported to potentially strengthen relationships and enhance closeness between partners (Knobloch, Basinger, Wehrman, Ebata, & McGlaughlin, 2016; Knobloch & Theiss, 2012). Military couples have reported on the importance of communication in preserving their romantic connection during a deployment (Knobloch et al., 2016), and leading models of relationship maintenance in geographically separated couples emphasize communication as integral (e.g., Merolla, 2010a; Stafford & Canary, 1991). To build on these previous studies, the main aim of the current study is to explore communication in military couples separated by deployment through assessing how different topics and types of communication media used relate to relationship satisfaction.

One theory that focuses on the effects of different types of communication media (e.g., phone calls, letters) is niche theory. Niche theory, as explored by Dimmick and colleagues (2000), posits that different communication media offer unique gratifications for users. One important feature that is posited to be related to the unique gratifications is synchronicity. Synchronicity refers to the degree of immediacy in communication. It is represented as a spectrum, with media such as phone and video calls being highly synchronous, as they are interactive and occur in real-time, and media such as letters and care packages having low synchronicity, since there is a significant delay between missives. Dimmick and colleagues (2000) found that synchronous media (e.g., phone calls) were heavily favored by users, due to the immediate exchange of information and the ability to hear another individual's voice. At the same time, less synchronous, written communication is theorized to offer more overall opportunities for gratification from communication. These opportunities arise from the convenience of writing (e.g., not needing to coordinate schedules) and the ability to re-read messages multiple times, meaning that less synchronous communication is more easily written and consumed. Furthermore, the variations in communication media result in differing suitability for specific topics of communication (e.g., discussing problems, chatting about the day). Significant civilian research supports that users may be choosing different communication media for distinct tasks, such as relying on phone calls for problem solving (e.g., Dainton & Aylor, 2002; Petrič, Petrovčič, & Vehovar, 2011). To date, however, little is known regarding what military couples communicate about during deployment and whether or how topics of communication vary across communication media (Carter & Renshaw, 2016). Thus, the first aim of the present study was to address the gap in empirical data on the presence of different topics of communication during deployment. Consistent with niche theory, we explored the proportions of communication topics discussed, differentiated across different types of communication media.

To our knowledge only one study (with a sample of six partners of deployed SMs) has directly assessed the content of communication during deployment (Kim et al., 2005). These six partners provided daily reports of communication, analyses of which indicated that the most frequent topic of conversations was small talk (e.g., recapping the day, chatting). Studies of more general relationship maintenance during deployments have yielded analogous reports of "debriefing talk," wherein couples discussed their day and chatted about miscellaneous current events in their lives (Maguire, Heinemann-LaFave, & Sahlstein, 2013; Merolla, 2010a). Similar findings also arise in studies of communication within non-military couples, where everyday talk (e.g., discussing the day, chit chat) appears to be an important part of maintaining interpersonal relationships (Goldsmith & Baxter, 1996; Schrodt, Soliz, & Braithwaite, 2008; Vangelisti & Banski, 1993). These overlapping constructs can be conceptually combined into friendship talk, wherein couples discuss a variety of positive topics, such as recapping the day's events, sharing positive news, or engaging in chit-chat (Markman, Stanley, & Blumberg, 2010). Based on qualitative and civilian research, it is likely that friendship talk is important to military couples during deployments, but little empirical information is available regarding its prominence in communication or association with relationship satisfaction.

Several studies of couples during military deployments have also emphasized the importance of love talk, such as expressing affection and making statements of love and assurance to one another. Kim and colleagues (2005) found that this type of communication was the most frequently reported second to small talk. Qualitative studies of military couples outside of deployment, and the relationship literature in general, have found that offering assurances and positivity is also related to higher relationship satisfaction (e.g., Maguire, Heinemann-LaFave, & Sahlstein, 2013; Merolla, 2010a; Weigel & Ballard-Reisch, 2008). Again, to our knowledge, there are no other empirical data available on the prominence of such communication during deployments or its association with marital satisfaction.

Even less is known about problem-focused communication, or problem talk, during deployments. Research with civilians has found the presence and intensity of conflict can negatively impact relationship satisfaction (e.g., Cramer, 2000; Stanley, Markman, & Whitton, 2002), suggesting that higher levels of problem-focused talk may be detrimental for deployed couples. On the other hand, civilian couples in highly satisfied long-distance relationships have reported that joint problem solving actually increases satisfaction, which indicates that discussing problems with a romantic partner is not inherently harmful to a relationship (Maguire & Kinney, 2010). To date, research on deployment communication has primarily explored reasons for avoiding problem talk (e.g., Joseph & Afifi, 2010; Rossetto, 2013), with no empirical examination of the prominence of such communication or its association with overall relationship satisfaction.

Qualitative literature of SMs and partners offer preliminary support for niche theory, in that couples may be choosing different communication media for distinct topics of communication. Some partners report that synchronous communication (e.g., phone calls, video calls) is often used for more shallow and quick conversations (Smith-Osborne & Jani, 2014). Also, synchronous communication may be less suited for love and problem talk, due to the lack of privacy that many SMs have when having phone or video conversations (MacDermid et al., 2005) and concerns of operation safety (Hinojosa, Hinojosa, & Hognas, 2012). The inherent and often lengthy delay between letters being written and received may make letters less suited for friendship and problem talk. Thus, letters may be more likely to be utilized for love talk. Indeed, one qualitative study quoted a non-deployed partner as saying "he would always write really relaxing letters, just about 'I can’t wait to see you, I miss you,' that kind of thing” (Faber, Willerton, Clymer, MacDermid, & Weiss, 2008, p. 226). Thus, we hypothesized that synchronous communication would contain significantly more friendship talk than love and problem talk (Hypothesis 1), whereas letters would contain significantly more love talk than friendship and problem talk (Hypothesis 2).

The second aim of this study was to examine the associations between topics of communication in letters and in synchronous communication with deployed SMs’ and partners’ marital satisfaction. Per niche theory, letters may not only be more likely than synchronous media to contain love talk, but they may be better suited for conveying love talk. For instance, despite the increasing availability of multiple forms of synchronous communication, sending and receiving letters has consistently been found to have a positive association with relationship satisfaction in military couples whereas synchronous communication has resulted in smaller or no associations (e.g., Carter & Renshaw, 2016a; Caska & Renshaw, 2011; Ponder & Aguirre, 2012). Letters can be read multiple times and when individuals are most in need of support, which may allow a single letter with statements of love and support to influence individuals on multiple occasions (see review by Carter & Renshaw, 2016b). Also, letters are one of the few media that non-deployed spouses can initiate; only SMs are able to initiate synchronous communication (e.g., phone calls, video calls). When partners take the time to initiate communication via letters, the statements therein may have more impact on SMs (Carter & Renshaw, 2016a). Thus, it is hypothesized that love talk via letters will have a stronger, positive association with relationship satisfaction than love talk via synchronous conversations (Hypothesis 3).

There may also be important differences in the impact of problem talk via different communication media. The unique features that make letters ideal for love talk (e.g., the ability to be reread, partners being able to initiate) may result in problem talk via letters being damaging for relationships. For instance, statements perceived as negative or challenging may be reread several times, rather than just being heard once. As noted above, letters also have a slower delivery rate that may simply make them less useful for problem talk, which could result in less problem talk in letters. These reasons may be why SMs have reported using the phone more often than letters to try to resolve disagreements (Schumm, Bell, Ender, & Rice, 2004). However, problem talk by phone may be problematic for SMs and their partners, as well, as the ability to hear or even see your partner during discussions may make problem talk more likely to escalate. For instance, it may be difficult to mask frustration, hurt, or anger over the phone (Pincus, House, Christenson, & Alder, 2007). Also, technical difficulties in real-time may escalate conflict. For instance, partners have described getting cut-off during a heated discussion and assuming the SM had hung up, and bad connections may increase existing frustrations (Hinojosa et al., 2012). Thus, while problem talk may have a negative relationship with satisfaction across media, it is hypothesized that problem talk via synchronous conversations will have a stronger, negative association with relationship satisfaction than problem talk via letters (Hypothesis 4).

Finally, the impact of friendship talk may also vary by the communication medium. Qualitative studies have consistently discussed friendship talk as being beneficial for relationships (e.g., MacDermid et al., 2005; Merolla, 2010a). However, the benefits of friendship talk are overwhelmingly described within the context of synchronous communication. For instance, partners have reported that making small talk about their day or discussing what is going on at home on phone calls allows SMs to be up to date and helps to create a sense of normalcy (Merolla, 2010a). Others have described synchronous communication as allowing SMs to feel as if they are actively participating in the family by hearing their children play or finding out what is happening back home (e.g., what homework is assigned that day) in real time (MacDermid et al., 2005). The ability to hear family members and communicate in real-time may make synchronous communication uniquely suited towards friendship talk. In contrast, given that letters are inherently delayed by days or weeks after they are written, such benefits may not occur from friendship talk in letters. Thus, it is hypothesized that friendship talk via synchronous conversations will have a stronger, positive association with relationship satisfaction than friendship talk via letters (Hypothesis 5).

Overall, the study seeks to advance the current understanding of couple communication during deployment. By first assessing how the topics that couples discuss during a deployment vary across communication media, we can better understand how couples are trying to connect during their separation. By next exploring how topics of communication within letters and synchronous media relate to both SM and partner marital satisfaction, we can reveal a more nuanced understanding of how deployment communication is associated with SM intimate relationship functioning.

Methods

Participants and Procedure

Data were drawn from a large, randomized clinical trial of a marriage education program for Army couples, in which 662 couples participated from two separate installations. To be eligible for the study, couples had to be (1) currently married, (2) age 18 or older, (3) fluent in English, and (4) have at least one active-duty Army SM in the couple. Study procedures were approved by the University of Denver Institutional Review Board (Protocol #471733, "Marriage Education and Risk Reduction for Army Families”). A detailed description of the recruitment and intervention procedures can be found in Stanley, Allen, Markman, Rhoades, & Prentice, 2010. Couples were given follow-up assessments every 6 to 9 months following the intervention, including couples that were divorced, separated from the military, or were physically separated due to training or deployment. Retention across time points was very strong (see Stanley et al., 2014).

During the eighth and ninth follow-up time points (approximately 5–6 years after the original intervention), couples in which the SM was deployed were given a separate set of deployment-specific measures. Of the overall sample, only 56 couples indicated that (1) the SM was deployed at the time of filling out the eighth or ninth follow-up questionnaire, and that (2) the couple was still married. Other participants from the larger study were ineligible mainly due to not being deployed during the eighth or ninth follow-up (including those that were not deployed due to leaving the Army altogether) or due to no longer being married to the partner they entered the study with. Data from these time points were collected between 2011 and 2012, and included deployments primarily in support of Operations Iraqi and Enduring Freedom. Of the 56 couples selected for the current study, 55 couples were comprised of a male service member and female civilian partner, and one couple was dual-military in which the female partner was currently deployed.

Deployed SMs were, on average, 34.27 years old (SD = 5.95), and non-deployed partners’ average age was 33.36 years old (SD = 5.77). Couples reported being married on average for 10.12 years (SD = 4.40). The majority of participants identified as Caucasian (61% of deployed partners; 66% of non-deployed partners), followed by Hispanic (18% of deployed partners; 16% of non-deployed partners), African-American (11% of deployed partners; 7% of non-deployed partners), Asian (2% of deployed partners), Hawaiian or Pacific Islander (4% of deployed and non-deployed partners), Native American (4% of non-deployed partners), or Mixed (5% of deployed partners; 4% of non-deployed partners). Of the deployed SMs, 2% were Specialists, 39% were Jr. NCOs, 20% were Sr. NCOs, 23% were Company Grade Officers, and 7% were Field Grade Officers. The remaining 9% did not report their rank. SMs had on average served 13.49 years in the military (SD = 6.29).

Measures

Communication Topics

Participants completed an 8-item measure to assess topics of communication during deployment. Four items assessed communication topics in real-time conversations and four parallel items assessed communication topics in letters. There were two items for each medium that assessed problem talk: “During deployment, what proportion of you and your spouse’s conversations [letters] focus on…”: (1) “minor, day-to-day problems”, (2) “major, significant problems”. A single proportion score was created by averaging responses to these two items. One item assessed love talk: “During deployment, what proportion of you and your spouse’s conversations [letters] focus on love talk (for example, expressions of love, affection, appreciation, and encouragement)?”. Finally, one item assessed friendship talk: “During deployment, what proportion of you and your spouse’s conversations [letters] focus on friendship talk (for example, discussing positive events, shared interests, and day-to-day chit-chat)?” Participants answered each item on a scale ranging from 0 (0%) to 10 (100%). However, the sum of responses to the above questions for each medium (real-time or letters) were not forced to sum up to 100%, resulting in the total proportion of communication within each medium often deviating from 100% (e.g., a participant indicating that 100% of their conversations were about all topics). Thus, in order to produce relative scores, we divided the score for each medium by the sum of all topic proportions for that medium. For instance, for real-time conversations, we divided the proportion score for problem talk by the sum of the proportion scores for problem talk, love talk, and friendship talk. Similar calculations were made for love talk in real-time conversations, friendship talk in real-time conversations, problem talk in letters, love talk in letters, and friendship talk in letters.

Communication Frequency

To measure communication frequency during deployment, participants indicated how often they were keeping in contact with their spouse while deployed via thirteen communication media, using a Likert scale (0 = never, 1 = 1–2 times per month, 2 = 1–2 times per week, 3 = 3–4 times per week, and 4 = daily). The frequency of letter communication was taken from a single item. Synchronous communication frequency included the frequency of phone calls, instant messaging, IM with video and/or webcam, and internet chat with and without video.

Marital satisfaction

The Kansas Marital Satisfaction Scale (KMS; Schumm, Paff-Bergen, Hatch, & Obiorah, 1986) is a 3-item measure designed to assess marital satisfaction. This scale was used to assess participants’ satisfaction with their partner and their relationship. Participants responded using a 7-point scale (1 = extremely dissatisfied, 7 = extremely satisfied). Items included: “How satisfied are you with your marriage?”, “How satisfied are you with your partner as a spouse?”, and “How satisfied are you with your relationship with your spouse?” In previous studies (e.g. Schumm et al., 1986) this measure has demonstrated strong reliability and validity. In the currently sample, items demonstrated excellent internal consistency for both the deployed partner (α = .98) and the non-deployed partner (α = .99).

Analyses

Prior to analyses, the data were assessed for outliers. In line with recommendations (e.g., Aguinis, Gottfredson, & Joo, 2013), standardized DFBeta scores were generated from separate regressions for SMs and partners of marital satisfaction on letter topics and on synchronous communication topics. The deployed SM of one couple was identified as a clear outlier, with four of six standardized DFBETAs > 1.00. Elimination of this couple yielded substantial changes in results, consistent with their undue influence. Thus, all results are reported without this couple. Additionally, to ensure that communication media were adequately utilized by participants, we evaluated frequencies of the different communication media.

For hypotheses 1 and 2, an omnibus non-parametric Friedman test of differences was conducted (for additional information, see García, Fernández, Luengo, & Herrera, 2010) that assessed differences in conversation topics (problem talk, friendship talk, love talk) within a single medium. A total of four Freidman tests were conducted: (1) SM reports of synchronous topics, (2) partner reports of synchronous topics, (3) SM reports of letter topics, and (4) partner reports of letter topics. For each statistically significant Freidman test, follow-up paired t-tests were completed to ascertain direct differences between topics.

To assess hypothesis 3 through 5, bivariate correlations were first run. Multilevel modeling was then utilized to assess individuals (Level 1) nested within couples (Level 2). Three separate models were created to assess the associations of relationship satisfaction with (1) love talk in synchronous communication and letters (2) problem talk in synchronous communication and letters, and (3) friendship talk in synchronous communication and letters. To control for any effects of the marriage education program on relationship satisfaction, a dichotomous Level 2 variable designating whether couples were in the intervention or control group was also included. The following model was utilized wherein SYNC represents the communication topic in synchronous communication and LETTER represents the communication topic in letters:

$${\mathit{\text{MARSAT}}}_{\mathit{\text{ij}}}={\beta }_{\mathit{0}j}+{\beta }_{\mathit{1}j}\ast ({\text{SYNC}}_{\mathit{\text{ij}}})+{\beta }_{\mathit{2}j}\ast ({\mathit{\text{LETTER}}}_{\mathit{\text{ij}}})+r_{\mathit{\text{ij}}}$$ Level-1 Model

$$\begin{array}{c}{\beta }_{\mathit{0}j}={\gamma }_{\mathit{00}}+{\gamma }_{\mathit{01}}\ast ({\mathit{\text{GROUP}}}_j)+u_{\mathit{0}j}\hfill \\ {\beta }_{\mathit{1}j}={\gamma }_{\mathit{10}}\hfill \\ {\beta }_{\mathit{2}j}={\gamma }_{\mathit{20}}\hfill \end{array}$$ Level-2 Model

Results

Frequencies of synchronous and letter communication are presented in Table 1. For all media, at least half of participants reported using them at least 1 to 2 times per month, with letters being utilized least often. Non-parametric Friedman test of differences indicated that the proportions with which each topic was discussed in synchronous communication were significantly different when reported by both SMs (χ²(2,46) = 29.20, p < .001) and partners (χ²(2,53) = 22.46, p < .001). Hypothesis 1, that synchronous communication would contain significantly more friendship talk than love and problem talk, was supported by follow-up paired t-tests that found synchronous communication consisted of significantly more friendship talk than love talk (SM: t(45) = 4.03, p < .001; Partner: t(52) = 2.95, p = .005) and problem talk (SM: t(45) = 5.09, p < .001; Partner: t(52) = 5.28, p < .001). Hypothesis 2, that letters would contain significantly more love talk than friendship and problem talk, was partially supported. A non-parametric Friedman test of differences indicated that the proportions with which each topic was discussed in letters were significantly different when reported by both SMs (χ²(2,27) = 27.68, p < .001) and partners (χ²(2,30) = 23.41, p < .001). Follow-up paired t-tests that found that hypothesis 2 was partially supported, as there was significantly more love talk (SM: t(26) = 3.16, p = .004; Partner: t(29) = 3.61, p = .001) and friendship talk (SM: t(26) = 5.90, p = .000; Partner: t(29) = 5.08, p < .001) than problem talk. However, contrary to hypothesis 2, there was no significant difference emerged between love and friendship talk (SM: t(26) = 0.04, p = .966; Partner: t(29) = 0.88, p = .384).

Table 1.

Frequency of Communication Media Use during Deployment

	Never		1–2 times per month		1–2 times per week		3–4 times per week		Daily
	N	%	N	%	N	%	N	%	N	%
Letters	19 (27)	39.6% (50.0%)	9 (8)	18.8% (14.8%)	2 (3)	4.2% (5.6%)	4 (5)	8.3% (9.3%)	14 (11)	29.2% (20.4%)
Phone Calls	4 (3)	8.3% (5.6%)	11 (17)	22.9% (31.5%)	12 (10)	25.0% (18.5%)	10 (14)	20.8% (25.9%)	11 (10)	22.9% (18.5%)
Instant Messaging	11 (17)	23.4% (31.5%)	6 (11)	12.8% (20.4%)	10 (9)	21.3% (16.7%)	7 (5)	14.9% (9.3%)	13 (12)	27.7% (22.2%)
IM with Video and/or Webcam	11 (15)	22.9% (28.3%)	12 (9)	25.0% (17.0%)	7 (9)	14.6% (17.0%)	9 (9)	18.8% (17.0%)	9 (11)	18.8% (20.8%)
Internet chat or voice calls with video	12 (13)	25.0% (24.1%)	10 (12)	20.8% (22.2%)	12 (8)	25.0% (14.8%)	7 (11)	14.6% (20.4%)	7 (10)	14.6% (18.5%)

Note. Service member reports are included first with partner reports following in parentheses.

To first assess hypotheses 3–5, bivariate correlations were conducted¹ (see Table 2) followed by an unconditional multilevel model in order to assess what proportion of the variance was accounted for by nesting within couples. An intraclass correlation coefficient of .49 indicated that a nested multilevel model was warranted. Results from the subsequent multilevel models are reported in Table 3.

Table 2.

Means, Standard Deviations, and Correlations

	Service Member Mean (SD)	Partner Mean (SD)	1	2	3	4	5	6	7
1. Marital Satisfaction	6.17 (1.28)	5.81 (1.46)	--	.24	.20	−.45**	.08	−.03	−.04
Synchronous Conversations
2. Friendship talk	4.47 (1.35)	4.51 (1.81)	.44**	--	−.52***	−.50***	.57**	−.22	−.31
3. Love Talk	3.34 (1.35)	3.23 (1.80)	.33*	.01	--	−.49***	−.16	.58**	−.52**
4. Problem Talk	2.20 (1.92)	2.26 (1.77)	−.54***	−.71***	−.71***	--	−.35	−.35	.75***
Letters
5. Friendship talk	4.34 (1.74)	4.03 (1.73)	−.45*	.28	−.35	.07	--	−.58**	−.32
6. Love Talk	4.37 (1.93)	4.56 (2.03)	.33	−.08	.44*	−.26	−.65***	--	−.59**
7. Problem Talk	1.30 (1.55)	1.41 (1.75)	.10	−.21	−.16	.25	−.32	−.52**	--

Note. Correlations below the diagonal are service member reports. Correlations above the diagonal are partner reports.

^*p < .05.

^**p < .01.

^***p < .001

Table 3.

Multilevel Model Results Predicting Relationship Satisfaction From Communication Topics

	Synchronous Communication1			Letters1			Group2
	Unstandardized Coefficient	SE	t	Unstandardized Coefficient	SE	t	Unstandardized Coefficient	SE	t
Love talk	0.26	0.10	2.51*	−0.08	0.09	−0.92	−0.31	0.30	−1.02
Friendship talk	0.15	0.11	1.34	−0.10	0.09	−1.03	−0.20	0.33	−0.62
Problem talk	−0.29	0.08	−3.49**	0.20	0.10	1.97	−0.33	0.29	−1.15

¹Indicates a level 1 variable.

²Indicates a level 2 variable.

^*p < .05.

^**p < .01.

Hypothesis 3 was that love talk via letters would have a stronger, positive association with relationship satisfaction than love talk via conversations. In contrast to this hypothesis, only the correlation between SMs’ report of love talk via synchronous communication and SMs’ report of relationship satisfaction was significant and positive. Love talk via letters was nonsignificantly correlated with satisfaction in both SMs and partners. Moreover, in the multilevel model assessing the associations of love talk via synchronous communication and via letters with relationship satisfaction, only love talk via synchronous communication was positively and significantly associated with relationship satisfaction. Thus, Hypothesis 3 was not supported.

Hypothesis 4 was that problem talk via synchronous conversations would have a stronger, negative association with relationship satisfaction than problem talk via letters. For both SMs and partners, problem talk via synchronous communication was significantly and negatively correlated with their relationship satisfaction. There was no significant correlation between problem talk via letters and relationship satisfaction. In the multilevel model assessing the associations of problem talk via synchronous communication and via letters with relationship satisfaction, problem talk via synchronous communication was significantly and negatively associated with relationship satisfaction, while problem talk via letters was not significantly associated. Thus, Hypothesis 4 was supported.

Finally, Hypothesis 5 stated that friendship talk via synchronous conversations would have a stronger, positive association with relationship satisfaction than friendship talk via letters. Although no significant correlations emerged for partners, there was a significant and positive correlation between friendship talk via synchronous communication and relationship satisfaction for SMs. Surprisingly, for SMs, there was also a significant and negative correlation between friendship talk via letters and relationship satisfaction. In the multilevel model, neither friendship talk in synchronous communication or friendship talk in letters were significantly associated with relationship satisfaction. Thus, Hypothesis 5 was not supported by the statistically rigorous multilevel model.

Discussion

To our knowledge, this study is the first to quantitatively assess specific topics of discussion across different communication media in military couples during a deployment. The frequency of use of communication media is consistent with previous studies which have found high levels of synchronous communication and regular use of letters (see review by Carter & Renshaw, 2016b). In line with niche theory, couples reported different proportions of communication topics across different media. Couples reported engaging in significantly more friendship talk than love or problem talk in synchronous communications (e.g., phone calls, video calls), with friendship talk making up almost half of synchronous communication. In contrast, couples reported similar levels of friendship talk and love talk, with very little problem talk, in letters. The high frequency of friendship talk in both synchronous communication and letters indicates that couples use communication, in large part, to catch up with each other, joke around, or simply chat. Overall, these patterns indicate that couples are employing communication media for differing uses, offering some of the first empirical support for the applicability of niche theory within the context of communication during a military deployment.

When assessing how communication topics may influence relationship satisfaction, love talk was found to be significantly and positively associated with relationship satisfaction, but only when communicated via synchronous communication. This finding is contrary to our hypothesis and to findings from previous research, which have consistently revealed love talk via synchronous communication to be less associated with relationship satisfaction than love talk via letters (Carter & Renshaw, 2016a; Caska & Renshaw, 2011; Ponder & Aguirre, 2012). The divergence from previous literature may be, in part, due to the current sample's high access to synchronous communication media. Given that SMs in the current study completed the online surveys while deployed, they were likely deployed to areas that had consistent and reliable internet access. Additionally, the current sample was deployed between 2011 and 2012. Previous studies utilized participants with earlier OIF/OEF deployments, which often had less consistent access to synchronous communication media (e.g., MacDermid et al., 2005). Overall, the current findings suggest that expressing feelings of love, affection, and encouragement via synchronous communication is important for maintaining or even improving satisfaction for deployed couples. Synchronous communication uniquely offers real-time conversation. It may be that words of love and affirmation are more potent when made in real time and can be immediately returned. Additionally, synchronous communication often has the benefit of being able to hear a partner's voice and, through video calls, potentially even see partners. Directly hearing or seeing one's partner express the sentiment may make it more impactful for relationship satisfaction. Due to our data being cross-sectional, it is also possible that highly satisfied couples may simply be more likely to engage in love talk than couples who are less satisfied with their relationship.

Problem talk was significantly and negatively related to relationship satisfaction, but only when communicated via synchronous media. Not only is problem talk more frequent in synchronous communication, it may be that negative communication patterns such as conflict escalation, invalidation, or withdrawal are more apt to occur when couples are interacting in real time. For instance, the escalation of conflict often occurs when couples react to each other's negative affect (e.g., irritation) building up negative affect and behaviors (e.g., Fincham, 2004). Being able to hear or see a partners' negative affect over video call or the telephone and then respond in real time may cause problem talk to have negative impacts on relationship satisfaction and functioning. Of note, it is also quite likely that couples who are more dissatisfied are more likely to engage in problem discussions; thus, no causality can be inferred by this association. Communication in letters, which does not take place in real time and does not include audio or visual cues, may be less prone to these problematic patterns. At the same time, the lack of association of marital satisfaction with problem talk via letters in our sample may well be due to the low levels (and, thus, low range) of this type of communication. Although the current study was unable to test the intensity, behavior, or specific content of the problem talk, future research would benefit from directly assessing how problem discussion may differ across communication media.

When assessing the association between friendship talk and relationship satisfaction, correlations showed that SMs did indicate that friendship talk via synchronous communication was positively correlated with their own relationship satisfaction. Surprisingly, SMs also reported a significant and negative association between friendship talk via letters and their own relationship satisfaction. It is possible that less satisfied couples may use friendship talk in letters to maintain a level of superficiality in content as a means of limiting emotional closeness. In qualitative studies, SMs have reported emotionally distancing themselves from their romantic partner and family to concentrate more fully on the mission (Hinojosa et al., 2012). It may also be the case that SMs are dissatisfied when letters are primarily filled with friendship talk rather than more meaningful statements (e.g., statements of love or of missing each other). Notably, when assessed within a multilevel model, friendship talk via synchronous communication or letters was not significantly related to relationship satisfaction. Thus, when accounting for the nested nature of the data and assessing friendship talk via both media and including partners simultaneously, the significant associations in the correlations disappear. Assessing the hypothesis in a multilevel model may have significantly reduced the power such that the ability to elucidate associations was minimized. Additionally, friendship talk between the two media were likely linked; thus, including them in a single analysis may have resulted in their accounting for overlapping variance, rendering them both nonsignificant. Overall, the hypothesis was supported by correlations but was not supported by the more statistically rigorous multilevel model. Previous research has found that geographically-close military couples rated everyday talk (a construct with significant conceptual overlap with friendship talk), as more important for their relationships than did civilian couples (Frisby et al., 2011). It may be that friendship talk is serving a purpose outside of relationship satisfaction. For instance, hearing about what is occurring at home and participating in family decisions may help SMs to maintain a place in the family, smoothing later reintegration.

Surprisingly, within multilevel models, no communication topic communication via letters was significantly related to marital satisfaction. This was contrary to current hypotheses and previous research (e.g., Carter & Renshaw, 2016a; Ponder & Aguirre, 2012). This unexpected finding may be due to the rapid shift over the last decade in communication technology that has resulted in a drastic reduction of letters being exchanged (review by Carter & Renshaw, 2016b). Fewer letters being sent may translate into fewer chances for letter-based communication to impact or reflect relationship satisfaction. It should also be considered that, as noted, changes in technology during the time of the study may have lead to a shift from previous findings. Future studies are integral to determine if the current findings replicate and represent a meaningful change. Letters may still benefit individuals through unique features like the ability to carry the letter to be re-read in times of need. It may be that re-reading letters does not directly impact relationship satisfaction, but instead helps with individual coping, stress, and mental health (e.g., Carter et al., 2011).

It is important to consider these findings within the context of study limitations. As noted, the data are cross-sectional. Although findings can elucidate patterns during deployment, they cannot speak to the directionality or long-term impacts of the findings. Additionally, we assessed only synchronous communication and letters, rather than exploring potential differences across multiple communication media (e.g., email, social media, texting, instant messages), only three topics of communication, and used only single-item questions. Some researchers have specified different categories (e.g., Stafford & Canary, 1991) or subcategories of communication (e.g., 29 domains of everyday talk; Goldsmith & Baxter, 1996). More fine-grained variables assessed via multiple questions may have yielded more nuanced findings. We also were not able to adequately account for potentially moderating variables, such as deployment history, combat exposure, and length of deployment. Access to communication media may differ across branch, rank, deployment location, and when the deployment took place (review by Carter & Renshaw, 2016b). Addressing environmental factors related to the deployment in future studies will help to create a more in-depth understanding of deployment communication.

The current sample may also have characteristics which reduce generalizability. All but one couple in the study were made up of a male SM and female partner. Although this composition reflects the most common couple formation in the military, it may not generalize to the gender and couple dynamics that may be present for female SMs, dual military couples, and same-sex couples. For instance, male and female service members have been found to react to deployments differently, and civilian males and females utilize communication media differently (Adler, Huffman, Bliese, & Castro, 2005; Kimbrough, Guadagno, & Muscanell, 2013). These differences could shift how couples communicate when the deployed SM is female and the at-home partner is male. The small sample size may have also reduced power to detect associations. Thus, the current findings should not be assumed to represent patterns taking place in the military or deployments at large.

These limitations notwithstanding, the current study offers the first quantitative assessment of communication topics and media during deployment, as well as their association with relationship satisfaction. The study is also the first to include data collected from both SMs and partners during a deployment, rather than retrospectively. Future research would benefit from further assessment of the communication patterns within deployment communication, such as how couples navigate problems and conflict while apart and a more in-depth assessment of friendship talk in separated military couples. However, given that real-time conversation topics had both positive and negative associations with relationship satisfaction, couples may benefit from communication skill-building through workshops or therapy prior to deployment to help couples to minimize escalation when discussing problems and maximize statements of love and support when communicating in real-time. Additional findings will help to further inform empirically based suggestions for how couples can best maintain their relationship while separated.