Abstract
This study drew on the emotional cycle of deployment model to track the content, valence, and sequence of relationship changes experienced by returning service members and at-home partners during the transition from deployment to reintegration. In a longitudinal study, 555 military couples (1,100 individuals) wrote 7,387 comments describing changes that had occurred in their relationship during the past month. A content analysis identified 10 substantive categories: emotional intimacy, sexual intimacy, spending time together, appraisals of the relationship, life changes, readjustment to daily life, conflict, family changes, commitment, and reports of no change. The frequency of changes reported in emotional intimacy, sexual intimacy, readjustment to daily life, and conflict declined across the transition. In contrast, reports of life changes, and comments stating that no change had occurred, increased over time. Independent coders judged each change as positive (42.1%), negative (32.4%), or neutral (25.5%) in valence. Participants described fewer positive changes as the transition progressed, although this tendency slowed over time. In contrast, the frequency of negative changes remained stable across the transition, and the frequency of neutral changes increased. The findings are used to advance theory, research, policy, and intervention designed to help military couples negotiate relationship changes across the post-deployment transition.
Both news media and social media portray homecoming after deployment as a fairytale moment (e.g., Howard & Prividera, 2015), but the reality is that a service member’s return marks only the first day of a long and important transition for military couples (Bommarito et al., 2017; Meadows et al., 2016; Sherman et al., 2015). Settling into new routines takes patience and perseverance: Returning service members need time to acclimate from deployment to domestic life, and at-home partners need time to adjust from independence to interdependence (Freytes et al., 2017; Karakurt et al., 2013; Sahlstein et al., 2009). In fact, when military couples are asked what advice they would give to others about reintegration, they recommend slowing down and not rushing the process (Knobloch, Basinger, et al., 2016). Because time is required to identify a new normal upon reunion following deployment, scholars have defined the transition as lasting up to 6 months in length (Pincus et al., 2001).
The emotional cycle of deployment model depicts the return home from deployment as a period filled with relationship changes of both positive and negative valence (Pincus et al., 2001), but theory and research have overlooked the sequence of relationship changes that unfold over time. Although empirical work provides a general description of the issues military couples face upon reunion (Balderrama-Durbin et al., 2015; Knobloch & Theiss, 2012), it does not identify the timing of relationship changes or the pivotal junctures that are essential for helping military couples during the months after homecoming.
A more systematic understanding of the trajectories of relationship changes that military couples experience during the transition from deployment to reintegration would benefit theory, research, policy, and practice. For example, it would facilitate theory development about relationship progression occurring within stages of the deployment cycle to complement existing theorizing about changes that happen across stages (e.g., Sahlstein Parcell & Maguire, 2014). Second, it would help organize empirical findings about the reunion period with respect to the timing and tenor of people’s experiences (e.g., Knobloch & Theiss, 2018). Third, it would inform military policy about the best time to offer support services to returning service members and their families. Finally, it would foster clinical recommendations designed to help military couples during the days, weeks, and months following homecoming.
Guided by the emotional cycle of deployment model (Pincus et al., 2001), we used a longitudinal design to investigate the reintegration experiences of returning service members and at-home partners in their own words. Couples described changes to their relationship in seven monthly assessments beginning 30 days after homecoming. Our goal was to document the content, valence, and sequence of the month-to-month relationship changes experienced by military couples across the transition.
Emotional cycle of deployment model
Trajectories are patterns of relationship change that unfold over time (Sahlstein Parcell & Maguire, 2014). Stage models of deployment speak directly to trajectories because they conceptualize deployment as a sequential process. The most prominent of these stage models, the emotional cycle of deployment model, was developed by military psychiatrists relying on their personal experiences, clinical observations, and review of the literature (Pincus et al., 2001).
The emotional cycle of deployment model segments the deployment cycle into five stages across the months of deployment (Pincus et al., 2001). Pre-deployment occurs when service members receive orders to deploy and prepare for departure, deployment refers to the first month after departure when individuals adjust to their new circumstances, sustainment includes the duration of the time apart when people settle into new routines, redeployment involves preparation for homecoming during the month before the service member’s return, and post-deployment focuses on the 6 months following reunion when military couples adjust to living together again.
The model also identifies specific issues military couples face at each stage (Pincus et al., 2001). With regard to the post-deployment transition, the model theorizes that homecoming begins with a honeymoon period stemming from the joy of being reunited. Tension emerges over time as service members seek to reassert their role in the family and at-home partners adjust to less autonomy. Other challenges include navigating sexual intimacy after the time apart and developing a new household routine that is functional for both partners. Beyond highlighting these key issues, however, the model has little to say about the timing or sequence of relationship changes during the transition from deployment to reintegration.
Research on the post-deployment transition has identified relationship changes consistent with the emotional cycle of deployment model. Such findings emphasize romanticized closeness (Karakurt et al., 2013), role strain (Karakurt et al., 2013; Knobloch, Basinger, et al., 2016), and shifts in independence (Karakurt et al., 2013; Knobloch & Theiss, 2012). Other results compatible with the model highlight changes in sexual intimacy (Baptist et al., 2011; Knobloch & Theiss, 2012) and domestic routines (Baptist et al., 2011; Knobloch, Basinger, et al., 2016) as relevant to the transition from deployment to reintegration. Although providing important descriptions of the relationship changes that occur after homecoming, these studies have been conducted with relatively small sample sizes and long retrospective recall windows.
Goals of the current study
The emotional cycle of deployment model offers valuable insight into the post-deployment period, but key questions remain. First, the set of relationship changes highlighted by the model is far from exhaustive. Investigations of the transition from deployment to reintegration have documented a variety of relationship changes not identified by the model, including relationship growth, enhanced or impaired communication, fluctuations in support patterns, and personality shifts (Baptist et al., 2011; Karakurt et al., 2013; Knobloch, Basinger, et al., 2016; Knobloch & Theiss, 2012). Our study advances beyond previous work by assessing relationship changes in a large sample of military couples for 7 months following reunion, using monthly intervals of data collection to shorten the recall window. Our first two research questions investigate the content and trajectories of relationship changes over time:
Other unanswered questions involve the valence of people’s experiences. Consistent with the logic of the emotional cycle of deployment model (Pincus et al., 2001), cross-sectional studies reveal a constellation of positive and negative dynamics upon reunion (Sahlstein et al., 2009; Sahlstein Parcell & Maguire, 2014). For example, some military couples report that their relationship grew stronger or that they value the relationship more than they did before deployment (Baptist et al., 2011; Knobloch, Basinger, et al., 2016; Knobloch & Theiss, 2012). On the other hand, nearly 20% of returning service members report moderate to severe problems with family reintegration after homecoming (Balderrama-Durbin et al., 2015). Nearly 30% of Army wives report decreasing levels of marital satisfaction during the post-deployment transition, and more than 12% of Army wives report a turbulent trajectory marked by several upturns and downturns of marital satisfaction (Sahlstein Parcell & Maguire, 2014). Another complexity is that relationship changes can have both positive and negative ramifications over time (Sahlstein et al., 2009; Sahlstein Parcell & Maguire, 2014). Such findings led Sahlstein Parcell and Maguire (2014) to call for future research to “carefully consider the postdeployment period by teasing out the differences between experiences [and trajectories] that trend more positive versus negative” (p. 142). Our final research questions evaluate these issues:
Method
We solicited online survey data from a nationwide sample of U.S. military couples (see also Knobloch, Knobloch-Fedders, & Yorgason, 2018, 2019). Recruitment occurred via (a) postings to social media channels frequented by military families; (b) advertising in military installation newspapers; and (c) publicizing the study to military family professionals. Eligibility criteria required (a) partners to have separate email accounts, (b) one or both partners to have recently returned home from deployment, and (c) both partners to complete the Wave 1 questionnaire within the first 7 days after homecoming.
Participants
Participants were 555 military couples (N = 1,110 individuals; n = 554 men and 556 women) who lived in 44 U.S. states, the District of Columbia, and Guam. Most couples were heterosexual (99%), married (95%), involved in their first marriage (77%), parents (71%), and living together upon homecoming (96%). Individuals ranged in age from 19 years to 59 years (M = 31.18 years, SD = 6.39 years), and their romantic relationship averaged 8.43 years in length (SD = 5.40 years).
The sample was 81% White, 10% Latinx, 4% African American, 3% Asian or Pacific Islander, and 2% American Indian or Alaskan Native. Level of education included some high school (1%), high school graduate (13%), some college (31%), associate’s degree (15%), bachelor’s degree (28%), or advanced graduate degree (12%). Modal household income in U.S. dollars was $21,000 to $40,000 (23%), $41,000 to $60,000 (32%), or $61,000 to $80,000 (18%).
Deployed military personnel were returning from a tour averaging 7.71 months (SD = 2.31 months) in length. They were part of the U.S. Army (40%), Navy, (21%), Marines (18%), Air Force (10%), Army National Guard (8%), Air National Guard (2%), and Coast Guard (1%). Most were male (n = 547; 99%). Approximately 60% deployed as part of a combat mission, but others completed assignments involving peacekeeping (17%), training (15%), relief (3%), or unreported (5%). Approximately 30% of service members had returned from their first deployment, but others had completed one (24%), two (17%), three (13%), four (8%), or five or more (8%) previous deployments. Most at-home partners were female (n = 548; 99%). The majority of at-home partners were civilians (88%), but others were currently serving (5%) or had previously served (7%) in the military themselves.
Procedures
After both partners replied to an email requesting their consent, we emailed each individual a link to the Wave 1 questionnaire, along with a unique login name and temporary password. Participants selected a permanent password for the remainder of the study. We sent reminder emails on the fourth and sixth days after homecoming to individuals who had not yet responded. After the Wave 1 logins expired on the seventh day, we eliminated 32 military couples who did not complete the Wave 1 questionnaire by the 7-day deadline.
The remaining 555 couples continued in the study at monthly intervals for the next 7 months. We used the emotional cycle of deployment model to inform both the interval and the length of our data collection. We chose monthly intervals because it is the metric of time used by the model, and we conducted assessments over 7 months to extend beyond the 6-month window the model assigns to the transition (Pincus et al., 2001).
Each person received an email with a link to the next survey on the anniversary of their reunion date, along with reminder emails on the fourth and sixth days. The logins expired on the seventh day. Upon completion of each questionnaire, individuals received a U.S. $15 e-gift card from a national retailer. Participants who completed all eight waves received a bonus U.S. $50 e-gift card.
Measures
Demographics
For descriptive purposes, participants at Wave 1 completed demographic measures of individual characteristics (e.g., sex, race, age, and education), relationship attributes (e.g., household income, marital status), and military variables (e.g., deployment mission, prior deployment experience).
Relationship satisfaction
We measured relationship satisfaction as a covariate at Wave 1 using the Couple Satisfaction Index (Funk & Rogge, 2007). Participants’ responses to the 4 items were summed (M = 17.20, SD = 3.32, range = 2.00–21.00, α = .83).
Relationship changes
Waves 2 through 8 began with an open-ended item that asked, “Has your romantic relationship changed in the past month? If so, list up to three ways your romantic relationship has changed.” 1 Accordingly, participants’ responses at Wave 2 indexed relationship changes from Wave 1 to Wave 2, their responses at Wave 3 indexed relationship changes from Wave 2 to Wave 3, and so on. Responses to the item generated a total of seven waves of data from Wave 2 through Wave 8.
Data analysis
Content analysis
We derived categories from the responses using content analytic methods (following Neuendorf, 2002). First, we read through the responses several times to familiarize ourselves with the data. Second, we employed open coding and axial coding to identify prominent categories (Strauss & Corbin, 1998). During the open coding stage, we examined the data line-by-line to identify and label key concepts. During the axial coding stage, we distinguished among the key concepts and collapsed categories with similar meanings.
Next, we created a codebook that described the categories and provided examples of positive, negative, and neutral changes within each category (see the online supplement). We used contextual information from the tone and content of the response to judge valence. For example, positive changes included, “I think we appreciate each other more now” (at-home partner in Wave 2), “We have gotten closer emotionally,” (at-home partner in Wave 6), and “We communicate better” (returning service member in Wave 6). Negatively valenced comments included “More hostile and tense” (at-home partner in Wave 3), “Mood swings are higher” (returning service member in Wave 7), and “I despise him” (at-home partner in Wave 7). Neutral responses either (a) contained both positive and negative elements, such as “It’s been up and down” (returning service member in Wave 3) and “Good one day, not so good the next” (at-home partner in Wave 2), or (b) provided no contextual details for judging valence. Examples of the latter included “Moved to a new house” (at-home partner in Wave 2) or “Left my job” (at-home partner in Wave 7). 2
We unitized responses conveying multiple ideas into thematic units containing one idea. A thematic unit is a unit of analysis suitable for open-ended text that ranges from one clause to multiple sentences in length (Krippendorff, 2004). Dividing the open-ended data in this way permitted each thematic unit to be coded into a single category. For example, at Wave 5, a returning service member wrote, “Our communication skills have improved and we seem to get along a little better than the previous few months.” Because this statement conveys separate ideas, it was divided into two thematic units (“Our communication skills have improved”/“and we seem to get along a little better than the previous few months”). The third and fourth authors unitized the responses into thematic units, and the first author reviewed their work for accuracy.
Coder training
We trained eight independent coders to classify the data by category and valence. The coders began by completing 10 hr of training, which included learning the codebook, discussing exemplar codes, coding practice data, and receiving feedback. Then, we divided the eight coders into two teams of four. Both teams coded one wave of data (14%) in common, and once reliability was established, each team coded 43% of the remaining data independently.
Coding procedures
The coders were blind to the research questions, the wave of data collection, and the participant’s gender and role as a returning service member or at-home partner. Coders worked independently to assign each thematic unit a category and a valence. To prevent coder drift, they met weekly as a group to review their work and receive feedback. Disagreements were resolved by selecting the majority choice. Ties (which occurred for 7% of the category judgments and 6% of the valence judgments) were resolved by the third and fourth authors, who managed the coding process under the supervision of the first author.
Coding reliability
We calculated reliability using Krippendorff’s α (Krippendorff, 2004). Krippendorff’s α is appropriate for any number of coders and all levels of measurement, is independent from the distribution of categories in the data, and has a computable sampling distribution (Hayes & Krippendorff, 2007). Krippendorff’s α ranges from 0.00 to 1.00, with marginal reliability > .67 and satisfactory reliability > .80 (Krippendorff, 2004). In this study, Krippendorff’s α = .85 for category judgments and .82 for valence judgments.
Results
In response to the open-ended item, 977 participants (468 returning service members and 509 at-home partners) wrote a total of 7,387 thematic units across the seven waves of data (M = 7.56 thematic units, SD = 5.54 thematic units, range = 1–41 thematic units). At-home partners (M = 8.42, SD = 5.99) wrote more thematic units than returning service members (M = 6.62, SD = 4.83), t(975) = 5.15, p < .001. A total of 6,270 thematic units (84.9%) described changes in the relationship, 1,104 (14.9%) indicated no changes had occurred, and 13 (0.2%) were uncodable fragments.
Changes to the relationship (RQ1)
RQ1 asked about the relationship changes reported by military couples during the transition (see Table 1). The content analysis generated 10 categories of substantive responses (see the online supplement for information regarding the frequency of each category by wave).
Frequency of relationship changes by valence.
Note. N = 7,387 thematic units across seven waves of data.
Emotional intimacy, closeness, and support
Responses in this category (n = 1,550) described changes in emotional intimacy, closeness, and support within the relationship. Themes included connecting or distancing; expressing feelings; listening; and conveying or withdrawing support, love, or understanding. With respect to valence, 68.9% of the units were coded as positive, 25.0% were coded as negative, and 6.1% were coded as neutral. Positively valenced responses included “We’re telling each other that we love each other more” (at-home partner in Wave 2), “We are understanding each other better than before” (returning service member in Wave 5), and “Being kinder to each other” (at-home partner in Wave 7). Negatively valenced comments included “We don’t seem to communicate like we used to” (returning service member in Wave 4), “My wife is having issues sharing her feelings with me” (returning service member in Wave 2), and “It feels more distant” (at-home partner in Wave 2). An example of a neutrally valenced comment was “Feels like in some ways we have gotten closer, but in some ways we have grown apart” (at-home partner in Wave 3).
Sexual intimacy and romance
These comments depicted changes in sexual or physical intimacy, affection, romance, or passion (n = 1,137). Positive valence was assigned to 42.5% of the units, including “More intimate sex” (at-home partner in Wave 8), “Much more romantic” (at-home partner in Wave 8), and “Wife makes more of an attempt to kiss daily” (returning service member in Wave 6). Negatively valenced comments (49.7%) included “Lost passion,” (service member in Wave 2), “Not much sex anymore” (at-home partner, Wave 3), and “My wife is not as affectionate as she used to be” (returning service member in Wave 3). Neutral changes (7.8%) included “Sex drive has wavered between low and high” (at-home partner in Wave 3) and “I felt that he was less affectionate towards me, but that has been improving” (at-home partner in Wave 2).
Spending time together
Responses in this category (n = 732) described partners spending quality time together or having problems doing so. Themes included going on dates or vacations; being separated due to travel, work, or additional deployments; and not spending time together because of busy schedules. In terms of valence, 53.6% of the units were coded as positive, 41.1% as negative, and 5.3% as neutral. Positive changes included “We don’t want to leave each other’s side” (at-home partner in Wave 2), “Have spent more time alone together, date nights” (returning service member in Wave 2), and “We’ve found a new common interest” (at-home partner in Wave 7). Negative comments included “Less time together” (returning service member in Wave 5), “I have been away from my spouse for military training” (returning service member in Wave 8), and “We have not been able to hang out together as much as we would like” (at-home partner in Wave 8).
Appraisals of the relationship
These comments (n = 677) described or evaluated the relationship, including observations that the relationship was different, had become stronger or weaker, or that efforts were being made to improve it. Positive valence was assigned to 71.0% of the units, including “I think we’re stronger than ever” (at-home partner in Wave 5), “Relationship is overall better than prior to deployment” (returning service member in Wave 2), and “Made some progress with the help of proactive counseling” (returning service member in Wave 5). Negatively valenced comments (18.0%) included “Become more hopeless since the last survey” (returning service member in Wave 8), “There is not a lot of effort being put into it by either of us” (at-home partner in Wave 3), and “Unhappy” (at-home partner in Wave 2). Responses judged as neutral in valence (11.0%) included “It has many ups and downs. Really high highs and low lows” (at-home partner in Wave 6) and “It’s OK” (returning service member in Wave 8).
Life changes
Statements in this category (n = 621) reported changes in major life domains, including work, home, school, physical health, mental health, spirituality, and finances. The valence was 11.3% positive, 43.3% negative, and 45.4% neutral. Positive changes included “We have begun to do devotions at night and afterward pray” (returning service member in Wave 4), “I started exercising again. I feel better about myself” (at-home partner in Wave 7), and “Drink alcohol less” (returning service member in Wave 2). Negative comments included “More financial stress” (at-home partner in Wave 5), “Wife had surgery” (returning service member in Wave 7), and “Work interfering more” (at-home partner in Wave 3). Neutral responses included “We got our orders to Hawaii” (at-home partner in Wave 6) and “Started a second job” (at-home partner in Wave 7).
Readjustment to daily life
Responses (n = 586) described acclimating to everyday schedules and patterns, including getting used to living with each other again, resuming daily activities, establishing routines, and reallocating household duties. Positive valence was assigned to 44.0% of the units, including “We feel more adjusted with each other. Developed a routine” (at-home partner in Wave 2), “Teamwork is better” (at-home partner in Wave 3), and “Finally settled in” (returning service member in Wave 3). Negative valence comprised 36.0% of the comments, including “He is still having challenges adjusting to home life” (at-home partner in Wave 3), “The honeymoon phase of being back is gone” (returning service member in Wave 4), and “Reintegration is proving to be a little difficult this time” (at-home partner in Wave 2). Neutral valence included 20.0% of the responses, including “We have drifted back into pre-deployment mode” (at-home partner in Wave 7) and “My expectations have changed” (at-home partner in Wave 8).
Conflict
Comments in this category (n = 452) explicitly referred to conflict and its resolution, such as addressing tensions in the relationship, arguing, fighting, resolving disagreements, and apologizing. With respect to valence, 21.9% of the units were coded as positive, 75.0% as negative, and 3.1% as neutral. Positive exemplars included “Getting along better” (returning service member in Wave 3), “Quicker conflict resolution” (at-home partner in Wave 2), and “It is easier to agree on things” (returning service member in Wave 2). Negative comments included “We fight more than we ever have before” (at-home partner in Wave 3), “Partner had angry outburst” (at-home partner in Wave 5), and “Get on each other’s nerves more” (returning service member in Wave 2). An example of a neutral response was “Normal amount of arguments, nothing major” (at-home partner in Wave 6).
Family changes
Responses (n = 359) alluded to changes within the family, including pregnancy, infertility, parenting, or the well-being of children or extended family members. Valence was 54.3% positive, 29.8% negative, and 15.9% neutral. Positive changes included “We are excited about the new baby” (at-home partner in Wave 8), “Attempting to get pregnant” (at-home partner in Wave 8), and “Getting closer as a family” (returning service member in Wave 3). Negative responses included “Our daughter’s behavior problems have put a strain on our relationship” (at-home partner in Wave 8), “Subversive mother-in-law” (returning service member in Wave 5), and “Miscarriage” (at-home partner in Wave 5). Neutral comments included “Back to reality of parenting together” (at-home partner in Wave 3), and “Home with kids during military school” (at-home partner in Wave 7).
Commitment to the relationship
This category contained responses (n = 156) reporting changes in relationship status, investment, or faithfulness. Themes included getting engaged or married, separating, divorcing, or confronting issues related to infidelity. In terms of valence, 37.8% of comments were positive, 57.1% negative, and 5.1% neutral. Positive responses included “We got engaged” (at-home partner in Wave 3), “Husband asked to renew our wedding vows” (at-home partner in Wave 8), and “We decided to restart our relationship” (at-home partner in Wave 8). Negative statements included “Wife moved out” (returning service member in Wave 6), “He asked for a divorce” (at-home partner in Wave 6), and “I think my wife is cheating” (returning service member in Wave 5). Neutral comments included “We have recently opened up our marriage” (returning service member in Wave 8) and “We have been talking much more about the direction of our relationship” (at-home partner in Wave 3).
No change
These responses (n = 1,104) indicated that no change had occurred in the relationship. Over 99% of these comments were assigned a neutral valence, including “It hasn’t changed much” (at-home partner in Wave 2), “It’s been about the same” (returning service member in Wave 4), and “Nothing has changed” (returning service member in Wave 6).
Trajectories of relationship changes (RQ2)
A second research question asked about the trajectories of relationship changes over time (RQ2a) and differences in the trajectories between returning service members and at-home partners (RQ2b). We conducted dyadic growth curve analyses (Kenny et al., 2006; Ledermann & Kenny, 2017; Singer & Willett, 2003) to model the frequency of relationship changes across the seven waves. We computed separate growth curve models predicting each of the 10 substantive categories as the dependent variable. The models contained five predictors, all treated as fixed effects: (a) the linear effect of time across the study, with the intercept (Time 0) designated as the first time we measured relationship changes at Wave 2; (b) the quadratic effect of time represented as time-squared (time2); (c) role (1 = returning service member, −1 = at-home partner); and (d) two cross-level interactions (Role × Time and Role × Time2). Based on prior research (Balderrama-Durbin et al., 2015; Meadows et al., 2016), we also controlled for four covariates as fixed effects: (a) relationship satisfaction; (b) marital status (0 = married, 1 = dating, engaged, or civil union); (c) deployment mission (0 = combat, 1 = peacekeeping, relief, training, or other); and (d) first deployment (0 = no, 1 = yes). 3
We treated the intercept as a random effect and time as a repeated effect across the seven waves of data from Wave 2 to Wave 8. The estimation approach was restricted maximum likelihood, the covariance structure for the random effect was variance components, and the calculation method was Type III sum of squares. All models were computed using SPSS v.26.
Results for RQ2 are summarized in Table 2. 4 With respect to the covariates, higher relationship satisfaction at Wave 1 corresponded with fewer changes in emotional intimacy, sexual intimacy, readjustment to daily life, conflict, and commitment at Wave 2 (the first time relationship changes were assessed). Higher relationship satisfaction at reunion also predicted more frequent reports of no change at Wave 2. Individuals who were married reported more changes in sexual intimacy, and fewer changes in commitment, at Wave 2. The two military covariates (deployment mission, first deployment) were not statistically significant predictors in any of the models.
Results of dyadic growth curve analyses predicting frequency of relationship changes over time.
Note. N = 7,387 thematic units across seven waves of data.
*p < .05; **p < .01; ***p < .001.
Trajectories over time (RQ2a)
As indicated by a negative linear effect of time, the frequency of changes reported in emotional intimacy, sexual intimacy, readjustment to daily life, and conflict decreased across the transition. For both sexual intimacy and readjustment to daily life, a positive quadratic trajectory also emerged such that the decline in the frequency of these categories leveled off over time.
The domain of life changes increased across the transition via a positive linear effect of time. Participants’ reports of no change increased as well, with a negative quadratic trajectory indicating a plateau in this increase over time.
Differences in the trajectories (RQ2b)
With respect to the intercepts, returning service members reported fewer changes in emotional intimacy, spending time together, and conflict at Wave 2 than at-home partners. Conversely, at-home partners reported fewer changes in appraisals of the relationship and fewer reports of no change at Wave 2 than returning service members.
Most of the slopes were not different for returning service members versus at-home partners, indicating similar rates of change over time. The only exception was that, compared to returning service members, at-home partners reported a steeper increase in life changes over time.
Valence of relationship changes (RQ3)
RQ3 examined the valence of relationship changes. Of the 7,387 total comments, 3,109 (42.1%) were assigned a positive valence; 2,392 (32.4%) were assigned a negative valence; and 1,886 (25.5%) were assigned a neutral valence (see Table 1).
Trajectories of the valence of relationship changes (RQ4)
RQ4 investigated the valence of relationship changes over time (RQ4a) and evaluated differences in valence trajectories between returning service members versus at-home partners (RQ4b). To examine RQ4, we repeated the growth curve models for RQ2 but substituted the frequency of positive, negative, or neutral changes from Wave 2 to Wave 8 as the dependent variable.
Findings for RQ4 appear in Table 3. 5 In terms of the covariates, people who were more satisfied with their relationship at reunion reported fewer negative changes at Wave 2, and individuals who were not married at reunion reported fewer neutral changes at Wave 2. Neither of the military covariates explained a statistically significant amount of variance in any of the models.
Results of dyadic growth curve analyses predicting frequency of positive, negative, and neutral changes over time.
Note. N = 7,387 thematic units across seven waves of data.
*p < .05; **p < .01; ***p < .001.
Positive changes
With respect to the intercept, participants described an average of 1.118 positive changes at Wave 2. With respect to the slopes, negative linear and positive quadratic effects of time were evident such that the number of positive changes people reported declined across the transition and leveled off over time (RQ4a). Returning service members reported fewer positive changes at Wave 2 than at-home partners (RQ4b), but no differences were apparent in their slopes, indicating similar rates of decline over time.
Negative changes
At Wave 2, participants reported an average of .999 negative changes. No linear or quadratic effects of time were apparent (RQ4a). Although returning service members reported fewer negative changes than at-home partners at Wave 2, their slopes did not differ, suggesting similar trajectories across the transition (RQ4b).
Neutral changes
An average of .178 neutral changes were reported by participants at Wave 2. A positive linear effect of time emerged such that the frequency of neutral changes increased across the transition (RQ4a). Compared to returning service members, at-home partners reported fewer neutral changes at Wave 2 (RQ4b), but no differences emerged in the slopes representing change over time. 6
Discussion
The transition from deployment to reintegration is far more than a picture-perfect welcome home ceremony (e.g., Howard & Prividera, 2015). Incorporating a returning service member back into family life is a lengthy process that can be more challenging than deployment itself (Huebner et al., 2007; Mmari et al., 2009; Pincus et al., 2001). Whereas the emotional cycle of deployment model (Pincus et al., 2001) emphasizes the changes that military couples experience between stages, it has less to say about changes within stages. Our study addressed that gap by investigating the relationship changes military couples experience during the months following homecoming.
Our research design advanced beyond prior work in several ways. First, to gain a more detailed understanding of the transition, we asked 1,110 returning service members and at-home partners to describe their experiences in their own words. Second, to facilitate the generalizability of our results, we enrolled a nationwide sample from all military branches and regions of the country. Third, to address calls to track relationship changes sequentially (e.g., Balderrama-Durbin et al., 2015; Knobloch & Theiss, 2012), we utilized a Wave 1 assessment of demographic information at reunion, followed by seven monthly assessments of relationship changes beginning 30 days after homecoming (7,387 total observations). In what follows, we describe the implications of our data, the limitations of our study, and directions for future work.
Relationship changes during the transition
We started by identifying the relationship changes returning service members and at-home partners reported during the transition (RQ1). In order of diminishing frequency, the relationship changes military couples described were (a) emotional intimacy, closeness, and support; (b) sexual intimacy and romance; (c) spending time together; (d) appraisals of the relationship; (e) life changes; (f) readjustment to daily life; (g) conflict; (h) family changes; and (i) commitment. Whereas some of these domains echo the emotional cycle of deployment model (e.g., sexual intimacy and readjustment to daily life; Pincus et al., 2001) and others triangulate prior cross-sectional findings (e.g., emotional intimacy, closeness, and support; Karakurt et al., 2013; Knobloch & Theiss, 2012), our full list offers a more comprehensive view than previously articulated by theory or research.
The data for RQ1 offer an important reminder that returning service members and at-home partners experience normative changes alongside reunion-specific shifts (e.g., Sahlstein Parcell & Maguire, 2014). Most of the relationship changes reported by military couples were directly linked to reunion after separation, but others involved life changes and family changes tied to the developmental life course more generally. Although changes in domains such as education, finances, health, and pregnancy are not unique to military couples at homecoming, couples who accumulate family life course changes on top of deployment-related changes may deplete their coping resources more quickly (e.g., Collins et al., 2017). Therefore, targeted support services may be particularly helpful to military couples who experience both normative and reunion-specific changes during the post-deployment transition.
Our longitudinal data allowed us to examine sequences of relationship changes over time (RQ2a). After controlling for relationship satisfaction, marital status, deployment mission, and prior deployment experience, people’s reports of changes in (a) emotional intimacy, closeness, and support; (b) conflict; (c) sexual intimacy and romance; and (d) readjustment to daily life declined across the transition, with the trajectories for the latter two categories leveling off over time. Conversely, reports of life changes and no change escalated across the transition, with the trajectory for the latter category plateauing over time. These descriptive findings imply that relationship changes happening between partners (i.e., issues of intimacy, conflict, and daily routines) are most prominent during the early months of the post-deployment transition, and changes happening to couples (i.e., life changes) are most prominent later on as the transition unfolds.
Beyond their descriptive value, the findings for RQ2a have relevance for both theory and intervention with military couples. For example, although the emotional cycle of deployment model provides a useful heuristic for understanding the issues returning service members and at-home partners encounter at each stage (Pincus et al., 2001), our results illuminate ways to expand the model to include the sequence of relationship changes that may occur within a particular stage (such as reintegration). From a clinical standpoint, because military couples reported some categories of relationship changes more frequently at certain points during the transition, our findings imply that sequential programming tailored to the progression of the transition over time may be effective. Such programming could leverage the open-ended comments in our data describing relationship changes in participants’ own words, because military couples express a preference for insight, guidance, and support from others with similar experiences (Rossetto, 2015).
Valence of relationship changes during the transition
RQ3 asked about the valence of relationship changes military couples experience upon homecoming following deployment. Among the 7,387 responses we gathered, 42.1% described positive changes, 32.4% portrayed negative changes, and 25.5% depicted neutral changes. Because our study is the first to prospectively quantify the valence of relationship changes during the months after reunion, it provides a more nuanced understanding of the tenor of the post-deployment transition than previously available. With almost a third of the responses describing negatively valenced changes, the transition appears less idyllic than the happily-ever-after storyline depicted in cultural narratives (e.g., Howard & Prividera, 2015). Conversely, with more than two-fifths of the responses describing positively valenced changes, the transition appears less gloomy than implied by studies that focus on the challenges of reunion after deployment (e.g., Knobloch & Theiss, 2018). Our quantitative findings triangulate qualitative work suggesting that the post-deployment transition is much more complex than pure joy or acute distress (Sahlstein et al., 2009; Sahlstein Parcell & Maguire, 2014).
We also tracked the trajectories of positive, negative, and neutral relationship changes over time (RQ4a). The emotional cycle of deployment model emphasizes the importance of attending to relationship changes during deployment and reunion (Pincus et al., 2001), but the model is less precise about the tenor of the transition beyond suggesting an initial honeymoon period replaced by the emergence of daily struggles. Our data show that the frequency of (a) positively valenced relationship changes declined across the transition but leveled off over time, (b) negatively valenced relationship changes remained stable across the transition, and (c) neutrally valenced relationship changes increased over time. Moreover, individuals who were more satisfied with their relationship at homecoming reported fewer negative changes during the first month after reunion.
The findings for RQ4a contribute to both theory and practice. With respect to theory, the frequency of positively valenced relationship changes declining and leveling off over time offers empirical evidence bolstering the emotional cycle of deployment model’s suggestion of a honeymoon phase during the early days of the transition (Pincus et al., 2001). On the other hand, the fact that the frequency of negatively valenced relationship changes remained stable over time suggests that the strains and stressors of reintegration do not emerge gradually after homecoming but steadily challenge military couples throughout the post-deployment transition.
The emotional cycle of deployment model also could be enriched by attending to relationship satisfaction at homecoming. Our data imply that individuals who are more satisfied with their relationship upon reunion may be buffered from negatively valenced relationship changes (see Theiss & Knobloch, 2014), or alternatively, less likely to view relationship changes in a negative light (see Fletcher, 2015). Extending these findings to practice suggests that military couples may benefit from prevention and intervention services designed to (a) prepare them for a decline in positively valenced relationship changes over time and (b) promote relationship satisfaction at reunion.
Finally, we compared the relationship changes reported by returning service members and at-home partners. Several differences emerged in our first assessment of relationship changes at Wave 2. With respect to content (RQ2b), at-home partners wrote more comments describing relationship changes in emotional intimacy, spending time together, and conflict. Returning service members wrote more responses describing appraisals of the relationship and stating that no change had occurred. With respect to valence (RQ4b), at-home partners reported more positive and negative changes, and fewer neutral changes, than returning service members. Because at-home partners functioned on their own during the deployment, directed the household independently, and often cared for children as a single parent (Rossetto, 2015; Sahlstein et al., 2009), they may be more sensitive to both the beneficial and detrimental relationship changes that occur right after homecoming. Despite these differences at Wave 2, the longitudinal trajectories of the content and valence of relationship changes were similar for returning service members and at-home partners over time, suggesting comparable sequences throughout the rest of the transition. These findings imply that the weeks immediately following reunion, rather than the duration of the transition, may be particularly ripe for misunderstanding between partners.
Our project as a whole, and particularly the findings for RQ2b and RQ4b, illuminate the benefits of combining self-report and observer data. Some studies have asked returning service members and at-home partners to describe their experiences at one point in time (e.g., Knobloch, Basinger, et al., 2016; Knobloch & Theiss, 2012) or interviewed military couples multiple times across the transition (Karakurt et al., 2013). Other work has solicited people’s closed-ended ratings of relationship functioning during the months after homecoming (e.g., Knobloch, McAninch, et al., 2016). In comparison, our data integrating people’s descriptive accounts of relationship changes with the judgments of independent coders offer a multi-perspective view of the transition as it unfolds over time.
Limitations and directions for future work
Several limitations of our study are important to note. First, our sample is not representative of the diversity of U.S. military couples. Particularly overrepresented in our sample are White individuals, male returning service members, heterosexual couples, and Army couples. Therefore, our results may not generalize to people from other racial or ethnic groups, female returning service members, same-sex dyads, or couples affiliated with other military branches. Second, our findings are subject to the drawbacks inherent in self-report methods. Fortunately, given the wide range of positive, negative, and neutral changes reported by participants, as well as the detailed and sensitive nature of many responses, our data do not appear to be unduly influenced by social desirability or careless responding. Third, because our first assessment of relationship changes occurred 1 month after reunion at Wave 2, we lack information about the changes military couples experienced during deployment. Finally, we judged the valence of relationship changes using the contextual information in people’s comments, but having participants themselves classify the valence of their relationship changes would have triangulated our coding data.
Our study highlights two avenues for future work. First, our results pave the way for a larger scale evaluation of relationship changes due to deployment. Such a study could follow military couples through the entire deployment cycle, with a baseline assessment at pre-deployment and long-term follow-up assessments after homecoming (e.g., Miller et al., 2018). Second, military children also are affected by changes to the family system across the deployment cycle (Houston et al., 2009; Huebner et al., 2007; Knobloch et al., 2015). Investigators could widen the scope of our focus on military couples to include the changes experienced by military parents and children (e.g., Lester et al., 2016; Strong & Lee, 2017). These directions for future work could advance theory, research, policy, and clinical intervention designed to help military couples successfully navigate the post-deployment transition.
Supplemental material
supplemental_material - Relationship changes of military couples during reintegration: A longitudinal analysis
supplemental_material for Relationship changes of military couples during reintegration: A longitudinal analysis by Lynne M. Knobloch-Fedders, Leanne K. Knobloch, Samantha Scott and Hannah Fiore in Journal of Social and Personal Relationships
Footnotes
Acknowledgments
The authors thank Jeremy B. Yorgason for his feedback on an earlier version of this article, and Kaitlyn Bellingar, Karl Briedrick, Chrishane Cunningham, Daphne Liu, Alexandra Maynard, Alexis Meade, Kathleen Pell-King, and Jacqueline Wong for their assistance in coding.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the Congressionally Directed Medical Research Programs through the Military Operational Medicine Research Program (Award W81XWH-14-2-0131). The U.S. Army Medical Research Acquisition Activity, 820 Chandler Street, Fort Detrick MD 21702-5014, was the awarding and administering acquisition office. Interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the U.S. Department of Defense.
Open research statement
As part of IARR’s encouragement of open research practices, the author(s) have provided the following information: This research was pre-registered. The aspects of the research that were pre-registered were the study design and methods. The registration was submitted via a funded grant application to the Military Operational Medicine Research Program of the U.S. Army Medical Research and Materiel Command (Award W81XWH-14-2-0131). The data used in the research are not available. The materials used in the research are available. The materials can be obtained by emailing:
Supplemental material
Supplemental material for this article is available online.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
