Content uploaded by Brian D. Doss
Author content
All content in this area was uploaded by Brian D. Doss on Feb 15, 2021
Content may be subject to copyright.
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 1
©American Psychological Association, 2020. This paper is not the copy of record and may not exactly
replicate the authoritative document published in the APA journal. Please do not copy or cite without
author's permission. The final article is available at DOI: 10.1037/ccp0000606
Maintenance of Relationship and Individual Functioning Gains
Following Online Relationship Programs for Low-Income Couples
McKenzie K. Roddy, M.S.1
Kayla Knopp, Ph.D.2
Emily Georgia Salivar, Ph.D.3
Brian D. Doss, Ph.D.1
1University of Miami, 2VA San Diego Healthcare System, 3Nova Southeastern University
Authors’ Note:
Dr. Brian Doss is a co-inventor of the intellectual property used in this study and an equity owner
in OurRelationship LCC. McKenzie Roddy completed this work at the University of Miami but
has since moved to Baylor College of Medicine.
This research was supported by the U.S. Department of Health and Human Services,
Administration for Children and Families, Grant Number 90FM0063 to the last author. This
project was also made possible by Grant Number 90PR0008 to first author from the Office of
Planning, Research and Evaluation, Administration for Children and Families, U.S. Department
of Health and Human Services. Its contents are solely the responsibility of the authors and do not
necessarily represent the official views of the Office of Planning, Research and Evaluation, the
Administration for Children and Families, or the U.S. Department of Health and Human
Services. Additionally, this work was supported by the Dean’s Summer Fellowship from the
University of Miami to the first author.
We would like to thank Clark Brinson and Samantha Joseph who served as Project Coordinators
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 2
for this study.
Correspondence concerning this article should be addressed to McKenzie K. Roddy, Department
of Psychiatry and Behavioral Sciences, Baylor College of Medicine, 1977 Butler Boulevard,
Houston, TX, 77030. E-mail: mkr19@miami.edu
Abstract
Initial evidence suggests that gains in relationship functioning from brief, web-based programs
are maintained through one year following the intervention; however, whether these results
generalize to a low-income sample is unclear. Furthermore, previous research from in-person
couple therapy suggests there may be different shapes of maintenance slopes for behavioral
versus acceptance-based techniques. This study contacted 668 individuals who enrolled in online
behavioral (ePREP) or acceptance-based (OurRelationship) programs one year following
completion of the program. Multilevel modeling was used to examine linear and quadratic rates
of change in the year following the online intervention as well as total amount of change from
pretreatment to 12-month follow-up for both relationship and individual functioning. The
majority of couples who responded continued to be in a relationship with the same partner
(68.3%). Examinations of relationship functioning indicated couples in both programs
maintained their gains over follow-up (i.e., no significant linear or quadratic changes), with
medium-to-large within-group effect sizes from pre- to one-year follow-up. There were no
significant differences in relationship outcomes between OurRelationship and ePREP. Similarly,
examinations of individual functioning outcomes indicated couples maintained their gains over
follow-up or continued to improve. In total, couples experienced small-to-medium within-group
effect sizes from pretreatment to one-year follow-up, with larger effects for individuals who were
initially distressed. These results suggest that online programs create lasting change for low-
income couples in relationship and individual functioning, with minimal differences between
behavioral and acceptance-based orientations.
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 3
Keywords: couples / romantic relationships, longitudinal data, web-based, self-help
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 4
Maintenance of Relationship and Individual Functioning Gains
Following Online Relationship Programs for Low-Income Couples
Distress in relationships adversely impacts many domains of life, including mental and
physical health, family functioning, and work (Foran, Whisman, & Beach, 2015). Relationship
distress is common, affecting approximately one-fifth of newlyweds (Beach, Fincham, Amir, &
Leonard, 2005) and one-third of couples overall (Whisman, Beach, & Snyder, 2008).
Relationship distress is especially pronounced for low-income couples, who are at greater risk of
adverse relationship outcomes including divorce (Bramlett & Mosher, 2002), intimate partner
violence (Cunradi, Caetano, & Schafer, 2002), and low satisfaction (Lundquist et al., 2014).
Both relationship education and couple therapy improve relationship distress (Hawkins,
Blanchard, Baldwin, & Fawcett, 2008; Lebow, Chambers, Christensen, & Johnson, 2012),
though unfortunately remain markedly underutilized, particularly among lower-income couples
(Doss, 2014; Halford, O'Donnell, Lizzio, & Wilson, 2006). This underutilization may be due in
part to the logistical barriers couples encounter in attending, including difficulties around
scheduling, childcare, transportation, and finances. Unfortunately, these barriers are exacerbated
for low-income couples (Gaubert, Gubits, Principe Alderson, & Knox, 2012; Williamson,
Karney, & Bradbury, 2019) because partners are often working more than one job (sometimes on
different schedules), they utilize public or unreliable transportation, and may need to rely on
family members for childcare.
Online programs may serve as effective, accessible alternatives to face-to-face
interventions. Indeed, many couples report that they would prefer a structured, self-directed
online program (Georgia & Doss, 2013). The two most well-researched online interventions for
couples, OurRelationship (OR; based on Integrative Behavioral Couple Therapy, IBCT; Jacobson
& Christensen, 1998) and ePREP (based on the Prevention and Relationship Enhancement
curriculum, PREP; Markman, Blumberg, & Stanley, 2010), are effective at reducing relationship
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 5
and individual distress. In an initial randomized controlled trial (RCT) among couples across all
income levels, OR improved relationship satisfaction and confidence as well as individual
functioning domains including depression, anxiety, perceived health, and overall quality of life
(Doss et al., 2016). Two trials of ePREP have also demonstrated positive impacts on relationship
commitment, communication, psychological aggression, physical violence, and depression
(Braithwaite & Fincham, 2011, 2014). Results from a recent RCT specifically for low-income
couples utilizing the same sample as the current work showed significant improvements across
relationship and individual wellbeing domains for both OR and ePREP that were largely
maintained through 4 months follow-up (Doss et al., 2020; Roddy et al., 2020). However,
although the effects of intensive couple therapy typically maintain over longer follow-up periods
(Christensen, Atkins, Yi, Baucom, & George, 2006; Shadish & Baldwin, 2003), it is possible that
brief, self-directed, online methods do not yield the same maintenance of gains.
Furthermore, even if some couples are able to maintain their gains, there remains the
possibility that relapse may be more pronounced for low-income couples – either because
contextual stressors (e.g., financial problems, mental health difficulties, racial discrimination,
fewer social supports) prevent full application of skills gained during the intervention or because
these additional stressors erode the initial gains more rapidly (Bradbury & Lavner, 2011;
Johnson, 2012). At the same time, some analyses suggest that low-income couples can
experience long-term gains from relationship interventions to at least the same degree as higher-
income couples. In the Building Strong Families (BSF) and Supporting Healthy Marriage (SHM)
relationship education initiatives, improvements in communication (SHM) and relationship
quality (BSF) were largest for couples who experienced the most socioeconomic disadvantage
(Amato, 2014; Williamson, Altman, Hsueh, & Bradbury, 2016). However, implications from
these relationship education programs are limited because they were delivered primarily to non-
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 6
distressed couples and had a focus on prevention, rather than treatment, of relationship distress.
Two previous studies partially speak to the potential maintenance of brief interventions
for distressed couples. First, in a study of relatively brief treatment-as-usual couple therapy (M =
8.7 treatment hours), the sample of primarily low-income couples maintained their gains across a
variety of relationship domains 18 months after treatment (Nowlan, Georgia, & Doss, 2017).
Second, the 4-hour Marriage Checkup intervention also demonstrated that its effects on
relationship satisfaction are relatively stable through a one-year follow-up, although the sample
consisted primarily of moderately-dissatisfied, higher-income couples (Cordova et al., 2014).
Most relevant to the present study, initial findings on the maintenance of initial gains
from online couples’ programs are also encouraging. Couples in an initial trial of OR in a sample
of socioeconomically diverse couples reported maintenance of relationship and individual gains
through 12 months post-intervention, and even continued to improve in positive relationship
quality and anxiety over follow-up (Doss, Roddy, Nowlan, Rothman, & Christensen, 2019).
Previous work on ePREP has also found reductions in physical and psychological aggression one
year after the intervention (Braithwaite & Fincham, 2014). However, it is not clear whether the
initial impacts on those variables were stronger than other domains, as the authors did not report
long-term effects on satisfaction or other outcomes. ePREP’s effects on other relationship and
individual outcomes have not yet been assessed over a long-term follow-up period (Braithwaite
& Fincham, 2011) and no studies to date have examined the effect of ePREP in primarily low-
income samples. Additionally, although household income did not moderate maintenance of
gains in the year following OR in a previous study (Doss et al., 2019), the small number of low-
income households meant the study was only powered to detect large-sized interactions.
Follow-up assessment timing and frequency may play an important role in the extent to
which intervention effects appear to be maintained. Large-scale studies of relationship education
for low-income couples have often collected outcome measurements only at pre-treatment and
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 7
long-term follow-ups (Hsueh et al., 2012; Moore, Avellar, Patnaik, Covington, & Wu, 2018;
Wood, McConnell, Moore, Clarkwest, & Hsueh, 2010), making it difficult to determine whether
small or nonsignificant effects are due to lack of initial treatment response or to difficulty
maintaining gains. Additionally, the sole study of couple therapy we are aware of with a
primarily low-income sample reported outcomes only at the end of treatment and at 18-month
follow-up (Nowlan et al., 2017), precluding an understanding of the timing of any relapse.
In samples of higher-income couples, there seems to be evidence of sudden relapse
immediately following the end of the intervention followed by recovery and then relative
stability. For example, in a study of intensive couple therapy (for primarily higher-income
couples), Christensen and colleagues (2006) identified a “hockey-stick” pattern of maintenance:
An initial drop immediately in the six months after treatment followed by slow improvement
over time (regaining post-treatment levels two years after the end of treatment). Similarly,
following the brief Marriage Checkup intervention, couples partially relapsed immediately in the
month after the intervention but then were able to stabilize by 6 months and maintain the
remaining gains for up to a year (Cordova et al., 2014).
An understanding of patterns of relapse is important because it can speak to the
underlying causes of that relapse. For example, a rapid relapse following the end of an
intervention can likely be best understood as being caused by the cessation of supports or
structure provided by the intervention. In contrast, a gradual relapse would point to a slow
deterioration of the skills/insight taught in the intervention, a resumption of couples to their
previous patterns, and/or to external stressors slowly wearing on the couple’s relationship.
Current Study
The current study has two aims. First, we determined whether improvements in
relationship and individual functioning in an RCT of an online intervention for low-income
couples were maintained for one year after the intervention. Additionally, we tested for any
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 8
differences in treatment maintenance between the acceptance-focused intervention (OR) and the
behaviorally-focused intervention (ePREP), extending previous work comparing in-person
couple therapies with similar theoretical foundations (Christensen et al., 2006). Second, by using
multiple assessments during the follow-up period, this study examined the overall shape of any
relapse and whether the two interventions show different shapes of relapse.
Methods
Participants
In the larger study that established efficacy for these programs compared to a waitlist
control group during the program and through 4 month follow-up (Doss et al., 2020; Roddy et
al., 2020), a total of 990 individuals were randomized to the OR or ePREP interventions (plus
494 individuals randomized to waitlist control but not used in the current study). Of the original
990, 668 individuals (67.5%) were contacted for the present study (NOR = 332; NePREP = 336)1.
There were no differences in gender, race, ethnicity, income, age, or relationship satisfaction at
baseline, throughout the interventions, or at post-assessment between those individuals who were
and were not contacted to complete the follow up assessment (ps > .122). However, married
couples were more likely to be contacted than engaged or cohabiting couples (χ2(3) = 11.282, p
=.010), perhaps due to fluctuations in the recruitment sources of the original study. Of the 668
individuals contacted, 544 individuals (81.4%) completed the 12-month follow up assessment
(NOR = 268 individuals; NePREP = 276 individuals). Of those contacted, women were more likely to
complete the assessment than men (χ2(1) = 10.444, p = .001). However, of those contacted to
complete the 12-month assessment, there were no differences in race, ethnicity, income, age,
relationship status, or relationship satisfaction at baseline, throughout the interventions, or at
post-assessment (ps > .063) between those who did and did not complete the follow up.
1 We were unable to contact the full intervention sample of 990 due to a limited amount of
supplemental funding for this study. Additionally, at the time of award, some couples were past
the follow-up period (>13 months post-treatment) and were not contacted. Waitlist couples were
not contacted as they were offered the intervention at the end of the waiting period.
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 9
Of the 668 individuals contacted to complete the 12-month follow up assessment,
participants were 52% female and 48% male, tended to be in their early 30s (range = 18-64, M =
32.81, SD = 8.56), and the average length of individuals’ current committed romantic
relationship was six years (M = 6.14, SD = 5.32). Participants were generally representative of
the United States in terms of race and ethnicity; the majority identified as White non-Hispanic
(58%), 22% identified as Black, 8% White Hispanic, 1% American Indian Alaskan Native, 1%
Asian, 0.9% Black Hispanic, 0.3% Native Hawaiian Pacific Islander, and 7% Biracial or Other.
Highest level of education varied in the current study: 8% had no degree or diploma, 12% had a
high school General Education Development certification, 17% earned their high school
diploma, 8% a vocational or technical certification, 26% some college, 10% associate’s degree,
14% bachelor’s degree, and 4% master’s degree or other advanced graduate degree. The majority
of participants worked full time (43%) followed by 28% unemployed, 16% part-time, and 13%
reporting variable, seasonal, or temporary employment. Regarding annual individual income,
27% reported earning <$5,000, 18% earned $5,000-$11,999, 11% earned $12,000-$15,999, 17%
earned $16,000-$24,999, 13% earned $25,000-$34,999, 10% earned $35,000-$49,999, 2%
earned $50,000-$74,999, 0.3% earned $75,000-$99,999, and 1% chose not to answer.
Procedures
To be eligible for the current study, participants had to have participated in the original
web-based relationship study and be initially randomized to one of the active interventions.
Couples randomly assigned to the waitlist were not eligible for the current study. Couples were
recruited through a variety of paid (Google search ads) and free (e.g., word of mouth)
recruitment strategies and were seeking help for their relationship. Couples needed to report a
household income within 200% of the federal poverty line to be eligible; full eligibility criteria
can be found in Doss and colleagues (2020). Participants were contacted 14 months following
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 10
the date they entered the original study (approximately 12 months after program completion) via
email, phone call, and text message and asked to complete an online questionnaire hosted by
Qualtrics. Participants were compensated for completing this survey with $25 electronic gift
cards (see online supplement Figure 1 for Consort diagram).
Interventions
The interventions are described in detail in Doss and colleagues (2020). Briefly,
intervention couples completed online intervention materials on their own, and met with a coach
at the start of the program and three times during the program for a total of one hour of contact.
ePREP. The ePREP program consists of six 60-minute modules focused primarily on
communication and conflict management skills, as well as psychoeducation about enhancing
relationships through commitment and positive bonding. Additionally, couples were asked to
complete one to two hours of homework outside of the program each week.
OurRelationship. The OurRelationship program is comprised of three phases designed
to be completed in seven hours over the course of six weeks. These phases help couples identify
a specific problem in their relationship, build an acceptance-focused understanding of that
problem consistent with IBCT’s DEEP formulation (see Jacobson & Christensen, 1998), and
commit to behavioral changes to improve that problem. Most activities are completed
individually, and couples come together for structured conversations at the end of each phase.
Measures
Relationship satisfaction. The Couple Satisfaction Index-four item (CSI-4) was used to
measure relationship satisfaction; it is widely used with strong psychometric properties (Funk &
Rogge, 2007). Scores on the CSI-4 range from 0 to 21, with higher scores representing greater
relationship satisfaction. In the current study, internal consistency was excellent (α = .92).
Relationship status. A single item dichotomous question was used to assess relationship
status: “What is your current relationship with the partner with whom you started the program?”
(0 = We broke up; 1 = We are still together). A relationship was considered broken up if either
member of the couple reported they were no longer in a relationship.
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 11
Breakup potential. A three-item scale, adapted from the Marital Instability Index
(Edwards, Johnson, & Booth, 1987), was used for participants to rate their perceived likelihood
of breakup (e.g., “The thought of ending my relationship has crossed my mind”). Items were
rated on a five-point scale with higher scores indicating greater potential. The measure had good
internal consistency in the current sample (α = .83).
Communication Skills. A seven-item measure developed for the ACF Supporting
Healthy Marriage initiative (Lundquist et al., 2014) was used to measure participants’
communication skills over the past month. A sample item was: “During arguments, my
partner/spouse and I were good at taking breaks when we needed them” using a four-point scale
(Never to Often). Internal consistency was good in the present sample (α = .89).
Communication conflict. A seven-item measure of negative communication developed
for the ACF Supporting Health Marriage project was used for the current study (Lundquist et al.,
2014). Participants rated how often (Never to Often) communication conflict happened over the
previous month; e.g., “Small issues suddenly became big arguments.” Internal consistency was
good in the present sample (α = .89).
Emotional support. Five items were used from a seven-item measure of emotional
support developed for the ACF Supporting Healthy Marriage project (Lundquist et al., 2014).
Participants rated agreement with statements such as, “I can count on my partner/spouse to be
there for me” on a four-point scale; higher scores indicated greater emotional intimacy. Internal
consistency was good (α = .83).
Intimate Partner Violence. In consultation with the National Domestic Violence
Hotline, seven items were developed to assess intimate partner violence over the last month.
Participants reported whether their current partner pushed or shoved them, grabbed them,
scratched them, slapped them, threw something at them, bit them, or punched them during the
past month. The frequency of each of these behaviors (0 – 20+ times in the last month) was
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 12
dichotomized to no IPV (0) or occurrence of IPV (1).
Psychological Distress. The six-item Kessler Psychology Distress scale assessed
psychological distress (Kessler et al., 2002). Items such as, “In the last 30 days, how often have
you felt nervous?” were rated on a scale from 1 (None of the time) to 5 (All of the time). The
69.6% of individuals whose sum score was greater than or equal to 13 at pre-treatment were
considered distressed (Kessler et al., 2002). Internal consistency was good (α = .86).
Perceived Stress. The four-item Perceived Stress Scale (Cohen & Williamson, 1988)
measured feelings of being overwhelmed and out of control. Participants rated the frequency of
feelings such as “In the last 30 days, how often have you felt that you were unable to control the
important things in your life?” on a 0 (Never) to 4 (Very often) scale. Individuals with sum scores
greater than or equal to 6.27 at pre-treatment were distressed (Cohen & Williamson, 1988);
74.3% were distressed at baseline. Internal consistency was adequate in this sample (α = .74.)
Anger. The NIH PROMIS five-item Anger-Short Form measure assessed anger in the
current study (Pilkonis et al., 2011), with instructions modified to ask participants to omit angry
feelings toward their partner. Participants rated items such as, “In the last 30 days, I was irritated
more than other people,” on a scale from 1 (Never) to 5 (Always). Internal consistency was
excellent in this sample (α = .92). Individuals with sum scores greater than or equal to 16 at pre-
treatment were coded as distressed (Pilkonis et al., 2011); 41.8% were distressed at baseline.
Alcohol Use. Problematic alcohol use over the past month was assessed using the seven-
item PROMIS Alcohol Use measure (Pilkonis et al., 2013). Participants endorsed alcohol-related
problems, such as, “Others complained about my drinking,” on a scale from 1 (Never) to 5
(Almost always). Internal consistency was excellent in this sample (α = .91). Individuals with
sum scores greater than or equal to 9 at pre-treatment were classified as distressed (Pilkonis et
al., 2013); 22.0% were distressed at baseline.
Perceived Health. The General Health Perceptions subscale from the SF-36 has five
items to assess overall physical health (Ware & Sherbourne, 1992). Participants rated items such
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 13
as, “I am as healthy as anyone I know,” from 1 (Definitely false) to 5 (Definitely true). Internal
consistency was good in this sample (α = .84). Individuals with sum scores less than or equal to
11.65 were coded as distressed (Ware & Sherbourne, 1992); 9% were distressed at baseline.
Sleep. The Insomnia Severity Index (ISI) has five-items to assess current sleep quality
(Bastien, Vallières, & Morin, 2001). Participants rate items such as, “How satisfied are you with
your current sleep pattern?” from 0 to 4, with higher numbers anchored at greater sleep difficulty.
The ISI had acceptable internal consistency in this sample (α = 0.86). Individuals with sum
scores greater than or equal to 10 were initially distressed; 55% were distressed at baseline.
Missing Data and Data Analysis
Some data were missing at posttreatment (10.3%), 2 months follow-up (12.5%), 4 months
follow-up (13.0%), and at 12 months follow-up (18.6%) among the 688 individuals contacted.
All missing data were imputed using the multilevel imputation software Blimp (Enders, Keller,
& Levey, 2017; Keller & Enders, 2018) to create 10 imputed datasets. Imputation models
accounted for shape of data and auxiliary variable related to missingness. Analyses report an
average result across the 10 imputed datasets.
Three-level models were run in HLM software (HLM 7.03; Raudenbush, Bryk, Cheong,
Congdon, & Toit, 2011) with time at level 1 nested within individuals at level 2 nested within
couples at level 3. Both linear and quadratic parameterizations of time were tested for all models
from posttreatment to one-year follow-up. Gender was grand mean centered and entered at level
2, and treatment condition (OR or ePREP) was entered uncentered at level 3 (Atkins, 2005).
Models were run twice for each outcome, once with each treatment group coded as zero in order
to obtain within-group estimates for each condition. All models were linear regressions except
IPV, which was a logistic regression. Effect sizes (Cohen’s d) were calculated by multiplying the
slope coefficient by the corresponding length of time and dividing that product by the pooled
pre-treatment standard deviation.
To calculate effect sizes for total change across the intervention and one-year follow-up
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 14
period, we ran a second set of three-level models in HLM including only the baseline (coded as
0) and one-year follow-up (coded as 1) assessment points at level 1. Gender and treatment
condition were included as described above in levels 2 and 3, respectively. In models for
individual functioning outcomes, a grand-mean centered dichotomous variable was included at
level 2 to indicate whether the participant was below established clinical cutoffs at baseline; this
is because previous analyses indicate significant moderation effects by initial distress for
individual (Roddy et al., 2020) but not relationship (Doss et al., 2020) outcomes during
treatment. Individuals, not couples, were classified as “distressed” on an outcome-by-outcome
basis using established norms for the measures of individual functioning in order to test for
differential outcomes for those above or below the established cutoffs at baseline.
Results
Supplementary Table 1 reports descriptive statistics, separated by condition. One year
following the program among the 668 individuals contacted, 68.3% of couples were still in a
relationship with the partner with whom they completed the program, 26.0% had broken up, and
we were unable to determine relationship status for the remaining 5.7% of couples (i.e., they
were still in a relationship at 4-month follow-up but no data was available at 12 months).
Long-term Relationship Functioning
From post-treatment to the 12-month follow-up, analyses showed that couples in OR and
ePREP did not report any significant linear or quadratic changes in any domains of relationship
functioning; further, there were no significant differences in maintenance of gains between the
two conditions (see Table 1 for full results). Moreover, when inspecting these non-significant
changes, there was not a consistent pattern of continued improvement versus deterioration across
relationship functioning in the year after the intervention.
Long-term Individual Functioning
Similar to findings for relationship functioning, couples in OR and ePREP did not
generally report statistically significant linear or quadratic change in individual functioning over
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 15
one-year follow-up; additionally, there were no statistically significant differences in
maintenance of gains between the two conditions. See Table 2 for full results. One notable
exception was couples in ePREP who reported significant linear increases in perceived health in
the 12-months following the program (b = 0.029, SE = 0.013, p = .025), unlike OR couples;
however, the difference between conditions was non-significant.
Pretreatment to Long-term Follow-up
From pretreatment to one-year follow-up, couples in both conditions experienced
significant improvements in all measured domains of relationship functioning (all effect sizes
reported are within-group d): relationship satisfaction (dePREP = 0.85; dOR = 0.83), emotional
intimacy (dePREP = 0.52; dOR = 0.55), and communication skills (dePREP = 1.04; dOR = 1.06), as well
as significant decreases in communication conflict (dePREP = -1.13; dOR = -1.29) and break-up
potential (dePREP = -0.85; dOR = -0.79). Couples in both groups reported non-significant decreases
in IPV from pretreatment to one-year follow-up (dOR = -0.26; dePREP = -0.16, calculated using a
odds ratio to d transformation; Sánchez-Meca, Marín-Martínez, & Chacón-Moscoso, 2003).
There were no significant differences between conditions from pretreatment to one-year follow-
up on any relationship outcomes (Figure 1).
Over the course of the intervention and one-year follow-up, couples in both conditions
reported significant improvements in alcohol use (dePREP = -0.19; dOR = -0.30), anger (dePREP =
-0.63; dOR = -0.52), insomnia (dePREP = -0.36; dOR = -0.27), psychological distress (dePREP = -0.77;
dOR = -0.68), and perceived stress (dePREP = -0.80; dOR = -0.73), and perceived health (dePREP =
0.27; dOR = 0.19); see Table 3 for detailed results. No significant differences emerged between
conditions on overall individual functioning outcomes. Further, couples in both conditions
reported significantly larger gains if they started the intervention in the clinically distressed range
for all individual outcomes (see Figure 2). For perceived health only, this interaction was further
moderated by treatment condition, such that clinically distressed participants in the OR condition
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 16
experienced the greatest gains in perceived health (b = -2.622, SE = 1.025, p = 0.011).
Discussion
From post-treatment to one year later, OR and ePREP participants did not report
significant changes in relationship functioning. Previous analyses from the RCT within this
sample demonstrated the programs were superior to the waitlist control group and couples
maintained gains through four months follow-up (Doss et al., 2020). This paper extends this
period of maintenance and replicates a previous trial of OR in primarily higher-income couples
(Doss et al., 2019). It is particularly valuable to show a replication of long-term maintenance in a
low-income sample. In contrast to previous studies of maintenance following in-person couple
therapy (i.e., an initial sharp drop followed by gradual increases, Christensen et al., 2006), or in-
person assessment and feedback interventions (i.e., steep drops that leveled with time, Cordova
et al., 2014), we found no linear or quadratic change during the year following treatment in
relationship outcomes, suggesting couples consistently maintained gains over follow-up.
Repeated measurements over follow-up allowed us to test for non-linear change, a limitation of
previous studies of maintenance of couple therapy (Doss et al., 2014) or relationship education
(Hsueh et al., 2012; Wood et al., 2010) in low-income samples. Online programs for relationship
distress are therefore a viable option for low-income couples.
Results for maintenance of individual functioning were largely similar to relationship
functioning, with the exception of perceived health. Couples in both conditions generally
maintained the gains in individual functioning over the year following the interventions, with
minimal differences between OR and ePREP, indicating brief, relationship-focused programs
have lasting spill-over effects into individual domains. The exception to this was perceived
health, which increased significantly in the year following the program for couples in ePREP.
Couples in OR did not significantly change following the program and the slopes of the two
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 17
programs were non-significantly different. Previous analyses of short-term follow-up did not
detect this change (Roddy et al., 2020), highlighting the need for longer follow-up to fully
understand the impacts of online relationship programs for low-income couples’ health.
There were relatively few differences between ePREP and OR in maintenance of
relationship and individual functioning during the year following the program. Previous research
comparing maintenance of behaviorally-based and acceptance-based in-person couple therapy
demonstrated different patterns over follow-up (Christensen et al., 2006). The lack of differences
here is encouraging in that both programs are impactful for low-income couples over long-term
follow-up. However, the general equivalence also suggests a need for more research on the
mechanisms by which these two approaches impact relationship functioning and if those
mechanisms are the same or different in an online context.
Magnitude of Effects of Interventions One Year After Completion
OR and ePREP produced meaningful, lasting change on a number of key relationship
domains. Of note, the effects described here are within-person changes, not changes compared to
a waitlist control group, and therefore may be larger than between-group comparisons. In the
larger study (Doss et al., 2020), couples in the waitlist control group reported non-significant
changes over the first four months of follow-up – generally in the direction of improvement.
Whereas previous studies of OR (Doss et al., 2016, 2019) demonstrated minimal changes in
positive relationship quality during the intervention and significant improvements over one-year
follow-up, this pattern was not replicated here. Given the current sample of lower-income
couples, it may be that couples with more socioeconomic disadvantage have potential for more
benefit from these online programs, and thus experience changes faster. Alternatively, the
different trajectories of change could be due to modifications made to OR between these two
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 18
trials to directly enhance positive relationship quality. Unfortunately, we are unable to directly
test these hypotheses due to limitations in study design. Finally, neither OR nor ePREP had a
lasting impact on incidence of IPV (despite a significant effect at the end of treatment in this
sample; Doss et al., 2020). A previous study of ePREP showed effects on physical and
psychological aggression at one-year post-treatment (Braithwaite & Fincham, 2014), results we
failed to replicated here. Thus, further work is needed to understand the mechanisms of effects of
both programs given some differences observed relative to prior studies of these programs.
There were small to medium effects for individual functioning at one-year post-treatment.
Additionally, larger effects were seen for individuals who began the intervention in the distressed
range, replicating previous work (Doss et al., 2016). Previous analyses of individual functioning
with this sample demonstrated moderation by initial distress during the intervention, but not over
short-term follow-up (Roddy et al., 2020); therefore, it is encouraging that we continue to see
benefits for the most distressed participants over a longer follow-up period. Specifically,
distressed individuals maintained the gains from the program and did not experience significant
relapse in the year following the intervention.
Strengths and Limitations
This is the first study to our knowledge to analyze the shape of change over long-term
follow-up from a large RCT of online relationship interventions. Additionally, this study had
good retention with a large sample size one year following the study which is especially notable
in a sample of low-income individuals. Finally, outcomes for both relationship and individual
functioning were included, demonstrating the wide impact the online programs had for up to a
year. However, this study is not without limitation. Only a subsample of the full RCT was
contacted at one-year follow-up and none of the waitlist control group was contacted (as some of
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 19
them had completed the interventions in the interim), limiting our ability to compare
maintenance of the intervention to naturally occurring trends in a control group. Second, a large
number of analyses were run in the current study, increasing the risk of a Type 1 error. Finally,
only self-report data were used; inclusion of objective measures of relationship functioning such
as observed communication measures and individual functioning such as cortisol or heart rate
variability would strengthen results found here.
Future Directions
Continued research in this area should collect longer follow-up data on ePREP and OR,
as has been done in in-person couple therapy (e.g., Christensen et al., 2010; Snyder, Wills, &
Grady-Fletcher, 1991) and relationship education programs (e.g., Hsueh et al., 2012).
Additionally, it is important for future research to work to understand the mechanisms of both
initial change and maintenance in these programs. Better knowledge of mechanisms could
provide information to help maximize benefits, such as identifying the need for brief booster
interventions over follow-up. This work highlights the ability of online programs, that by design
decrease barriers to entry, to impact relationship and individual functioning up to a year
following treatment within a sample of low-income couples.
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 20
References
Amato, P. R. (2014). Does social and economic disadvantage moderate the effects of relationship
education on unwed couples? An analysis of data from the 15-month Building Strong
Families evaluation. Family Relations, 63(3), 343-355. doi:10.1111/fare.12069
Atkins, D. C. (2005). Using multilevel models to analyze couple and family treatment data: basic
and advanced issues. Journal of Family Psychology, 19(1), 98-110. doi: 10.1037/0893-
3200.19.1.98
Bastien, C. H., Vallières, A., & Morin, C. M. (2001). Validation of the Insomnia Severity Index
as an outcome measure for insomnia research. Sleep medicine, 2(4), 297-307. doi:
10.1016/S1389-9457(00)00065-4
Beach, S. R. H., Fincham, F. D., Amir, N., & Leonard, K. E. (2005). The taxometrics of
marriage. Journal of Family Psychology, 19(2), 276-285. doi:10.1037/0893-3200.19.2.276
Bradbury, T. N., & Lavner, J. A. (2011). How can we improve preventive and educational
interventions for intimate relationships? Behavior Therapy, 43(1), 113-122.
doi:10.1016/j.beth.2011.02.008
Braithwaite, S. R., & Fincham, F. D. (2011). Computer-based dissemination: A randomized
clinical trial of ePREP using the actor partner interdependence model. Behaviour Research
and Therapy, 49(2), 126-131. doi:10.1016/j.brat.2010.11.002
Braithwaite, S. R., & Fincham, F. D. (2014). Computer-based prevention of intimate partner
violence in marriage. Behaviour Research and Therapy, 54, 12-21.
doi:10.1016/j.brat.2013.12.006
Bramlett, M. D., & Mosher, W. D. (2002). Cohabitation, marriage, divorce, and remarriage in the
United States. Vital and Health Statistics. Series 23, Data from the National Survey of
Family Growth, (22), 1-93. Retrieved from
https://www.ncbi.nlm.nih.gov/pubmed/12183886
Christensen, A., Atkins, D. C., Baucom, B., & Yi, J. (2010). Marital status and satisfaction five
years following a randomized clinical trial comparing traditional versus integrative
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 21
behavioral couple therapy. Journal of Consulting and Clinical Psychology, 78(2), 225-235.
doi:10.1037/a0018132
Christensen, A., Atkins, D. C., Yi, J., Baucom, D. H., & George, W. H. (2006). Couple and
individual adjustment for 2 years following a randomized clinical trial comparing traditional
versus integrative behavioral couple therapy. Journal of Consulting and Clinical
Psychology, 74(6), 1180-1191. doi:10.1037/0022-006X.74.6.1180
Cohen, S., & Williamson, G. (1988). Perceived stress in a probability sample of the United
States. In S. Spacapan & S. Oskamp (Eds.), The social psychology of health (pp. 31–68).
Newbury Park, CA: Sage.
Cordova, J. V., Fleming, C. J. E., Morrill, M. I., Hawrilenko, M., Sollenberger, J. W., Harp, A.
G., . . . Wachs, K. (2014). The Marriage Checkup: A randomized controlled trial of annual
relationship health checkups. Journal of Consulting and Clinical Psychology, 82(4), 592-
604. doi:10.1037/a0037097
Cunradi, C., Caetano, R., & Schafer, J. (2002). Socioeconomic predictors of intimate partner
violence among White, Black, and Hispanic couples in the United States. Journal of Family
Violence, 17(4), 377-389. doi:1020374617328
Doss, B. (2014). Oklahoma marriage initiative, baseline report (unpublished analyses).
Doss, B. D., Cicila, L. N., Georgia, E. J., Roddy, M. K., Nowlan, K. M., Benson, L. A., &
Christensen, A. (2016). A randomized controlled trial of the web-based OurRelationship
program: Effects on relationship and individual functioning. Journal of Consulting and
Clinical Psychology, 84(4), 285-296. doi:10.1037/ccp0000063
Doss, B. D., Roddy, M. K., Nowlan, K. M., Rothman, K., & Christensen, A. (2019).
Maintenance of gains in relationship and individual functioning following the online
OurRelationship program. Behavior Therapy, 50(1), 73-86. doi:10.1016/j.beth.2018.03.011
Edwards, J. N., Johnson, D. R., & Booth, A. (1987). Coming apart: A prognostic instrument of
marital breakup. Family Relations, 168-170. doi: 10.2307/583948
Enders, C. K., Keller, B. T., & Levy, R. (2017). A chained equations imputation approach for
multi-level data with categorical and continuous variables. Psychological Methods, Advance
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 22
online publication. http://dx.doi.org/10.1037/met0000148.
Foran, H. M., Whisman, M. A., & Beach, S. R. H. (2015). Intimate partner relationship distress
in the DSM‐5. Family Process, 54(1), 48-63. doi:10.1111/famp.12122
Funk, J. L., & Rogge, R. D. (2007). Testing the ruler with item response theory: increasing
precision of measurement for relationship satisfaction with the Couples Satisfaction
Index. Journal of Family Psychology, 21(4), 572. doi: 10.1037/0893-3200.21.4.572
Gaubert, J. M., Gubits, D., Principe Alderson, D., & Knox, V. (2012). The Supporting Healthy
Marriage evaluation: Final implementation findings. OPRE Report, 2012-12.
Georgia, E. J., & Doss, B. D. (2013). Web-based couple interventions: Do they have a future?
Journal of Couple and Relationship Therapy, 12(2), 168-185.
doi:10.1080/15332691.2013.779101
Halford, W. K., O'Donnell, C., Lizzio, A., & Wilson, K. L. (2006). Do couples at high risk of
relationship problems attend premarriage education? Journal of Family Psychology, 20(1),
160-163. doi:10.1037/0893-3200.20.1.160
Hawkins, A. J., Blanchard, V. L., Baldwin, S. A., & Fawcett, E. B. (2008). Does marriage and
relationship education work? A meta-analytic review. Journal of Consulting and Clinical
Psychology, 76, 723-734. doi: 10.1037/a0012584
Hsueh, J., Alderson, D. P., Lundquist, E., Michalopoulos, C., Gubits, D., Fein, D., & Knox, V.
(2012). The Supporting Healthy Marriage evaluation: Early impacts on low-income
families. OPRE Report, 2012-11.
Jacobson, N. S., & Christensen, A. (1998). Acceptance and change in couple therapy: A
therapist's guide to transforming relationships. New York: Norton.
Johnson, M. D. (2012). Healthy marriage initiatives: On the need for empiricism in policy
implementation. The American Psychologist, 67(4), 296-308. doi:10.1037/a0027743
Keller, B. T., & Enders, C. K. (2018). Blimp User's Manual (Version 1.1). Los Angeles, CA.
Kessler, R. C., Andrews, G., Colpe, L. J., Hiripi, E., Mroczek, D. K., Normand, S. L., ... &
Zaslavsky, A. M. (2002). Short screening scales to monitor population prevalences and
trends in non-specific psychological distress. Psychological medicine, 32(6), 959-976. doi:
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 23
10.1017/S0033291702006074
Lundquist, E., Hsueh, J., Lowenstein, A. E., Faucetta, K., Gubits, D., Michalopoulos, C., &
Knox, V. (2014). A family-strengthening program for low-income families: Final impacts
from the supporting healthy marriage evaluation. OPRE Report, 2014-09A.
Markman, H. J., Blumberg, S. L., & Stanley, S. M. (2010). Fighting for your marriage (3rd ed.).
Hoboken: Jossey-Bass.
Moore, Q., Avellar, S., Patnaik, A., Covington, R. and Wu, A. (2018). Parents and Children
Together: Effects of Two Healthy Marriage Programs for Low-Income Couples. OPRE
Report Number 2018-58. Washington, DC: Office of Planning, Research, and Evaluation,
Administration for Children and Families, U.S. Department of Health and Human Services.
Nowlan, K. M., Georgia, E. J., & Doss, B. D. (2017). Long-term effectiveness of treatment-as-
usual couple therapy for military veterans. Behavior Therapy, 48(6), 847-859.
doi:10.1016/j.beth.2017.05.007
Pilkonis, P. A., Choi, S. W., Reise, S. P., Stover, A. M., Riley, W. T., Cella, D., & PROMIS
Cooperative Group. (2011). Item banks for measuring emotional distress from the Patient-
Reported Outcomes Measurement Information System (PROMIS®): depression, anxiety,
and anger. Assessment, 18(3), 263-283. doi: 10.1177/1073191111411667
Raudenbush, S. W., Bryk, A. S., Cheong, Y. F., Congdon Jr., R. T., & Toit, M. D. (2011). HLM7:
Hierarchical Linear and Nonlinear Modeling. Scientific Software International:
Lincolnwood, IL.
Sánchez-Meca, J., Marín-Martínez, F., & Chacón-Moscoso, S. (2003). Effect-size indices for
dichotomized outcomes in meta-analysis. Psychological methods, 8(4), 448-467. doi:
10.1037/1082-989X.8.4.448
Shadish, W. R., & Baldwin, S. A. (2003). Meta-analysis of MFT interventions. Journal of
Marital and Family Therapy, 29(4), 547-570. doi:10.1111/j.1752-0606.2003.tb01694.x
Snyder, D. K., Wills, R. M., & Grady-Fletcher, A. (1991). Long-term effectiveness of behavioral
versus insight-oriented marital therapy: A 4-year follow-up study. Journal of Consulting
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 24
and Clinical Psychology, 59(1), 138-141. doi:10.1037/0022-006X.59.1.138
Ware Jr, J. E., & Sherbourne, C. D. (1992). The MOS 36-item short-form health survey (SF-36):
I. Conceptual framework and item selection. Medical care, 473-483.
https://www.jstor.org/stable/3765916
Whisman, M. A., Beach, S. R. H., & Snyder, D. K. (2008). Is marital discord taxonic and can
taxonic status be assessed reliably? Results from a national, representative sample of
married couples. Journal of Consulting and Clinical Psychology, 76(5), 745-755.
doi:10.1037/0022-006X.76.5.745
Williamson, H. C., Altman, N., Hsueh, J., & Bradbury, T. N. (2016). Effects of relationship
education on couple communication and satisfaction: A randomized controlled trial with
low-income couples. Journal of Consulting and Clinical Psychology, 84(2), 156-166.
doi:10.1037/ccp0000056
Williamson, H. C., Karney, B. R., & Bradbury, T. N. (2019). Barriers and facilitators of
relationship help-seeking among low-income couples. Journal of Family Psychology, 33(2),
234-239. doi:10.1037/fam0000485
Wood, R. G., McConnell, S., Moore, Q., Clarkwest, A., & Hsueh, J. (2010). The Building Strong
Families project: Strengthening unmarried parents' relationships: The early impacts of
building strong families. Washington, DC: Office of Planning, Research, and Evaluation,
Administration for Children and Families, U.S. Department of Health and Human Services.
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 25
Table 1
HLM Relationship Functioning Results
Measure
Post to follow-up Pre to follow-up
Linear slope Quadratic slope Effect sizeaChange Effect sizea
b SE pb SE pCohen’s db SE pCohen’s d
Main effect OR
Satisfaction -0.003 0.024 0.884 0.000 0.000 0.777 0.03 3.896 0.439 <0.001 0.83
Conflict 0.012 0.027 0.655 0.000 0.000 0.492 -0.05 -6.282 0.472 <0.001 -1.29
Emotional Support 0.009 0.015 0.545 0.000 0.000 0.756 0.08 1.767 0.283 <0.001 0.52
Breakup -0.009 0.007 0.182 0.000 0.000 0.206 0.04 -0.872 0.093 <0.001 -0.79
Communication
Skills -0.013 0.020 0.534 0.000 0.000 0.511 -0.01 4.497 0.384 <0.001 1.06
IPV 0.006 0.014 0.656 0.000 0.000 0.884 0.15 -0.430 0.221 0.055 -0.26
Main effect ePREP
Satisfaction 0.025 0.022 0.259 0.000 0.000 0.698 0.21 3.975 0.423 <0.001 0.85
Conflict -0.018 0.026 0.471 0.000 0.000 0.830 -0.15 -5.481 0.519 <0.001 -1.13
Emotional Support 0.003 0.014 0.859 0.000 0.000 0.637 -0.06 1.848 0.306 <0.001 0.55
Breakup -0.011 0.006 0.076 0.000 0.000 0.221 -0.09 -0.940 0.083 <0.001 -0.85
Communication
Skills -0.029 0.021 0.170 0.001 0.000 0.088 0.06 4.382 0.380 <0.001 1.04
IPV 0.022 0.017 0.208 0.000 0.000 0.272 0.17 -0.260 0.254 0.311 -0.16
Difference between conditions (ePREP = 1)
Satisfaction 0.028 0.033 0.390 0.000 0.001 0.639 0.17 0.079 0.605 0.896 0.02
Conflict -0.031 0.037 0.411 0.000 0.001 0.515 -0.11 0.801 0.689 0.245 0.16
Emotional Support -0.007 0.021 0.751 0.000 0.000 0.926 -0.14 0.082 0.431 0.850 0.02
Breakup -0.001 0.009 0.884 0.000 0.000 0.864 -0.14 -0.068 0.126 0.590 -0.06
Communication
Skills -0.016 0.030 0.585 0.000 0.000 0.433 0.07 -0.115 0.524 0.827 -0.03
IPV 0.016 0.020 0.425 0.000 0.000 0.402 0.02 0.169 0.333 0.613 0.10
Note: Bolded terms indicate significant findings at p < .05. aWithin-group effect sizes for the main effects or OR and ePREP as well as
the between group effect sizes were calculated from posttreatment to 12 months follow-up using the full sample pretreatment standard
deviation.
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 26
Table 2
HLM Individual Functioning Results
Measure Linear slope Quadratic slope
Post to follow-
up effect sizea
b SE pb SE pCohen’s d
Main effect OR
Psychological Distress 0.004 0.023 0.876 0.000 0.000 0.982 0.03
Perceived Stress 0.008 0.016 0.598 0.000 0.000 0.618 0.02
Anger 0.026 0.023 0.247 0.000 0.000 0.362 0.10
Alcohol Use 0.006 0.022 0.791 0.000 0.000 0.680 -0.02
Perceived Health 0.005 0.014 0.731 0.000 0.000 0.718 0.00
Insomnia 0.025 0.023 0.280 0.000 0.000 0.678 0.15
Main effect ePREP
Psychological Distress -0.003 0.025 0.892 0.000 0.000 0.713 -0.12
Perceived Stress 0.007 0.017 0.670 0.000 0.000 0.353 -0.14
Anger -0.002 0.023 0.922 0.000 0.000 0.701 -0.13
Alcohol Use -0.005 0.021 0.822 0.000 0.000 0.807 0.00
Perceived Health 0.029 0.013 0.025 0.000 0.000 0.091 0.13
Insomnia 0.039 0.023 0.093 -0.001 0.000 0.157 0.07
Difference between conditions (ePREP = 1)
Psychological Distress -0.007 0.034 0.833 0.000 0.001 0.786 -0.15
Perceived Stress -0.001 0.023 0.962 0.000 0.000 0.719 -0.16
Anger -0.029 0.032 0.376 0.000 0.001 0.728 -0.22
Alcohol Use -0.011 0.031 0.728 0.000 0.000 0.641 0.02
Perceived Health 0.025 0.019 0.202 0.000 0.000 0.379 0.13
Insomnia 0.014 0.033 0.673 0.000 0.001 0.471 -0.08
Note: Bolded terms indicate significant findings at p < .05. aWithin-group effect sized for the main effects or OR and ePREP as well as
the between group effect size were calculated from posttreatment to 12 months follow-up using the full sample pretreatment standard
deviation.
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 27
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 28
Table 3
HLM Individual Functioning Results from Pre to One-year Follow-up Moderated by Initial Distress
Difference Initial distress Cohen’s d effect sizea
Measure
b SE pb SE pOverall
Without
distress
With
distress
Main effect OR
Psychological Distress -3.849 0.416 <0.001 -6.040 0.753 <0.001 -0.68 0.06 -1.01
Perceived Stress -2.278 0.240 <0.001 -2.827 0.463 <0.001 -0.73 -0.06 -0.96
Anger -2.435 0.309 <0.001 -5.380 0.609 <0.001 -0.52 -0.04 -1.19
Alcohol Use -1.818 0.292 <0.001 -6.917 1.000 <0.001 -0.30 -0.05 -1.19
Perceived Health 0.820 0.238 <0.001 4.194 0.785 <0.001 0.19 0.10 1.08
Insomnia -1.670 0.381 <0.001 -5.976 0.718 <0.001 -0.27 0.26 -0.70
Main effect ePREP
Psychological Distress -4.324 0.370 <0.001 -6.049 0.670 <0.001 -0.77 -0.02 -1.09
Perceived Stress -2.502 0.243 <0.001 -3.421 0.499 <0.001 -0.80 0.01 -1.08
Anger -2.955 0.340 <0.001 -5.469 0.636 <0.001 -0.63 -0.14 -1.31
Alcohol Use -1.163 0.357 0.001 -6.065 1.177 <0.001 -0.19 0.03 -0.97
Perceived Health 1.166 0.231 <0.001 1.572 0.627 0.012 0.27 0.24 0.60
Insomnia -2.252 0.391 <0.001 -5.183 0.686 <0.001 -0.36 0.10 -0.73
Difference between conditions (ePREP = 1)
Psychological Distress -0.476 0.566 0.401 -0.010 1.015 0.992 -0.08 -0.08 -0.08
Perceived Stress -0.224 0.334 0.502 -0.594 0.672 0.377 -0.07 0.07 -0.12
Anger -0.520 0.453 0.253 0.089 0.870 0.918 -0.11 -0.12 -0.10
Alcohol Use 0.654 0.453 0.150 0.852 1.580 0.590 0.11 0.08 0.22
Perceived Health 0.347 0.337 0.304 -2.622 1.025 0.011 0.08 0.14 -0.47
Insomnia -0.582 0.544 0.285 0.793 0.999 0.428 -0.09 -0.16 -0.04
Note: Bolded terms indicate statistically significant findings at p < .05. aWithin-group effect sizes for the main effects or OR and
ePREP as well as the between group effect sizes were calculated from posttreatment to 12 months follow-up using the full sample
pretreatment standard deviation.
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 29
Figure 1
Pretreatment to One-year Follow-up Within-group Effect Sizes for Relationship Functioning
Relationship Satisfaction
Positive Communication
Communication Conflict
IPV
Breakup Potential
Emotional Support
-1.50
-1.00
-0.50
0.00
0.50
1.00
1.50
0.85
1.04
-1.13
-0.16
-0.85
0.55
0.83
1.06
-1.29
-0.26
-0.79
0.52
ePREP OurRelationship
Within-group Cohen's d
Note: *** p < .001
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 30
Figure 2
Pretreatment to Long-term Follow-up Within-group Effect Sizes for Individual Functioning
-1.50
-1.00
-0.50
0.00
0.50
1.00
1.50
-0.30
-0.52
0.19
-0.27
-0.68 -0.73
-1.19 -1.19
1.08
-0.70
-1.01 -0.96
OurRelationship
Overall With Initial Distress
Cohen's d
-1.50
-1.00
-0.50
0.00
0.50
1.00
-0.19
-0.63
0.27
-0.36
-0.77 -0.80
-0.97
-1.31
0.60
-0.73
-1.09 -1.08
ePREP
Overall With Initial Distress
Cohen's d
Note: Asterisks on overall effect sizes indicate significant changes from pretreatment to one-year
follow-up. Asterisks on “With Initial Distress” effect sizes indicate significant moderation by
initial distress from pretreatment to one-year follow-up. **p <.01; *** p < .001
Perceived
Health
Figure 2a
Alcohol Use Anger Insomnia Psychological
Distress
Perceived Stress
Perceived
Health
Perceived Stress
Psychological
Distress
Insomnia
Anger
Alcohol Use
Figure 2b
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 31
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 32
Supplementary Table 1
Means and Standard Deviations by Timepoint and Condition
Measure Pre Post
2 months
follow-up
4 months
follow-up
12 months
follow-up
Mean SD Mean SD Mean SD Mean SD Mean SD
OurRelationship
Relationship
Satisfaction 9.13 4.69 12.80 5.07 12.93 4.76 12.78 5.03 13.03 5.51
Conflict 21.98 4.87 16.07 5.87 15.92 5.56 16.11 5.61 15.70 6.16
Emotional
Support 13.99 3.30 15.93 3.43 16.08 3.35 15.81 3.48 15.77 3.78
Breakup 2.77 1.10 1.85 1.01 1.72 0.94 1.76 0.98 1.89 1.04
Positive
Communication 17.06 4.23 21.49 4.50 21.60 4.39 21.37 4.36 21.58 4.99
IPV 0.31 0.46 0.17 0.38 0.20 0.40 0.19 0.39 0.22 0.42
Individual
Psych Distress 16.09 5.63 12.20 5.54 12.03 5.37 12.25 5.42 12.28 5.76
Perceived Stress 8.36 3.12 6.00 3.49 6.23 3.22 6.02 3.44 6.10 3.62
Anger 14.50 4.69 11.62 4.52 11.97 4.35 11.91 4.62 12.06 4.75
Alcohol Use 5.91 6.05 4.15 4.69 4.14 4.85 4.21 5.22 3.96 5.11
Health 17.87 4.29 18.69 4.18 18.70 4.35 18.81 4.39 18.70 4.28
Insomnia 10.85 6.22 8.03 6.48 8.53 6.60 8.45 6.19 9.03 6.59
ePREP
Relationship
Satisfaction 9.10 4.69 12.16 5.00 12.29 4.77 12.63 5.09 13.04 5.10
MAINTENANCE OF ONLINE PROGRAM FOR LOW-INCOME DYADS 33
Conflict 21.02 4.87 16.30 5.65 16.13 5.66 15.93 5.94 15.55 6.09
Emotional
Support 13.80 3.46 15.43 3.61 15.47 3.54 15.43 3.72 15.62 3.83
Breakup 2.72 1.10 1.87 1.02 1.78 0.97 1.73 1.00 1.79 0.98
Positive
Communication 17.43 4.23 21.55 4.20 21.33 4.57 21.32 4.90 21.80 4.96
IPV 0.25 0.44 0.14 0.35 0.16 0.37 0.17 0.38 0.18 0.38
Individual
Psych Distress 16.08 5.63 12.55 5.66 12.19 5.41 12.42 5.91 11.73 5.55
Perceived Stress 8.40 3.12 6.29 3.34 6.46 3.31 6.25 3.47 5.89 3.71
Anger 14.61 4.69 12.25 4.61 12.35 4.64 12.06 4.79 11.66 4.71
Alcohol Use 5.26 6.05 4.24 4.96 4.13 4.83 4.22 5.45 4.20 5.16
Health 17.65 4.29 18.26 4.07 18.52 4.01 18.72 4.14 18.81 4.26
Insomnia 10.71 6.22 8.21 5.96 8.57 6.03 8.71 6.44 8.59 6.36
Note: Means and standard deviations are averaged across the 10 imputed datasets for pre to one-
year follow-up. Ns = 3320 for OurRelationship, Ns= 3360 for ePREP. Ns = 1,484 for full sample
at Pre mean and standard deviation. Means for IPV represent the proportion of respondents
endorsing IPV.
