A randomized clinical trial of online-biblio relationship education for expectant couples.
ABSTRACT This study evaluated the efficacy of a self-paced, Internet-based marriage and relationship skills education program ("Power of Two Online"). The program integrated an online intervention with print supplemental resources. New and expectant parents (n = 79) were randomly assigned to the 2-month intervention or placebo-control group. Assessments were conducted at baseline, 1-month, and 2-month follow-up intervals. Latent growth curve modeling was used to examine differences between conditions for marital satisfaction and conflict management. Participants who received the intervention reported trajectories of improved marital satisfaction and improved marital conflict management over time relative to controls. Implications for widespread dissemination of marriage and relationship education as a primary prevention tool are discussed.
- [Show abstract] [Hide abstract]
ABSTRACT: The authors discuss findings pertaining to male pornography use and female partners' relationship satisfaction and distress. The authors investigated honesty regarding pornography use and mutual consumption between partners, along with honesty and mutual use as predictors of satisfaction. Female participants (N = 340) in committed relationships completed the Pornography Distress Scale and Couples Satisfaction Index online. Participants reporting more honesty showed higher satisfaction and lower levels of distress, and participants disclosing mutual use showed lower levels of distress, although no differences were reported in satisfaction. Honesty regarding pornography use significantly predicted relationship dissatisfaction. Directions for future research and counseling implications are discussed.Journal of Sex and Marital Therapy 06/2013; · 1.27 Impact Factor
A Randomized Clinical Trial of Online–Biblio Relationship Education for
Christina J. Kalinka
University of Colorado, Denver
Frank D. Fincham
The Florida State University
Abigail H. Hirsch
TherapyHelp, Inc., Denver, Colorado
This study evaluated the efficacy of a self-paced, Internet-based marriage and relationship skills education
program (“Power of Two Online”). The program integrated an online intervention with print supplemental
resources. New and expectant parents (n ? 79) were randomly assigned to the 2-month intervention or
placebo-control group. Assessments were conducted at baseline, 1-month, and 2-month follow-up intervals.
conflict management. Participants who received the intervention reported trajectories of improved marital
satisfaction and improved marital conflict management over time relative to controls. Implications for
widespread dissemination of marriage and relationship education as a primary prevention tool are discussed.
Keywords: computer-based interventions, prevention, relationship education, marriage satisfaction
Supplemental materials: http://dx.doi.org/10.1037/a0026398.supp
Marriage and relationship education (MRE) has been widely
adopted as a tool to help couples create happy, healthy rela-
tionships and to decrease divorce rates. Recent meta-analyses
show that MRE increases marital satisfaction and communica-
tion skills while also reducing aggression between partners and
lowering divorce rates (Hawkins, Blanchard, Baldwin, & Faw-
cett, 2008; Blanchard, Hawkins, Baldwin, & Fawcett, 2009). As
the benefit of MRE services becomes clear, calls for ways to
provide a more diverse demographic of couples with empiri-
cally based services in a broader spectrum of settings have
grown (Braithwaite & Fincham, 2011). In light of the need for
new delivery mechanisms and with the explosion of Internet use
and capabilities online resources emerge as a logical vehicle for
MRE delivery. Existing trials of online MRE appear promising,
though they have used only the most basic of technological
advances (e.g., Braithwaite & Fincham, 2011; Duncan, Steed, &
MRE is also especially promising as a vehicle for primary
prevention during high-risk periods. Given the well-documented
stresses and declines in marital satisfaction associated with the
transition to parenthood and birth of a child, we have chosen to
target this intervention to the pre- and postpartum period (e.g.,
Doss, Rhoades, Stanley, & Markman, 2009). Meta-analysis sug-
gests that in-person MRE during this period is able to create
positive impact on marital communication (Pinquart & Teubert,
In order to increase the reach of empirically validated relationship
education, research done outside the laboratory is critical (Markman
et al., 2004). Effectiveness trials, like this study, which examine
treatment outcomes under the more practical conditions of everyday
implementation, are essential for understanding how a program works
in real-world conditions. Because we wanted to evaluate the potential
of online MRE to function as a broad-based primary prevention
mechanism, we recruited a community-based sample of new and
expectant parents who utilized the intervention from the privacy and
comfort of their own homes and chose when and how much they used
the program. Although such an approach could limit the magnitude of
the program’s effect, the information gleaned from testing the impact
of the program in the real conditions that any Internet-based MRE
would have to function is critical.
In summary, we believe a self-paced, skills-based, Internet-
delivered MRE program for new and expectant parents will create
a scalable and accessible vehicle for leveraging MRE’s primary-
prevention potential, enabling participants to counter the potential
slide into declining marital satisfaction over the pre- and postpreg-
Christina J. Kalinka, Department of Clinical Psychology, University of
Colorado Denver; Frank D. Fincham, Family Institute, The Florida State
University; Abigail H. Hirsch, TherapyHelp, Inc., Denver, Colorado.
This research was funded by the United States Department of Health and
Human Services, Administration for Children and Families, Grant 90-FE-
Abigail H. Hirsch is an owner of the Power of Two Online program.
Correspondence concerning this article should be addressed to Christina
J. Kalinka, Department of Clinical Psychology, University of Colorado
Denver, Campus Box 173, 1200 Larimer Street, Room 5005D, Denver, CO
80217-3364. E-mail: firstname.lastname@example.org
Journal of Family Psychology
2012, Vol. 26, No. 1, 159–164
© 2012 American Psychological Association
0893-3200/12/$12.00 DOI: 10.1037/a0026398
Participants and Procedure
We recruited a community sample of heterosexual individuals in
committed relationships at metropolitan obstetrics/gynecology of-
fices and via Internet advertising. Participants were required to be
trying to conceive, currently pregnant, or parenting an infant under
age one. They also needed to have no major mental health con-
cerns, be proficient in English, and be interested in receiving an
online relationship skills intervention in accordance with our ap-
proved Western Institutional Review Board protocol. All assess-
ment, including informed consent, was conducted entirely online.
Participants were randomly assigned to the intervention condi-
tion or a placebo-control condition using a computer-generated,
randomization list (see Figure 1 for subject flow). At baseline
measurement (pretreatment), there were no significant between
group differences on demographic or independent variables (i.e.,
age, gender, education, race/ethnicity, relationship status, time in
relationship, living arrangement, number of children, pregnancy
phase), or dependent variables (i.e., relationship satisfaction, con-
flict resolution, depression) when comparing subjects assigned to
the intervention condition with those assigned in the placebo-
In addition, there were no significant differences at baseline
measurement on demographic independent variables or dependent
variables when comparing subjects that completed the study to
those who were lost during follow-up, both within their respective
condition (i.e., comparing the specific condition’s completers to
the specific condition’s drop outs) and across all conditions (i.e.,
comparing everyone who completed the study to all those who
dropped out of the study). We also computed change from baseline
to 1 month and found no substantial differences between those
who dropped out with those who completed the study.
The intervention condition received access to an interactive
MRE intervention teaching broad relationship skills (“Power of
Two Online”) and associated resource materials after completing
the baseline online assessment. Power of Two Online consisted of
12 different modules grouped into five areas: Overview, Commu-
nication, Emotion Regulation, Decision Making, and Positivity
and Intimacy. See Table 1 for more details on the types of skills
taught in each unit. Each module had 8 to 12 specific activities that
could be completed in any order, such as interactive flash games,
videos, real-world challenges, and assignments related to sections
from the Power of Two print and DVD resources also sent to each
participant (Heitler, 1997; Heitler & Hirsch, 2003, 2006). More
details on the intervention are provided in the online supplemental
materials. During the first active intervention month, participants
received biweekly reminders to use the online materials via their
choice of text-message or e-mail. After this, they receive biweekly
generic messages expressing appreciation for participation or shar-
ing generic relationship statistics.
The control group was asked to complete the same online
assessments as the intervention condition. Participants in the con-
trol condition were told that they would be participating in a study
of relationship education and would receive regular communica-
tion from our team with information about marriage as well as
when they would be asked to take assessments. During this period,
they received the same kinds of generic messages that the inter-
vention group received during their second month. After the study
was completed, control group participants were offered the online
Participants were generally in their late twenties (N ? 79; mean
[M]age? 28 years, standard deviation [SD]age? 5.75 years) and
female (85.7%). The ethnic breakdown of participants was as
follows: 84.3% White (non-Hispanic), 7.1% Hispanic, 5.7% bira-
cial, and 2.9% Black. Participants were relatively diverse in terms
of educational background—approximately half held a bachelor’s
degree or higher (51.5%), 47.1% completed high school, and 1.4%
reported that they did not have a high school education. Partici-
pants mostly were having their first or second child (42.9% and
40.0%, respectively), and were married (72.9%). On average,
subjects had been with their partner/spouse for 5.41 years (SD ?
3.39 years). Of the 27.1% of participants who reported that they
were not currently married, 78.9% said they were currently cohab-
Randomized (n = 118)
Assessed for eligibility (n = 151)
Excluded for not meeting inclusion criteria (n = 33)
Other condition not used in this study (n = 39)
Placebo Control (n = 34)
Number of Completed Assessments:
Baseline - Time 1: 31
One month - Time 2: 28
Two month - Time 3: 27
Reasons for dropout: Did not respond to
our multiple attempts to contact
Number of Completed Assessments:
Baseline - Time 1: 39
One month - Time 2: 29
Two month - Time 3: 21
Reasons for dropout: Did not respond to
our multiple attempts to contact
Power of Two Online Intervention (n = 45)
clinical trial. Please note that the other condition not used in this study was
a 1-month waitlist control group without follow-up data, therefore not
appropriate for the latent growth curve model used in our analyses.
Sampling and flow of participants through the randomized
Power of Two Online Intervention Skills
Topic areaPower of two skills covered
Overview 1. Why marriage skills?
2. Does marriage matter?
3. Introduction to skill areas
7. Understanding anger
8. Managing anger
9. Decision making patterns
10. Win–win decision making
11. Being a supportive spouse
12. Building intimacy
13. Maintaining safety
Positivity and intimacy
For more details about the skills covered in the Power of Two
MRE Intervention,see http://www.PowerofTwo
KALINKA, FINCHAM, AND HIRSCH
iting with their significant other. Although including participants
from across the pregnancy/young baby demographic might lessen
the program’s overall effect, in the spirit of an effectiveness trial,
we opted for broad inclusion criteria. Only one member of a couple
was allowed to complete study surveys, although subjects were
encouraged to share materials with their spouse/partner.
Participants completed the Ineffective Arguing Inventory
(IAI: Kurdek, 1994), which is a measure of conflict resolution;
the Couples Satisfaction Index (CSI: Funk & Rogge, 2007),
which is a measure of relationship satisfaction; and the Center
for Epidemiologic Studies Depression scale (CESD; Radloff,
1977), which is a measure of depressive symptom severity.
Coefficient alphas for these measures in the present sample
were .95 for the IAI, .95 for the CSI, and .86 for the CESD.
Means and standard deviations at baseline, 1 month and 2
month follow-ups are presented in Table 2.
Latent growth curve (LGC) analysis was used to examine the
data. Please note all references related to statistical methodology
are provided in the Appendix. LGC represents repeated measures
of a given concept as a function of time. Each time point mea-
surement is an indicator of two latent growth factors, initial status
and change of slope, on which individuals may vary. Because
initial status is similar to the intercept in a regression equation, the
unstandardized loadings of all indicators (the three time points for
each subscale) on initial status were fixed to one. To specify a
linear trend, the loading of T1 on slope was fixed to 0, T2 was
fixed at 1, and T3 was fixed at 2; this is analogous to centering in
hierarchical linear modeling.
Because postpartum depression impacts 10–15% of the post-
partum population, is likely to correlate with marital satisfaction/
functioning and would likely vary over the course of our interven-
tion, we chose to include depression scores as a time varying
covariate (Robertson, Grace, Wallington, & Stewart, 2004).
Condition was included in the model as a fixed covariate, and
the variances from the latent variables (intercept and slope) were
allowed to covary; thus, providing information about the relation-
ship between initial levels of the dependent variable and the rate of
change across time. Specifying the model as described provides
information about the impact of the intervention on change or
“growth” over time. By using maximum likelihood estimation,
missing data could be handled so that a growth curve could be fit
for all respondents even if they did not have complete data.
When ineffective arguing was entered into the model, a negative
variance contributed to a nonpositive definite matrix (a Heywood
case). Per standard practice, the offending variance was set to zero
and the model was rerun. The corrected model provided a good fit
to the data that was not significantly different from the uncorrected
model, ?2(9) ? 10.79, p ? .29, comparative fit index (CFI) ? .99,
Tucker Lewis Index (TLI) ? .87; root mean square error approx-
imation (RMSEA) ? 0.054. In this model, condition predicted
slope such that those who received the intervention showed more
effective arguing over time relative to the control group, (? ?
?.40, p ? .005, Hedges’ g ? .42). Slope and intercept were
significantly inversely associated (r ? ?.39, p ? .05) with one
another, suggesting that higher initial levels of ineffective arguing
were associated with a lower rate of change in ineffective arguing
over time. See Figure 2 for path diagram.
When scores for the CSI were entered into the model, a negative
variance was generated, thus the offending estimate was set to 0
and the model was rerun. The corrected model provided an excel-
lent fit to the data that was not significantly different from the
uncorrected model, ?2(9) ? 4.84, p ? .84, CFI ? 1.00, TLI ?
1.07; RMSEA ? 0.00. In this model, condition only predicted
slope such that those who received the intervention had greater
satisfaction over time relative to the control group (? ? .420, p ?
.005, Hedges’ g ? .24).
Slope and intercept were not significantly associated, suggesting
that initial levels of relationship satisfaction were not associated
with the rate of change in satisfaction over time. See Figure 3 for
This effectiveness study investigated the impact of a self-paced,
skills-based, Internet-delivered MRE program administered to new
and expectant parents. The Power of Two Online appears to be
Couples Satisfaction Index and Ineffective Arguing Inventory Means and Standard Deviations
Baseline 1-month F/U2-month F/U
scores ? higher relationship satisfaction. Higher IAI scores ? higher level of ineffective arguing skills.
CSI ? Couples Satisfaction Index; IAI ? Ineffective Arguing Inventory. F/U ? follow-up. Higher CSI
SELF-PACED ONLINE RELATIONSHIP EDUCATION
viable as a primary prevention tool for preventing relationship
decline during the period around the birth of a child. Specifically,
we found that individuals in the intervention condition showed
more productive management of conflict situations and experi-
enced less of a decline in relationship satisfaction relative to
These findings are particularly striking when looked at in com-
parison with other studies presented in the meta-analysis of in-
person MRE for expectant parents. Though our observed effect
sizes both fall in the small range (Hedges g of .2 to .49), they are
also comparable to far more intensive and expensive workshop-
based interventions. For example, Pinquart and Teubert (2010)
found average effect sizes of around .3 for impact on couple
adjustment and communications for interventions with six or more
in person sessions. This work finds effect sizes of a similar
magnitude for this self-paced, online–bibliotherapy intervention.
Limitations and Future Directions
Our study has several limitations. First, it focused on partici-
pants at a unique life-stage transition. Although having more
resources in the empirically validated toolbox for expectant par-
ents is most certainly an asset, research extending this online MRE
intervention to a wider cross-section of couples would be benefi-
cial in determining its overall scope of effectiveness.
Second, it would have been ideal to have a long-term follow-up.
Future work investigating long-term impact and best practices for
increasing duration of gains from an online MRE resource would
also be advantageous.
A third limitation in this work concerns questions raised by the
mixed Internet and print materials provided to participants. Be-
cause data are not available on the degree to which participants
used the print versus the Internet materials, we were not able to
determine the relative impact of the two aspects of the interven-
tion. In addition, the literature on the efficacy of bibliotherapy for
MRE is also strikingly lacking—to the best of our knowledge, the
current literature includes no studies that directly address this
question. At the same time, there is some evidence from a related
domain, parenting skills, that self-paced bibliotherapy does appear
effective (Sanders, Markie-Dadds, Tully, & Bor, 2000), suggesting
that this aspect of our combined Internet–bibliotherapy interven-
tion may have been a potent contributor to the overall demon-
Two lines of future work would be merited here. First, would be
to tease out the relative impact of the print versus Internet mate-
rials—especially, as more of the print content is made available in
a digital format. Findings by Greg et al. (1999) about limitations of
self-directed bibliotherapy for treating panic attacks as compared
to bibliotherapy combined with periodic check-ins with a clinician
suggest a second direction for future work—that the programs
overall efficacy might be enhanced if a light-touch personalized
check-in could be added to the program.
Fourth, the relatively high attrition levels, especially in the
intervention condition are less than ideal. It is certainly possible
that attrition in the intervention condition was driven by people not
finding the intervention helpful, whereas similar people did not
drop out from the control condition. This would result in the
appearance of more impact than was actually created and the
findings of this initial report should be read with some caution in
light of this possibility. While such loss is likely normal for any
real-world application, future work should look both to identify
ways to reduce attrition and equally important must be designed to
test if attrition is truly at random or has systematic bias.
Because this was an effectiveness study, we did not have the
level of experimental control of a laboratory-based intervention.
Unfortunately, we know nothing about what materials participants
focused on most and how regularly they used the materials, mak-
ing it impossible to distinguish which components of the interven-
tion were the most helpful (e.g., online course components, video,
supplemental print resources). Future work should address more
specifically best practices for Internet-based MRE. How much
material and at what time periods? How important is repeated
exposure over time? What types of materials create the most robust
changes? The possible digital MRE delivery formats are numerous
and rapidly expanding, so it is crucial that we increase our sophis-
IAI (time 1) IAI (time 2) IAI (time 3)
Ineffective Arguing Inventory (IAI).
CSI (time 1) CSI (time 2) CSI (time 3)
Couples Satisfaction Index (CSI) model diagram.
KALINKA, FINCHAM, AND HIRSCH
tication about the strengths and limitations of each type of media
on its own and in conjunction with other media types.
Likewise, data gathered in the study also does not include
information about dosage rates for each participant. While our
qualitative impressions based on e-mails with participants prior to
sending them compensation checks left our team with the belief
that dosage rates varied considerably, we are unable to report on
connections between dosage and impact. It is possible that a
just-in-time approach like this can create considerable impact by
providing even just a very brief dose of learning on a highly
relevant topic. At the same time, it may be that the program’s
impact comes from the sustained, consistent presence that is pos-
sible with an online–biblio approach. Sorting out questions about
optimal dosage would most certainly be an essential component
for future work.
Implications and Conclusions
Our findings support the premise that it is possible to use the
Internet and supplemental print materials to deliver effective MRE.
Specifically, this work indicates that the characteristics of tradi-
tional MRE, such as live teachers and group settings, can be offset
through the use of engaging and interactive digital and print
self-study materials. This study provides a hopeful chord in the
search for a way to make primary prevention MRE available to all
interested individuals and couples. The scalability and cost-
effectiveness of such an approach are most exciting—a course like
Power of Two Online can be delivered for cents per user and is not
constrained by geography. Likewise, the potential to move MRE
into the online world so central in most young couples’ lives is also
encouraging. Indeed, as society becomes increasingly tied to our
digital existence, the future of MRE may depend on our ability to
successfully transition these tools into the digital sphere.
Blanchard, V. L., Hawkins, A. J., Baldwin, S. A., & Fawcett, E. B. (2009).
Investigating the effects of marriage and relationship education on
couples’ communication skills: A meta-analytic study. [Review]. Jour-
nal of Family Psychology, 23, 203–214. doi:10.1037/a0015211
Braithwaite, S. R., & Fincham, F. D. (2011). Computer-based dissemina-
tion: A randomized clinical trial of ePREP using the actor partner
interdependence model. Behaviour Research and Therapy, 49, 126–131.
Doss, B. D., Rhoades, G. K., Stanley, S. M., & Markman, H. J. (2009). The
effect of the transition to parenthood on relationship quality: An 8-year
prospective study. Journal of Personality and Social Psychology, 96,
Duncan, S. F., Steed, A., & Needham, C. M. (2009). A comparison
evaluation study of web-based and traditional marriage and relationship
education. Journal of Couple & Relationship Therapy, 8, 162–180.
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,
Greg A. R., Febbraro, G. C., Roodman, A., & Wright, J. H. (1999). The
limits of bibliotherapy: A study of the differential effectiveness of
self-administered interventions in individuals with panic attacks. Behav-
ior Therapy, 30, 209–222. doi:10.1016/S0005-7894(99)80004-5
Hawkins, A. J., Blanchard, V. L., Baldwin, S. A., & Fawcett, E. B. (2008).
Does marriage and relationship education work? A meta-analytic study.
Journal of Consulting and Clinical Psychology, 76, 723–734. doi:
Heitler, S., & Hirsch, A. (2003). The power of two workbook: Communi-
cation skills for a strong and loving marriage. Oakland, CA: New
Heitler, S., & Hirsch, A. (2006). The win-win waltz [DVD]. Denver, CO:
Heitler, S. (1997). The power of two: Secrets of a strong and loving
marriage. Oakland, CA: New Harbinger Publications.
Kurdek, L. A. (1994). Conflict resolution styles in gay, lesbian, heterosex-
ual nonparent, and heterosexual parent couples. Journal of Marriage and
the Family, 56, 705–722. doi:10.2307/352880
Markman, H. J., Whitton, S. W., Kline, G. H., Stanley, S. M., Thompson,
H., St Peters, M., . . . Cordova, A. (2004). Use of an empirically based
marriage education program by religious organizations: Results of a
dissemination trial. Family Relations, 53, 504–512. doi:10.1111/j.0197-
Pinquart, M., & Teubert, D. (2010). A meta-analytic study of couple
interventions during the transition to parenthood. Family Relations, 59,
Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale
for research in the general population. Journal of Applied Psycho-
Robertson, E., Grace, S., Wallington, T., & Stewart, D. (2004). Antenatal
risk factors for postpartum depression: A synthesis of recent literature.
General Hospital Psychiatry, 26, 289–295. doi:10.1016/j.genhosp-
Sanders, M., Markie–Dadds, C., Tully, L., & Bor, W. (2000). The triple
p-positive parenting program: A comparison of enhanced, standard, and
self-directed behavioral family intervention for parents of children with
early onset conduct problems. Journal of Consulting and Clinical Psy-
chology, 68, 624–640. doi:10.1037/0022-006X.68.4.624
SELF-PACED ONLINE RELATIONSHIP EDUCATION
Additional Statistical References
Dillon, W. R., Kumar, A., & Mulani, N. (1987). Offending estimates in
covariance structure analysis: Comments on the causes of and solutions
to Heywood cases. Psychological Bulletin, 101, 126–135. doi:10.1037//
Hancock, G. R., & Lawrence, F. R. (2006). Using latent growth models to
evaluate longitudinal change. In G. R. Hancock & R. O. Mueller (Eds.),
Structural equation modeling: A second course (pp. 171–197). Green-
wich, CT: Information Age Publishing.
Kline, R. B. (2005). Principles and practices of structural equation mod-
eling (2nd ed.). New York, NY: Guilford Press.
Little, R. J., & Rubin, D. B. (2002). Statistical analysis with missing data
(2nd ed.). New York, NY: Wiley.
van Driel, O. P. (1978). On various causes of improper solutions in
maximum likelihood factor analysis. Psychometrika, 43, 225–243.
Willett, J. B., & Sayer, A. G. (1994). Using covariance structure analysis
to detect correlates and predictors of individual change over time.
Psychological Bulletin, 116, 363–381. doi:10.1037/0033-2909
Received February 2, 2011
Revision received October 12, 2011
Accepted October 20, 2011 ?
KALINKA, FINCHAM, AND HIRSCH