MALLORY LUCIER-GREER Auburn University

FRANCESCA ADLER-BAEDER Auburn University*

Does Couple and Relationship Education Work

for Individuals in Stepfamilies? A Meta-analytic

Study

Recent meta-analytic efforts have documentedhow couple and relationship education (CRE)programs promote healthy relationship andfamily functioning. The current meta-analysiscontributes to this body of literature by exam-ining stepfamily couples, an at-risk, subpop-ulation of participants, and assessing theeffectiveness of CRE programs for individualsin these relationships. Findings are aggregatedfrom 14 evaluation studies of CRE programsdesigned for stepcouples. Analyses examineeffects by study design (i.e., comparison-groupand one-group/pre-post) and within specificdomains (i.e., family, parental, and couple func-tioning). The interventions have, on average,small effects overall (comparison-group: d =.20; one-group/pre-post: d = .23), and slightlylarger effects in family and parental function-ing when examining specific outcomes (d = .20to .35). An examination of effectiveness overtime reveals that earlier studies (i.e., 1980sand 1990s) show larger effects than more recentstudies. Implications for CRE programs for step-families and evaluation needs are discussed.

Family life education programs date back tothe 1800s, when the U.S. government recognized

Current address: 123 Dawson Hall, The University ofGeorgia, Athens, GA 30602 (mlgreer@uga.edu).*Department of Human Development and Family Studies,Auburn University, 203 Spidle Hall, Auburn, AL 36849.

Key Words: couple and relationship education, meta-analysis, stepfamily couples.

the value in providing families with resourcesto enhance the home environment as a meansto strengthen the community and well-being ofsociety (Duncan & Goddard, 2005). Programsfocused on parenting, financial management,and nutrition have developed through the Coop-erative Extension System, general university-based outreach, community movements, andgrassroots organizations (Duncan & Goddard,2005). It was not until recently, however, thatthe Administration for Children and Families(ACF), a federal agency in the U.S. Departmentof Health and Human Services responsible forprograms that enhance the economic and socialwell-being of communities, earmarked compet-itive funds to specifically support community-based couple and relationship education (CRE)programs. These funds were provided in an effortto promote healthy relationships and marriagesthrough skills-based education (U.S. Departmentof Health and Human Services, 2010).

Early on, researchers noted the importanceof documenting the effects of family life pro-grams in general (Lehman, 1947) and CREprograms specifically. Individual studies pro-vide indications of how educational programsaffect the well-being and functioning of partic-ipants, yet questions remain about the overallutility of programs such as CRE across studies.Meta-analytic procedures provide researchersa tool to systematically integrate the findingsfrom individual quantitative, evaluation studies,allowing researchers to assess overall effective-ness (Borenstein, Hedges, Higgins, & Roth-stein, 2009). It also provides study population

756 Family Relations 61 (December 2012): 756 – 769DOI:10.1111/j.1741-3729.2012.00728.x

CRE for Individuals in Stepfamilies 757

effect size(s) against which current and futureindividual evaluation studies can compare themagnitude of their results.

One of the earliest meta-analytic efforts ofCRE examined 85 premarital, marital, and fam-ily enrichment programs and found positivemoderate effect sizes (d = .42 to .55) whenassessing measures of relationship satisfac-tion, relational skills, and individual functioning(Giblin, Sprenkle, & Sheehan, 1985). It was alsonoted that programs with longer interventionshad stronger effects. Using meta-analytic strate-gies, Butler and Wampler (1999) found thatacross 16 studies of CRE, couple communica-tion and satisfaction improved immediately afterengaging in the program (d = .21 to .74), yeteffects were not sustained at follow-up. Carrolland Doherty (2003) used meta-analysis to exam-ine 23 published evaluation studies of premaritalprograms and concluded that premarital preven-tion programs are generally effective in produc-ing gains in interpersonal skills and relationshipquality with a large mean effect size (d = .8).Reardon-Anderson, Stagner, Macomber, andMurray (2004) conducted a meta-analysis onthe effectiveness of 39 diverse types of marriageprograms. Findings described in their techni-cal report suggested that program participantsreported moderate improvements in relationshipsatisfaction and small improvements in commu-nication. Additionally, therapy type programsproduced large effect sizes (d = .8), whereaseducation and enrichment programs producedsmall to moderate effect sizes (d = .2 to .5).

Recently, a flurry of ‘‘second generation’’meta-analytic efforts have aimed to furtherassess CRE program effectiveness by examin-ing diverse outcomes and including unpublishedstudies. Hawkins, Blanchard, Baldwin, andFawcett (2008) examined 117 studies to under-stand how characteristics of the participantsinfluenced program effects. They found that par-ticipating in CRE programs strengthened rela-tionship quality (d = .30 to .36) and communi-cation (d = .43 to .45) but did not find evidenceof differential effects by race, economic status,or sex of the participant. Blanchard, Hawkins,Baldwin, and Fawcett (2009) conducted a meta-analysis to compare program effects on com-munication for couples with different levelsof distress using observational and self-reportmethods. Results from this meta-analysis of 97studies suggest that CRE positively influencescommunication for distressed and nondistressed

couples (d = .23 to 1.57) with larger effectsdocumented from observational assessments.A follow-up analysis of premarital programsby Fawcett, Hawkins, Blanchard, and Car-roll (2010) included published and unpublishedstudies (47 total). Unlike Carroll and Doherty(2003), mixed results for effectiveness sur-faced; there was no direct effect on relationalquality, but a modest effect on communication(d = .45 to .54).

This next step in meta-analytic studiesincludes the exploration of subpopulations tounderstand how at-risk or high-need groupsbenefit from CRE programs. Hawkins andFackrell (2010) examined the effectiveness ofCRE for lower income couples; findings across15 studies suggest that these programs producedeffect sizes fairly similar to those of previousmeta-analyses using middle-income participants(d = .25 to .29). Pinquart and Teubert (2010)examined 21 CRE programs for parents duringthe transition to parenthood; small effectswere found for individual functioning andcommunication (d = .21 to .28) with strongereffects emerging for programs that were 5+sessions, taught by professionals, and includedboth prenatal and postnatal components.

To contribute to this evolving literature, thecurrent study is a first attempt to investigatethe effectiveness of CRE programs for anothersubpopulation of participants identified byACF as high need and as a population thatwould benefit from specialized educationalprogramming: couples who form stepfamiliesor ‘‘stepcouples’’ (Adler-Baeder, Robertson,& Schramm, 2010). Recently published meta-analytic efforts may have included stepcouplesas they examined broad programmatic effectsand even in their examination of effects ofgeneralized CRE; however, there has been noattention given to meta-analyzing evaluationstudies of CRE programs designed specificallyto address the needs and stressors of stepcouples.

Unique stressors for stepcouples not associ-ated with first marital relationships often lead tohigher rates of marital instability and greaterrisks of marital dissolution (Adler-Baeder,Robertson, et al., 2010; Coleman, Ganong, &Fine, 2000). Estimates suggest that second andhigher order marriages are often less stable thanfirst marriages, such that 40% of remarriages endin divorce within a decade compared to 32% offirst marriages (Bumpass & Raley, 2007). Thesehigher rates of instability and dissolution are

758 Family Relations

noteworthy given that between one third and onehalf of all marriages each year are remarriagesfor at least one partner (Kreider & Ellis, 2011).In addition, an increasing number of first mar-riages form stepfamilies because of nonmaritalbirths; thus, these families are faced with similarchallenges (Sweeney, 2010). As such, the needfor effective programming is highly relevantand, perhaps, key to increasing stepfamily sta-bility (Halford, 2004). Recommendations fromthe empirical literature suggest that in order tomeet the needs of stepcouples, CRE programsshould include basic relationship skills as wellas specialized information pertaining to parent-ing stepchildren, negotiating stepfamily rolesand rules, navigating relationships with formerpartners, debunking myths, and gaining realis-tic expectations about stepfamily development(Adler-Baeder, Robertson, et al., 2010). Effortshave been made to design classes that focus onthese specific needs related to stepcouple func-tioning (Adler-Baeder & Higginbotham, 2004).The studies in this meta-analysis used curric-ula with similar program content and consistentwith the design recommended for CRE for step-couples (Adler-Baeder, Robertson, et al., 2010).These classes have been implemented and docu-mented since the late 1970s (Messinger, Walker,& Freeman, 1978) and the prevalence of classesfor stepfamilies has grown in part because ofincreased federal funding. Yet, we are still in theearly stages of examining the effectiveness ofthese programs.

In a recent literature review on stepfamilyprograms, Whitton, Nicholson, and Markman(2008) examined evaluation studies of pro-grams designed for stepcouples by describinginformation on the context, content, and find-ings of 20 studies from the extant literature.They reported that interventions appeared wellgrounded in empirical literature and predom-inantly used educational models targeted atprevention. The researchers reported challengesin drawing conclusions on the effectiveness ofprograms because of the limited number of avail-able evaluation studies, relatively small samplesizes in existing studies, and diverse evalua-tion measures used across studies. They offera ‘‘vote count’’ method for summarizing pro-gram effectiveness by listing whether findingswere significant or not for each indicator bystudy. Their review provides a meaningful ini-tial step by describing evaluative work on CREfor stepcouples through 2007, but it does not

provide comprehensive information on programeffectiveness or an understanding of the magni-tude of effects or information on the variationin effects based on study context and outcomefocus.

Current Study

Meta-analysis is designed to address theconcerns presented by Whitton et al. (2008)pertaining to the amount of available data, thesample sizes of studies, and diverse evaluationmeasures by utilizing appropriate methods foraggregating quantitative data across relevantstudies and incorporating study features inanalyses (Borenstein et al., 2009). Through theuse of meta-analytic techniques, this studyextends recent efforts to meta-analyze CREfor subpopulations and efforts to organizeinformation on stepfamily programs. The focuson individuals in stepcouple relationships isparticularly relevant given the greater risk ofmarital dissolution and high prevalence rateof stepcouples. Despite over three decades ofprogram implementation, evaluation studies ofCRE for stepcouples have only now accumulatedto a point at which a meaningful meta-analysiscan be conducted. Early studies of stepfamilyprograms relied heavily on field notes andobserver judgment of participants’ satisfactionwith the course (e.g., Messinger et al., 1978; Pill,1981). Since that time, methodological advanceshave been implemented, moving to pre-post,quasi-experimental, and experimental designsthat utilize standardized family science measuresas indicators of individual and relationalfunctioning (e.g., Gellat, Adler-Baeder, &Seeley, 2010). These advances have resultedin a small but adequate sample of studies thatnow allow for the aggregation of findings. Thisstudy is an initial attempt to move beyond votecounting methods and conceptual syntheses tometa-analyze the effectiveness of CRE programsfor stepcouples.

Research Questions

The following research questions were ad-dressed:

RQ1: How effective is CRE that targets individ-uals in stepcouple relationships? For this ques-tion, effectiveness was broadly defined (similar torecent studies; e.g., Fawcett et al., 2010; Holmes,

CRE for Individuals in Stepfamilies 759

Galovan, Yoshida, & Hawkins, 2010). Variablesrelated to family, parental, couple, and individualfunctioning were standardized and aggregated togauge the overall sense of functioning reported byparticipants prior to and following participation ina stepfamily focused CRE program.

RQ2: How effective is CRE that targets individ-uals in stepcouple relationships in specific targetoutcome domains, including family functioning,parenting, couple quality, and individual func-tioning? This second set of analyses utilized anuanced perspective on effectiveness, as severalrelevant outcomes of stepfamily functioning weredistinguished and examined separately.

RQ3: Have CRE programs that target stepcouplesimproved in their effectiveness over time? Acumulative effects procedure was conducted toassess how the effectiveness of stepfamily-specific CRE programs has changed or evolvedover time. For this meta-analysis, stepfamilyprogram evaluation studies span three decadesof research ranging from 1982 to 2011. Thestudies in the current meta-analysis do not usedifferent iterations of the same program, yetthe assumption is that more recent programsbenefit from recent research and theory relatedto stepfamily functioning. A review of CREprograms for stepcouples illustrates that, although‘‘core’’ content is similar, earlier programs wereoften more narrow in focus, whereas recentprograms tend to use a more comprehensive lensin terms of content, explicit theoretical base, andoutcome goals (Adler-Baeder & Higginbotham,2004). Through the use of available evaluationstudies, this research question examined whetherprograms designed for stepcouples improved oversociohistorical time.

METHOD

Meta-analytic techniques provide transparentand objective tools that allow researchers tomove beyond descriptive synthesizing by quan-titatively aggregating outcome data from indi-vidual studies to produce an overall effect size.Effect sizes communicate the magnitude anddirection of the treatment by standardizing andintegrating outcome data, sample size informa-tion, and the variance of measured outcomesfrom studies that propose a similar set of researchhypotheses (Borenstein et al., 2009). Effect sizesare standardized in the metric of standard devia-tion units and in this study communicate the

impact of the CRE program on participantfunctioning. Each study contributes an individ-ual effect size, which represents the standardizedfindings for that study. The assumption is thatstudies are drawn from the same underlying pop-ulation—stepcouples in the United States—and,thus, can be statistically combined as indica-tors of the same phenomenon. The aggregationof individual effect sizes provides researcherswith a summary effect, which represents theweighted mean of the individual effects. In meta-analysis, random-effects models assume that thetrue effect size varies across studies based onstudy context; thus the summary effect repre-sents the distribution of effects and an estimateof the mean of these effects.

Search Procedure and Inclusion Criteria

The process of examining stepfamily programeffectiveness began by our conducting an exten-sive search for published and unpublishedintervention studies focused on stepfamilies.Stepfamilies were defined as a family unitcontaining a couple in which one or bothpartners had a child or children from a pre-vious relationship. This definition is descrip-tive of family dynamics and moves beyondclassifications by marital status. We utilizedthe book chapter by Whitton and colleagues(2008) as a starting point. Then, systematicsearches of electronic databases (e.g., EBSCO-Host, PsycINFO, Academic Search Premier,Psyc-ARTICLES, and ProQuest Dissertationand Theses database) were conducted usingcombinations of the following terms: stepfamily,evaluation, remarriage, education, intervention,and program. We also reviewed content-specificjournals (e.g., Family Relations and Journal ofDivorce & Remarriage) and reference pagesfrom relevant articles. Direct appeals were alsosent to relevant listservs (e.g., certified familylife educators list and extension human sci-ences network) and known program developers.Finally, we reviewed 226 abstracts of healthymarriage and fatherhood programs funded byACF and contacted sites that reported using astepfamily-specific CRE curriculum. The inclu-sion of unpublished studies avoids publicationbias, the assumption that only studies withsignificant effects are published (Borensteinet al., 2009).

Twenty-eight reports of studies evaluatinginterventions for stepfamilies were located.

760 Family Relations

Inclusion criteria for the current study were that(a) adult participants were part of a stepfamilycouple and programs utilized educational com-ponents—not therapeutic interventions—withstepfamily-specific content (4 studies excluded)and (b) quantitative pre-post data were col-lected on self-reported psychosocial measures(10 studies excluded). Studies that utilizeda comparison-group design (experimental andquasi-experimental) and those that utilized aone-group/pre-post design were included. Four-teen studies met all inclusion criteria (see Table 1for the complete list). Although small, this num-ber is adequate, as analyses can be conductedwith as few as three studies (Valentine, Pigott,& Rothstein, 2010). The sample size also isfairly consistent with recently published meta-analyses focused on family life education (e.g.,Hawkins & Fackrell, 2010; Holmes et al., 2010).

Coding Procedure

Studies were coded by the first author to assessthe inclusion criteria and to document avail-able data. A codebook was created to guidethe coding process. In general, classes weretaught over 4 to 8 weeks with an averageof 13 hours of program participation. Differentprograms were used across most of the stud-ies, but all programs addressed (a) normalizingthe stepfamily experience, (b) issues of par-enting children (yours, mine, and ours) andthe links to couple functioning, (c) positivecouple communication, and (d) couple conflictmanagement. For early studies that reportedparticipant demographics, samples were oftensmall (M = 32 participants) and homogeneousin terms of the racial makeup of participantsranging from 85% to 92% European Ameri-can. Samples after 2001 were larger (M = 833participants) and reported more racial hetero-geneity: 65% to 73% European American. Moststudies (71%) included only married coupleswith a broad range of years married. Four stud-ies, all produced after 2001, included cohabitingstepcouples. All participants had children, butdata were not available to consistently iden-tify child demographic information. It shouldbe noted that over half of the studies—mostlyearlier studies—did not include full sampledescriptions.

Measured outcome variables were dividedinto target outcome domains: family, parent-ing, couple, and individual functioning. Family

functioning was captured with measures of fam-ily cohesion, conflict, organization, adaptabil-ity, and harmony. Parenting was measured viainstruments assessing parental support, conflictover children, parenting practices, coparentingquality, and parental efficacy. The couple rela-tionship was examined using measures of adjust-ment, communication, commitment, trust, andstability. Individual functioning was measuredvia scales assessing depression, self-esteem,ability to cope, and individual empowerment(i.e., confidence in personal capabilities andintentionality in the use of personal strengths).Overall effectiveness was determined by aggre-gating these target outcomes. An expert onstepfamily education programs and evaluationreviewed half of the studies to assess codingreliability. Disagreements were resolved by dis-cussion. Any resulting changes in coding wereextended to the entire data set.

Each of the studies reported pretest data (i.e.,means, standard deviations, and sample size)prior to program participation and posttest dataimmediately after program completion. Programevaluators were contacted directly when thestudy report did not provide statistics neededfor the meta-analysis. When coding the one-group/pre-post studies, correlation assessmentsbetween the pretest and posttest were neededto calculate precise effect sizes. These wereavailable in most studies, but when they wereunavailable and could not be obtained, weestimated them to be .50, an estimate that doesnot bias the overall effect size (Hawkins &Fackrell, 2010; Nowak & Heinrichs, 2008).

Effect Size Computation

Summary effect sizes, which will be referredto simply as effect sizes, are calculateddifferently by study design. Effect sizes forcomparison-group design studies represent theoverall standardized mean difference betweenthe treatment and comparison group; effectsizes for one-group/pre-post studies representthe standardized mean gain score between thepretest and posttest. Because the effect sizesacross study design represent different typesof comparisons (Lipsey & Wilson, 2001), theanalyses in the current study were conductedseparately by study design.

Of the comparison-group studies, four utilizedan experimental design (randomized treatmentand control) and three studies utilized a

CRE for Individuals in Stepfamilies 761

Tab

le1.

Ove

rvie

wof

Incl

uded

Inte

rven

tion

Stud

ies:

Stud

yPr

ogra

mN

o.of

Res

earc

hPa

rtic

ipan

tsSt

udy

Des

ign

Out

com

es

Adl

er-B

aede

r&

Luc

ier-

Gre

er(2

011)

aSm

artS

teps

(Adl

er-B

aede

r,20

07)

orTo

geth

erW

eC

an(S

hire

r,20

09)

E=

539

C=

61Pr

e-po

stF,

P,C

,I

Bie

lenb

erg

(199

1)a

Stre

ngth

enin

gYo

urSt

epfa

mily

(Ein

stei

n&

Alb

ert,

1986

)15

Pre-

post

&8-

wee

kfo

llow

-up

F,P,

I

Bra

dy&

Am

bler

(198

2)a

Prog

ram

crea

ted

byau

thor

base

don

rese

arch

33Pr

e-po

stF

Cud

deby

(198

4)Pr

ogra

mcr

eate

dby

auth

orba

sed

onre

sear

ch24

Pre-

post

&4

wee

kqu

alita

tive

follo

w-u

p;co

ntro

lgro

upbu

tout

com

eda

tano

tre

port

ed

F,P

Gel

late

tal.

(201

0)a

Pare

ntin

gT

oolk

it:Sk

ills

for

Step

fam

ilies

E=

140

C=

148

Pre-

post

&8

wee

kfo

llow

-up;

rand

omiz

edco

ntro

lgro

upF,

P,C

Gib

bard

(199

8)E

ffect

ive

Step

pare

ntin

g26

Pre-

post

F,P,

CH

ende

rson

(200

1)St

reng

then

ing

Your

Step

fam

ilies

(Alb

ert&

Ein

stei

n,19

86)

E=

30C

=30

Pre-

post

;ran

dom

ized

cont

rolg

roup

F

Hig

bie

(199

4)E

mpo

wer

ing

Step

fam

ilies

Pro

gram

(n=

22);

OR

Supp

ortg

roup

form

at(n

=21

)E

=21

C=

22Pr

e-po

st&

6w

eek

follo

w-u

p;qu

asi-

expe

rim

enta

lcom

pari

son-

grou

pF,

C

Hig

ginb

otha

m(2

010)

aSm

artS

teps

(Adl

er-B

aede

r,20

07)

2,33

2Pr

e-po

st;6

-wee

kfo

llow

-up

F,P,

CN

elso

n&

Lev

ant(

1991

)aC

ombi

natio

nof

reco

mm

ende

dsk

ills

(lis

teni

ng,

self

-aw

aren

ess,

com

mun

icat

ion)

E=

14C

=20

Pre-

post

;qua

si-e

xper

imen

talc

ompa

riso

ngr

oup

F,P

Nic

hols

on&

Sand

ers

(199

9)a

Prog

ram

ofad

dres

sing

child

beha

vior

prob

lem

sin

rem

arri

edfa

mili

es(L

awto

n&

Sand

ers,

1992

)

E=

52C

=32

Pre-

post

&6-

wee

kfo

llow

-up;

rand

omiz

edco

ntro

lgro

upP

Ore

gon

Soci

alL

earn

ing

Cen

ter

(201

0)a

Pare

ntM

anag

emen

tTra

inin

gfr

omO

rego

n(P

MT

O)

E=

67C

=43

Pre-

post

&12

-mon

thfo

llow

-up;

rand

omiz

edco

ntro

lgro

upP,

C

Stro

up(1

982)

Mul

tiple

Fam

ilyG

roup

Tre

atm

entw

ithco

nten

tse

lect

eddu

ring

pilo

twor

k14

Pre-

post

F,C

Web

ber,

Shar

pley

,&R

owle

y(1

988)

aLi

ving

ina

Step

fam

ily(W

ebbe

r)20

Pre-

post

C,I

Not

e:E

=ex

peri

men

tal/t

reat

men

tgro

up;C

=co

ntro

l/com

pari

son

grou

p;F

=fa

mily

,P=

pare

ntin

g,C

=co

uple

,I=

indi

vidu

al.

a Peer

-rev

iew

edpu

blic

atio

n.

762 Family Relations

quasi-experimental design (sorting participantsinto the treatment or comparison-group oftenbased on participant availability). Combiningexperimental and quasi-experimental studiesis a fairly common practice in programmaticmeta-analysis with smaller sample sizes (e.g.,Hawkins & Fackrell, 2010; Holmes et al., 2010),but to ensure that this was appropriate, we testedwhether the effect sizes for the experimentaland quasi-experimental studies were statisticallydifferent. No statistically significant differencesin the effect sizes (Q = .70, p = .40) werefound. Q represents the test of homogeneityof the effect size or, in other words, teststhe null hypothesis that groups share acommon effect size. Significant effects indicateheterogeneity or evidence that the true effectsvary between groups. Therefore, the pooling ofthe experimental and quasi-experimental studiesin our analyses appears justified.

Effect sizes were calculated using a randomeffects model with Hedges’ (1981) correction,which adjusts effect size analyses accountingfor small sample bias, as many of the studiesreported relatively small sample sizes (i.e., <50participants). Each effect size was weightedby the inverse of its variance (which includesboth within-study and between-study variance)and averaged to create the overall summaryeffect size. Weights based on inverse variancerather than sample size provide a more nuancedmeasure, as the variance takes into accountthe total sample across studies as well as thesample size of each group. All analyses wereconducted using Comprehensive Meta Analysis(CMA) Version 2.

RESULTS

RQ1: Effects on Global Functioning

To address RQ1, the overall effectiveness ofCRE programs for individuals in stepcouplerelationships was examined to assess changesin participant functioning. Self-reports onmeasures of family functioning, parentingor the coparenting relationship, the couple,and the individual were combined usingstandardized scores. For the studies thatemployed a comparison-group design, theoverall standardized mean difference effectsize was calculated. A small but significantchange in functioning was found (d = .20,p = .008, k = 7, 95% CI: .05 – .35, Z = 2.63,

p = .008), indicating that participants reportedimprovements in functioning after participatingin the CRE class. The Q test for homogeneitywas conducted to ascertain whether the effectsizes from the individual studies estimated thesame population effect size; statistics indicatedthat this was a homogeneous set of studies(Q = 4.57, p = .60). The range of effect sizeswas fairly large, from d = .07 to .72. Publicationbias did not appear to be an issue, as there wasno significant difference between published andunpublished studies (Q = .44, p = .51).

As a general guide of effect size interpretation,a Cohen’s d of .20 is considered to be a smallchange, .50 is moderate, and .80 is large (Cohen,1977). Because participants in an educationalprogram (a) may already be at a higher level offunctioning and (b) experience the program for arelatively brief period (4 – 8 weeks, on average),Wolf (1986) suggested that an effect size around.25 represents a practical educational difference.Using the Rosenthal and Rubin (1982) BinomialEffect Size Display (BESD), a standardizedmean difference is translated into a success rateindicator noting the percentage of participantsabove the success threshold at posttest for boththe treatment and comparison groups. Accordingto the BESD, an effect size of .20 can also beinterpreted to mean that about 55% of treatmentgroup participants and 45% of those in thecontrol condition would show above medianimprovements in functioning.

Next, program evaluation studies that utilizeda one-group/pre-post design were examined,comparing scores of the treatment group atpretest to the scores of that same group atposttest. The overall effect of relationship edu-cation for individuals in stepfamily couple rela-tionships was significant (d = .44, p = .006,k = 8, 95% CI: .13 – .76, Z = 2.75, p = .006),yet the Q test for the hypothesis of homogene-ity had to be rejected, indicating that there wassubstantial unexplained variance (Q = 24.19,p < .001). An outlier test was conducted, andthe meta-analysis was then conducted againwithout the outlier (Stroup, 1982). The effectsize of the Stroup study, an unpublished dis-sertation, was d = 3.25, p < .001, indicating ashift in the mean score of over three standarddeviations. When we excluded this study, therange of effect sizes was more modest, fromd = .1 to d = .52, and the overall effect sizeacross studies decreased but remained significant(d = .23, p = .004, k = 7, 95% CI: .07 – .38,

CRE for Individuals in Stepfamilies 763

Z = 2.88, p = .004). With the outlier studyremoved, the Q test indicated that these werehomogenous studies (Q = 7.43, p = .28). Afterprogram participation, about 59% of partici-pants were above the overall preprogram medianscore. For one-group/pre-post studies, the dif-ference between published and unpublishedstudies was significant when the Stroup (1982)study was included (Q = 18.51, p < .001), butno differences were found when this studywas removed from the analysis (Q = 1.739,p = .19).

RQ2: Effects in Specific Domains

RQ2 examined effects of specific target out-comes; thus, separate analyses for each ofthe outcomes were conducted. Unfortunately,not enough data were available to test for aneffect of CRE on individual functioning (kshould be ≥3; Valentine et al., 2010). For thecomparison-group studies, small but significanteffect sizes were found for family functioningmeasures (d = .20, p = .018, k = 5, 95% CI:.03 – .37, Z = 2.36, p = .018) and parentingvariables (d = .24, p = .004, k = 5, 95% CI:.08 – .41, Z = 2.84, p = .004). The effect sizefor enhancements in the couple relationship wasnonsignificant (d = .07, p = .38, k = 4, 95%CI: −.09 – .24, Z = .88, p = .38).

For the one-group/pre-post studies, resultsare reported with the identified outlier(Stroup, 1982) excluded. Significant effect sizeswere found for family functioning measures(d = .24, p = .014, k = 6, 95% CI: .05 – .43,Z = 2.45, p = .014) and parenting measures(d = .35, p = .008, k = 4, 95% CI: .09 – .61,Z = 2.65, p = .008). The effect size of the

couple measures was also significant, but verysmall (d = .08, p = .017, k = 4, 95% CI:.02 – .15, Z = 2.39, p = .017).

RQ3: Cumulative Effects

RQ3 addressed cumulative effects over timeacross the studies. Cumulative effects proce-dures are used as a mechanism for displayingor illustrating a series of analyses in a sin-gle plot. The order of these analyses is basedon some factor, in this case, time. Thus, thedisplay illustrates how the estimate of theeffect size shifts as a function of time (Boren-stein et al., 2009). Separate cumulative effectsplots were created for comparison-group andone-group/pre-post studies (see Figure 1). Forboth the comparison-group and one-group/pre-post studies, there appears to be a trend ofsmaller, although still statistically significant,effect sizes across time, suggesting that morerecent programmatic evaluations appear lesseffective than those implemented three decadesago. Post hoc meta-regression analyses confirmthis trend (B = −.02, Z = −1.81, p = .07),such that, on average, for every year, there is acorresponding −.02 decrease in the effect size.As depicted in Figure 1, however, confidenceintervals become more precise over time, indi-cating higher quality studies and larger samplesizes. This trend is also found for one-group/pre-post studies (B = −.01, Z = −2.70, p = .007).

DISCUSSION

In this meta-analysis, 14 studies spanningfour decades were examined to assess theeffectiveness of CRE for stepcouples. These

FIGURE 1. CUMULATIVE EFFECTS FOR COMPARISON-GROUP STUDIES ARRANGED CHRONOLOGICALLY.

764 Family Relations

programs address the compounded needs ofthis at-risk population of couples who arerepartnered with a child or children from one ormore previous relationships and utilize a familysystems approach to couple strengthening bytargeting several subsystems of the stepfamily(Adler-Baeder, Robertson, et al., 2010). Theanalyses reveal small yet significant resultsof overall effectiveness for individuals instepcouples (comparison-group studies: d =.20; one-group/pre-post studies: d = .23). Thisindicates that the mean level of reportedfunctioning improved just over a fifth of astandard deviation after participants engaged inCRE programs.

These effects are only slightly smaller thanthose found for lower income samples (d = .25to .29 for overall effects; Hawkins & Fack-rell, 2010) and couples transitioning to parent-hood (d = .28 for communication; d = .21 forpsychological well-being; Pinquart & Teubert,2010) but considerably smaller than those foundfor studies of generalized CRE (d = .30 to .36for relationship quality; Hawkins et al., 2008)and distressed couples (d = .23 to .57 for com-munication; Blanchard et al., 2009). The effectsizes in the current meta-analysis are close tothe threshold noted as ‘‘practically meaningfulchange’’ for participants in educational settings(i.e., d ≥ .25; Wolf, 1986) but are still con-sidered comparatively small. This may indicatethat more needs to be done to improve theeffectiveness of programming for stepcouples,but it also may be because of the focus ofthe measures of functioning. Only 2 of the 14studies used stepfamily-specific measures (e.g.,stepfamily adjustment). It could be that variablessuch as couple quality and family functioninglook different in stepfamilies and include otherdimensions not assessed. Stepfamily-specificmeasures may capture a truer picture of thechange that may be occurring for these partici-pants.

When examining specific target outcomes,small, positive effect sizes were found for fam-ily functioning in comparison-group (d = .20)and one-group/pre-post design studies (Stroup,1982, excluded: d = .24) and for parenting incomparison-group (d = .24) and one-group/pre-post design studies (d = .35). The effect oncouple functioning at postprogram was not sig-nificant for comparison-group design studies andwas very small for one-group/pre-post stud-ies (Stroup, 1982, excluded: d = .08). From

this pattern of findings, a few suppositions areoffered. Perhaps CRE programs may currentlybe more effective at targeting the broader familydynamics and parenting issues related to step-family development compared to issues relatedspecifically to the couple relationship. Althoughit is shown that the quality and stability of thecouple relationship in the stepfamily is directlyrelated to the quality of other family rela-tionships, particularly the stepparent-stepchildrelationship (Hetherington & Kelly, 2002), itmay also be that participants are coming tothese classes looking for solutions to issuesrelated to the family unit and parenting ratherthan couple dynamics. Cartwright (2010) foundthat, in general, stepcouples often do not reportthe couple’s relationship as a primary sourceof concern. More focus tends to be placed onroles within the family, the stepparent-stepchildrelationship, and parenting concerns, particu-larly in the early years of stepfamily formation(Ganong & Coleman, 2004). This finding mayalso indicate a process of change among sub-systems in the family rather than concurrentchange. That is, enhancements to the couplerelationship may follow enhancements in otherareas. Unfortunately, follow-up data were notsufficient in the existing studies to examine thispossibility.

Finally, a cumulative effects procedure wasconducted. This is a common procedure inmeta-analysis, especially in medical research,yet this appears to be a unique addition to thestudy of CRE. As previously stated, cumula-tive effects models are used as a mechanismto display trends across sociohistorical time,from the beginning of CRE programming to thecurrent day. The assumption was that programdevelopment has occurred within the contextof a developing empirical literature such thatprogram enhancements and the potential forgreater effects would be expected over timeas new empirical literature is contributed (Hig-ginbotham, Henderson, & Adler-Baeder, 2007;Hughes, 1994; Jacobs, 1988). Although pro-grams over this 30-year span cover much ofthe same core programmatic content (Whit-ton et al., 2008), the more recent programsinclude additional content and an explicit the-oretical base (Adler-Baeder & Higginbotham,2004).

Interestingly and contrary to expectations,trends from the cumulative effects models forthe comparison-group and one-group/pre-post

CRE for Individuals in Stepfamilies 765

studies suggest that more recent studies pro-duce smaller but more precise effect sizes.Meta-regression analyses confirmed this trend.The post hoc nature of this research question onlyallows for speculation as to why more recentprograms are less effective than those imple-mented three decades ago. Because stepfamilieshave become increasingly prevalent over theyears, it may be that the experience has becomecomparatively more ‘‘normalized,’’ resulting inentry into recent programs of less distressedcouples. Another observation is that earlierstudies served predominantly European Ameri-can (often >90%) and middle-income marriedcouples, whereas studies conducted after 2001served more ethnically and economically diversesamples (approximately 65% European Ameri-can) of both married and nonmarried couples.The research basis for content of the CRE curric-ula used in the studies is the empirical literatureon stepfamilies, which continues to be domi-nated by study samples of married, EuropeanAmerican, middle-class stepfamilies (Colemanet al., 2000; Sweeney, 2010). Perhaps programdesign adjustments are necessary to providea better match with the current populationserved.

Research and Practice Implications

The results of the current study serve to informthe research literature on CRE and stepfam-ily interventions as well as several investedgroups, including funding agencies, practition-ers, and stepcouples themselves. From this initialmeta-analytic study, we can conclude that CREprograms that target participants in stepfami-lies, both married and nonmarried, are modestlyeffective in influencing overall participant func-tioning as well as specific target outcomes,including family functioning and parenting, andappear worthy of support. We suggest, however,that this not be the final step in the evaluationprocess. Evaluation is intended to be an itera-tive process in which research informs practiceto benefit the experiences of those who seekeducational services. Thus, CRE program devel-opment should be a dynamic process fueled byexisting research results, including the currentmeta-analysis, and next steps include continuedevaluations that inform the refinement of CREprograms (Higginbotham et al., 2007; Hughes,1994; Jacobs, 1988). Individual evaluation stud-ies with strong research designs are essential

as they dig deeper, evaluating factors relatedto program effects, such as the context of theimplementation setting, including ethnic andeconomic diversity of participants. Understand-ing both broad effects and more specific pro-cesses will serve to inform research and practice.Specifically, it is important to explore further thelack of evidence of program effects in the coupledomain for CRE targeting stepfamilies.

It is not possible to determine from the currentstudy if improvements in family functioning andparenting would eventually lead to couple-levelchanges because of the lack of studies withlong-term follow-up data. Understanding of thepatterns of change among multiple outcomeareas across time for CRE participants isneeded (Adler-Baeder, Bradford, et al., 2010). Inaddition, evaluations of general CRE programswould benefit from the examination of changepatterns by couple type (e.g., first marriage,stepfamily, cohabiting couple) and relationshipstage (e.g., newlywed, early marriage, latermarriage).

The cumulative effects procedure raisesimportant questions for the field of CREand stepfamily programming specifically. Re-searchers are encouraged to examine trends inCRE effectiveness with other subpopulations ofparticipants. Early CRE programs for stepcou-ples appear more effective than recent studies,but the findings of the more recent studies sug-gest more precise estimates likely because oflarger sample sizes being examined. Efforts tounderstand this apparent trend are critical to thedevelopment of best practices. Researchers areencouraged to implement evaluations that uti-lize an action research approach coupled with anecocultural lens. This requires that researchersand program developers collaborate with com-munity partners, educators, and participants tounderstand program context and needed changesas a means to inform scholarship and prac-tice and lead to better outcomes for participants(Small & Uttal, 2005). It is also critical thatthe empirical study of stepfamilies include sam-ples that more closely resemble characteristicsof the current population of stepfamily couplesattending CRE.

Study Limitations

This meta-analysis provides novel informationon the effectiveness of stepfamily educationalprogramming; however, limitations should be

766 Family Relations

noted. First is the limited sample size. It is evi-dent that further work is needed to build on thisbody of research, as inclusion of a greater num-ber of studies allows researchers the ability toconduct analyses of specific target outcomes. Inthe current study, it was not possible to assessindividual functioning or examine moderatorssuch as program context because of the lim-ited number of studies that contributed relevantdata. Most notably, the long-term effects ofCRE for stepcouples could not be addressedwith the current sample of evaluation studies.The lack of studies with follow-up data, in gen-eral, and data that goes beyond 2 months, morespecifically, leaves a gap in our understandingof these long-term effects. This deficit limitsour ability to fully evaluate programs for theirsustained or delayed effects or both and lim-its potential policy implications. At present, themost we can say is that the findings are infor-mative of initial effects of CRE programs forstepfamilies.

It is important also to acknowledge that therelatively small number of studies includedin this meta-analysis may not represent thepopulation of effects for all stepcouples servedin CRE. It is not clear how representativethese 14 studies are of all the educationalprograms offered to couples in stepfamiliesover the past four decades; thus, interpretationsshould be made with some caution. Previouswork in this field and methodological guidelinessuggest, however, that this is an adequatenumber of studies to meaningfully contributeto the body of literature on stepfamilies andCRE. Valentine et al. (2010) noted that ‘‘giventhe need for some kind of synthesis, all theavailable alternatives are worse than meta-analysis (even those with few studies), in thatthey are likely to be based on less defensibleassumptions and on less transparent processes’’(p. 239). A larger sample of studies that collectpre-, post-, and follow-up data using empiricallyvalidated measures can further develop ourknowledge of the sustained impact of CREfor stepcouples and the variations in outcomesthat may exist based on stepfamily couplecharacteristics.

Finally, in an attempt to be inclusive ofall evaluation studies, this study includedboth published and unpublished work. Thisis beneficial to avoid publication bias andthe overestimating of effects as noted inprevious meta-analyses (e.g., Fawcett et al.,

2010). Although no statistical differences werefound, it is still difficult to determine thequality of work contributed by unpublishedstudies, as these studies have not undergonea peer-reviewed evaluation process.

Conclusions and Future Directions

CRE programs for stepcouples have, on aver-age, small but significant effects overall; how-ever, larger effects were found for changes infamily and parental functioning following pro-gram participation. Suggested future directionsof this work include empirical and evaluativework as well as critical programmatic assess-ments. Increasing the empirical knowledge ofstepfamily development and functioning amongmore diverse populations is essential (Sweeney,2010). In addition, researchers and programdevelopers are encouraged to continue imple-menting an iterative approach to program eval-uation and refinement using individual researchfindings and the findings of the current study,as these can provide insight into the strengthsand areas of needed development for CRE pro-grams for stepcouples. Quality research designsframed with an ecocultural lens using con-trol groups and long-term follow-up proceduresare needed. Additionally, using a variety ofmeasures specific to couples in stepfamiliesand stepfamily functioning, including steppar-enting behaviors and quality and coparentingbehaviors and quality, will enhance our under-standing of these CRE programs’ effects andcontinue to move the study of CRE programsforward.

REFERENCES

References marked with an asterisk indicate studiesincluded in the meta-analysis.

Adler-Baeder, F. (2007). Smart Steps: Embrace theJourney. Auburn, AL: The National StepfamilyResource Center.

Adler-Baeder, F., Bradford, A., Skuban, E., Lucier-Greer, M., Ketring, S., & Smith, T. (2010). Demo-graphic predictors of relationship and marriageeducation participants’ pre- and post-programrelational and individual functioning. Journal ofCouple & Relationship Therapy, 9, 113 – 132.doi:10.1080/15332691003694885

Adler-Baeder, F., & Higginbotham, B. (2004). Impli-cations of remarriage and stepfamily forma-tion for marriage education. Family Relations,53, 448 – 458. doi:10.1111/j.0197 – 6664.2004.00053.x

CRE for Individuals in Stepfamilies 767

*Adler-Baeder, F., & Lucier-Greer, M. (2011).Alabama Community Healthy Marriage Initiative:Findings from stepfamily couples. Auburn, AL:Auburn University.

Adler-Baeder, F., Robertson, A., & Schramm, D.(2010). Community education programs servingcouples in stepfamilies: A qualitative study offormat, content, and service delivery. Journalof Extension, 48, Article 5RIB1. Retrieved fromhttp://www.joe.org/joe/2010october/rb1.php

*Bielenberg, L. T. (1991). A task-centered pre-ventive group approach to create cohesion inthe new stepfamily: A preliminary evaluation.Research on Social Work Practice, 1, 416 – 433.doi:10.1177/104973159100100406

Blanchard, V., Hawkins, A., Baldwin, S., & Fawcett,E. (2009). Investigating the effects of marriage andrelationship education on couples’ communicationskills: A meta-analytic study. Journal of FamilyPsychology, 23, 203 – 214. doi:10.1037/a0015211

Borenstein, M., Hedges, L. V., Higgins, J., & Roth-stein, H. R. (2009). Introduction to meta-analysis.Chichester, UK: Wiley.

*Brady, C. A., & Ambler, J. (1982). Therapy withremarriage families: X. Use of group educationaltechniques with remarried couples. Family Ther-apy Collections, 2, 145 – 157.

Bumpass, L., & Raley, K. (2007). Measuring separa-tion and divorce. In S. L. Hofferth & L. M. Casper(Eds.), Handbook of measurement issues in familyresearch (pp. 125 – 143). Mahwah, NJ: Erlbaum.

Butler, M. H., & Wampler, K. S. (1999). A meta-analytical update of research on the couple com-munication program. American Journal of FamilyTherapy, 27, 223 – 237. doi:10.1080/0192618992-61943

Carroll, J. S., & Doherty, W. J. (2003). Evaluating theeffectiveness of premarital prevention programs:A meta-analytic review of outcome research.Family Relations, 52, 105 – 118. doi:10.1111/j.1741 – 3729.2003.00105.x

Cartwright, C. (2010). Preparing to repartner andlive in a stepfamily: An exploratory investiga-tion. Journal of Family Studies, 16, 237 – 250.doi:10.5172/jfs.16.3.237

Cohen, J. (1977). Statistical power analysis for behav-ioral sciences (rev. ed.). New York: AcademicPress.

Coleman, M., Ganong, L., & Fine, M. (2000). Rein-vestigating remarriage: Another decade ofprogress. Journal of Marriage & the Family,62, 1288 – 1307. doi:10.1111/j.1741 – 3737.2000.01288.x

*Cuddeby, G. (1984). The effects of stepparent edu-cation on perceived family cohesion, organization,and conflict. (Unpublished doctoral dissertation).Arizona State University, Phoenix, Arizona.

Duncan, S. F., & Goddard, H. W. (2005). Family lifeeducation: Principles and practices for effectiveoutreach. London: Sage.

Einstein, E., & Albert, L. (1986). Strengthening yourstepfamily. Circle Pines, MN: American GuidanceService, Inc.

Fawcett, E. B., Hawkins, A. J., Blanchard, V. L., &Carroll, J. S. (2010). Do premarital educationprograms really work? A meta-analytic study.Family Relations, 59, 232 – 239. doi:10.1111/j.1741 – 3729.2010.00598.x

Forgatch, M. S., & Rains, L. (1997). MAPS: Marriageand Parenting in Stepfamilies [Parent trainingmanual]. Eugene, OR: Oregon Social LearningCenter.

Ganong, L., & Coleman, M. (2004). Stepfamily rela-tionships: Development, dynamics and interven-tions. New York: Kluwer Academic/Plenum.

*Gellat, V., Adler-Baeder, F., & Seeley, J. (2010).An interactive web-based program for parentsin stepfamilies: Development and evaluationof efficacy. Family Relations, 59, 572 – 586.doi:10.1111/j.1741 – 3729.2010.00624.x

*Gibbard, S. (1998). The effectiveness of an eight-week psychoeducational seminar on family envi-ronment and marital satisfaction in stepfamilies.(Unpublished doctoral dissertation). University ofMississippi, Oxford, Mississippi.

Giblin, P., Sprenkle, D. H., & Sheehan, R. (1985).Enrichment outcome research: A meta-analysis ofpremarital, marital and family interventions. Jour-nal of Marital and Family Therapy, 11, 257 – 271.doi:10.1111/j.1752 – 0606.1985.tb00619.x

Halford, W. (2004). The future of couple relationshipeducation: Suggestions on how it can makea difference. Family Relations, 53, 559 – 566.doi:10.1111/j.0197-6664.2004.00065.x

Hawkins, A., Blanchard, V., Baldwin, S., & Fawcett,E. (2008). Does marriage and relationship edu-cation work? A meta-analytic study. Journal ofConsulting and Clinical Psychology, 76, 723 – 734.doi:10.1037/a0012584

Hawkins, A. J., & Fackrell, T. A. (2010). Doesrelationship and marriage education for lower-income couples work? A meta-analytic studyof emerging research. Journal of Couple &Relationship Therapy, 9, 181 – 191. doi:10.1080/15332691003694927

Hedges, L. V. (1981). Distribution theory for Glass’sestimator of effect size and related estimators.Journal of Educational Statistics, 6, 107 – 128.doi:10.2307/1164588

*Henderson, H. (2001). Impact of a court-orderedpsychoeducational course on stepfamily adapt-ability and cohesion. (Unpublished doctoral dis-sertation). Indiana State University, Terre Haute,Indiana.

768 Family Relations

Hetherington, E. M., & Kelly, J. (2002). For betteror for worse: Divorce reconsidered. New York:Norton.

*Higbie, D. (1994). Empowering stepfamilies: Eval-uation of an educational curriculum for stepfamilycouples. (Unpublished doctoral dissertation). Mar-quette University, Milwaukee, Wisconsin.

*Higginbotham, B. (2010). Utah’s stepfamily educa-tion database as of Dec 9, 2010. Logan, UT: UtahState University.

Higginbotham, B. J., Henderson, K., & Adler-Baeder,F. (2007). Using research in marriage and rela-tionship education programming. The Forumfor Family and Consumer Issues, 12. Retrievedfrom http://ncsu.edu/ffci/publications/2007/v12-n1-2007-spring/higginbotham/fa-4-higginbotham.php

Holmes, E. K., Galovan, A. M., Yoshida, K., &Hawkins, A. J. (2010). Meta-analysis of the effec-tiveness of resident fathering programs: Are fam-ily life educators interested in fathers? FamilyRelations, 59, 240 – 252. doi:10.1111/j.1741-3729.2010.00599.x

Hughes, R. (1994). A framework for developingfamily life education programs. Family Relations,43, 74 – 80.

Jacobs, F. H. (1988). The five-tiered approach toevaluation: Context and implementation. InH. B. Weiss, F. H. Jacobs, H. B. Weiss, F. H.Jacobs (Eds.), Evaluating family programs(pp. 37 – 68). Hawthorne, NY: Aldine de Gruyter.

Kreider, R. M., & Ellis, R. (2011). Number, timing,and duration of marriages and divorces: 2009(Current Population Reports, P70-125). Retrievedfrom http://www.census.gov/prod/2011pubs/p70-125.pdf

Larson, J. H., Anderson, J. O., & Morgan, A. (1984).Effective stepparenting. New York: Family Serviceof America.

Lawton, J. M., & Sanders, M. R. (1992). The step-families self-help program. Unpublished treatmentmanual. Department of Psychiatry, The Universityof Queensland, Brisbane, Australia.

Lehman, R. T. (1947). Home and family life edu-cation. Review of Educational Research, 17,209 – 221. doi:10.2307/1168663

Lipsey, M. W., & Wilson, D. B. (2001). Practicalmeta-analysis. Thousand Oaks, CA: Sage.

Messinger, L., Walker, K. N., & Freeman, S. J.(1978). Preparation for remarriage followingdivorce: The use of group techniques. AmericanJournal of Orthopsychiatry, 48, 263 – 272.

*Nelson, W., & Levant, R. (1991). An evaluationof a skills training program for parents instepfamilies. Family Relations, 40, 291 – 296.doi:10.2307/585014

*Nicholson, J., & Sanders, M. (1999). Randomizedcontrolled trial of behavioral family interventionfor the treatment of child behavior problems in

stepfamilies. Journal of Divorce & Remarriage,30, 1 – 23. doi:10.1300/J087v30n03_01

Nowak, C., & Heinrichs, N. (2008). A comprehensivemeta-analysis of Triple-P Positive Parentingprogram using hierarchical linear modeling:Effectiveness and moderating variables. ClinicalChild and Family Psychological Review, 11,114 – 144. doi:10.1007/s10567-008-0033-0

*Oregon Social Learning Center. (2010). Marriageand Parenting in Stepfamilies (MAPS) report2002 – 2005. Eugene, OR: Oregon Social LearningCenter.

Pill, C. J. (1981). A family life education groupworking with stepparents. Social Casework, 62,159 – 166.

Pinquart, M., & Teubert, D. (2010). A meta-analyticstudy of couple interventions during the transitionto parenthood. Family Relations, 59, 221 – 231.doi:10.1111/j.1741-3729.2010.00597.x

Reardon-Anderson, J., Stagner, M., Macomber, J. E.,& Murray, J. (2005). Systematic review of theimpact of marriage and relationship programs.Washington, DC: Urban Institute. Retrievedfrom http://www.urban.org/UploadedPDF/411142_impact_ marriage.pdf

Rosenthal, R., & Rubin, D. B. (1982). A simple, gen-eral purpose display of magnitude of experimentaleffect. Journal of Educational Psychology, 74,166 – 169. doi:10.1037/0022-0663.74.2.166

Shirer, K. (2009). Together we can: Creating a healthyfuture for our family. Lansing, MI: Michigan StateUniversity Board of Trustees.

Small, S. A., & Uttal, L. (2005). Action-orientedresearch: Strategies for engaged scholarship.Journal of Marriage and Family, 67, 936 – 948.doi:10.1111/j.1741-3737.2005.00185.x

*Stroup, M. (1982). A preliminary study comparingtwo structured group treatments for stepfamilies:Couple/family treatment and family treatment.(Unpublished doctoral dissertation). BrighamYoung University, Provo, Utah.

Sweeney, M. M. (2010). Remarriage and stepfam-ilies: Strategic sites for family scholarship inthe 21st century. Journal of Marriage and Fam-ily, 72, 667 – 684. doi:10.1111/j.1741-3737.2010.00724.x

U.S. Department of Health and Human Ser-vices: Administration for Children & Families.(2010). Healthy marriage matters. Retrieved fromhttp://www.acf.hhs.gov/healthymarriage/about/factsheets_hm_matters.html

Valentine, J. C., Pigott, T. D., & Rothstein, H. R.(2010). How many studies do you need? A primeron statistical power for meta-analysis. Journalof Educational and Behavioral Statistics, 35,215 – 247. doi:10.3102/1076998609346961

Webber, R. (1986). Living in a Stepfamily: Aneducative program for step-parents. Unpublished

CRE for Individuals in Stepfamilies 769

treatment manual, Monash University, Melbourne,Australia.

*Webber, R., Sharpley, C., & Rowley, G. (1988).Living in a stepfamily. Australian Journal of Sex,Marriage & Family, 9, 21 – 29.

Whitton, S., Nicholson, J., & Markman, H. (2008).Research on interventions for stepfamily couples:

The state of the field. In J. Pryor (Ed.), The inter-national handbook of stepfamilies: Policy andpractice in legal, research, and clinical environ-ments (pp. 455 – 484). Hoboken, NJ: Wiley.

Wolf, F. M. (1986). Meta-analysis: Quantitativemethods for research synthesis. Beverly Hills, CA:Sage.

The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked to publications on ResearchGate.The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked to publications on ResearchGate.