Journal of Technology in Human Services, 30:14–31, 2012

Copyright # Taylor & Francis Group, LLC

ISSN: 1522-8835 print=1522-8991 online
DOI: 10.1080/15228835.2012.662848

A Preliminary Comparison Study

of Online and Face-to-Face Counseling:
Client Perceptions of Three Factors

COURTNEY HOLMES
Bowling Green State University, Bowling Green, Ohio

VICTORIA FOSTER
The College of William & Mary, Williamsburg, Virginia

This preliminary study examined the similarities and differences

between online and face-to-face counseling with regard to general
mental health, working alliance, and social presence. Participants
included 50 current counseling clients (37 face-to-face, 13 online)
who all were recruited through online mediums. Results indicated
that online counseling clients perceived a significantly stronger
working alliance on the total Working Alliance Inventory-Short
Form as well as the Goal subscale than did those who received
face-to-face only counseling. No significant differences existed
between the groups on general mental health or social presence.
These results suggest that a more controlled study with a larger
sample might find that a strong working alliance can be developed
through online counseling.

KEYWORDS online counseling, social presence, therapeutic

relationship

Clients and counselors alike are using technology to supplement face-to-face

services, or provide stand-alone services solely through computer-mediated
communication. Traditionally, counseling is structured around the human
experience and relies, in part, on the therapeutic relationship developed

Received October 20, 2011; revised January 22, 2012; accepted January 30, 2012.
Address correspondence to Courtney M. Holmes, PhD, NCC, Assistant Professor, Mental
Health and School Counseling Program, Bowling Green State University, 451 Education
Building, Bowling Green, OH 43551. E-mail: courtmh@bgsu.edu

14
 Online and F2F Counseling 15

between the client and counselor (Rogers, 1957). This human experience as
well as the therapeutic relationship between the client and counselor is sup-
ported by research (Duncan, Miller, Wampold, & Hubble, 2010; Hubble,
Duncan, & Miller, 1999). As technology shifts the way humans interact with
each other and their reality (Borgmann, 1984), online engagement continues
to infiltrate the field of counseling. Can a meaningful clinical relationship be
developed if a client and counselor do not share the same physical space
(Chester & Glass, 2006)? Surprisingly little research has evaluated therapy
provided through online or technological modalities in conjunction with
the therapeutic relationship.
As the field endeavors to match research with emerging practices,
Lundberg (2001) aptly notes that, ‘‘human service professionals are still trying
to understand just how human relationships change when they are mediated
by computers’’ (p. 142). The impact of a positive working alliance in tra-
ditional counseling is well documented; however, little is known about
how this alliance functions in online counseling (Knaevelsrud & Maercker,
2006). Whether technology-assisted counseling can construct an authentic
therapeutic relationship remains largely unsubstantiated.

BRIEF OVERVIEW OF ONLINE COUNSELING

E-mail, synchronous chat, videoconferencing and Internet phone (tele-

phone) are employed under the umbrella of computer-mediated communi-
cation or online counseling (Ainsworth, 2002). Internet phone (telephone)
counseling has been used as a supplemental form of face-to-face counsel-
ing for years (i.e., scheduling, consultation, crisis management) and has
served as a predecessor to current computer mediated technologies
(Mallen, Vogel, Rochlen, & Day, 2005). Synchronous chat and e-mail are
the two most commonly used forms of Internet counseling (Pelling,
2009). Synchronous chat allows for a real-time, text-based communication
between participants, however, the participants do not see or hear one
another through the conversation. E-mail counseling exchanges are similar
to synchronous chat communication, in that they also lack the ability to
retrieve verbal and nonverbal information. However, e-mail takes place at
different times so the participants do not share the same cyber space sim-
ultaneously. E-mail and synchronous chat are the most easily available to
online users requiring only an e-mail account and access to a computer
with the Internet. Videoconferencing, a lesser-utilized tool, allows each
participant to speak with one another with the additional benefit of sup-
porting verbal and nonverbal cues within the conversation. Although video-
conferencing is becoming increasingly more affordable, many clients do not
have access to this type of counseling due to monetary constraints (Mallen
et al., 2005).
16 C. Holmes and V. Foster

Current literature touts the ability to reach clients not otherwise served
by traditional counseling as a major benefit of online counseling (Layne &
Hohenshil, 2005). However, whether clients who are physically, geographi-
cally, or financially compromised are indeed those seeking services offered
through the Internet has not yet been documented in current literature.
Several studies that examined the therapeutic alliance built through online
counseling cited the homogenous samples of Caucasian, college-educated,
women as a methodological weakness (Cook & Doyle, 2002; Leibert, Archer,
Munson, & York, 2006).

THE WORKING ALLIANCE THROUGH ONLINE COUNSELING

The construct of the working alliance describes the degree to which the client
and counselor join in a mutual counseling relationship to promote client
change. Three constructs are defined within the working alliance: bonds
(i.e., the degree of mutual trust, acceptance, and confidence between the
client and counselor), goals (i.e., how closely the client and counselor agree
on how to reach the treatment goals), and tasks (i.e., how closely the client
and counselor are mutually engaged in the goals of therapy) (Horvath &
Greenberg, 1989).
The Hanley and Reynolds (2009) completed a meta-analysis of five stu-
dies specifically targeting the therapeutic alliance in online counseling. These
five studies, totaling 161 participants, examined human-to-human ongoing
therapeutic relationships that were maintained via computer-mediated com-
munication. Four of the five studies used previously established face-to-face
comparison groups. Three of these four studies showed that in comparison,
the online alliance was rated higher than in the face-to-face comparison
group. Overall, the review supported the contention that a measurable thera-
peutic alliance can be formed through computer-mediated therapy. However,
the small numbers of studies available and methodological weaknesses
such as homogenous, convenience samples impact the generalizability of
the results. This scant base of literature demonstrates the need for more
research as the field forges ahead with the integration of computer-mediated
counseling, regardless of the lack of research to fully substantiate such use.

SOCIAL PRESENCE

Mediated social presence or telepresence, the moment-by-moment awareness

of the co-presence of another sentient being accompanied by a sense of
engagement with the other (Biocca, Harms, & Gregg, 2001, p. 2), encompasses
the sense of awareness and engagement that one person feels when com-
municating via technology with another person. According to Short, Williams,
and Christie (1976) ‘‘social presence is an important key to understanding
 Online and F2F Counseling 17

person-to person telecommunications and varies between different media,

affecting the nature of the interaction’’ (p. 65). Social presence is impacted
by the capacity of each particular type of mediated communication to portray
particular nonverbal information between the communicating partners
(Gunawardena & Zittle, 1997). Essentially, this working definition includes
the sense of awareness and engagement that one person feels when com-
municating via technology with another person. Social presence is one of
the most significant factors in building a sense of community through online
communication (Aragon, 2003).
Users of computer-mediated communication reported feeling a similar
sense of social presence as users rating face-to-face communication (Walther,
1995). Extending this line of research the current study examined the concepts
of social presence and working alliance within the context of an ongoing
therapeutic relationship, including how these constructs are similar or differ-
ent through online counseling and face-to-face counseling. According to
Grohol (1998), ‘‘as telepresence increases, clinicians and researchers need
to allow that the subjective line between ‘being there’ in real-life and ‘being
there’ via technology will become increasingly blurred’’ (p. 127). Ultimately,
with advances in technology, communication may take place through
modalities that simulate a face-to-face environment. As humans continue to
utilize as well as refine these technologies, a clear boundary of what consti-
tutes meaningful communication will be vague.
This study examined the similarities and differences between online and
face-to-face counseling with regard to the working alliance, client mental
health status, and social presence. Earlier studies have laid important ground-
work for the current study although one potential methodological weakness
was shared by several of those studies (Cook & Doyle, 2002; Leibert, et al.,
2006): the comparison of collected online data with original norming data
from the Working Alliance Inventory (WAI). The authors attempted to address
this weakness by recruiting face-to-face and online volunteers simultaneously
to participate in the same study for comparison. Also, the exploration of social
presence for the study of online counseling is relatively new and expands the
current body of literature related to investigations of online communities and
classrooms. Finally, if results were meaningful, it was hoped that this prelimi-
nary study could be replicated with a larger sample size to produce more
conclusive results.

METHOD

Participants
Eligible participants were at least 18 years of age and currently involved in
either face-to-face (i.e., individual, group, couples, or family) or online coun-
seling (i.e., telephone, synchronous chat, asynchronous e-mail exchanges, or
18 C. Holmes and V. Foster

videoconferencing). In addition, all participants had completed at least three

sessions with their respective counselors. A minimum of three complete
sessions was used as the benchmark as the literature notes an early, measur-
able working alliance can be developed within three sessions (Horvath &
Greenberg, 1989).
A convenience sample of 50 current counseling participants, enrolled
in either online (n ¼ 13) or face-to-face counseling (n ¼ 37) was obtained.
The sample was obtained through online recruitment, and at the time of par-
ticipation clients identified the type of counseling he or she was receiving.
Participants were separated into two categories: face-to-face and online
clients. Participants were allowed to choose all categories that fit their current
counseling situation, and many clients were involved in several different types
of therapy through the same online or face-to-face medium (i.e., family and
individual counseling). However, only two participants were involved in both
online and face-to-face counseling. For the purposes of the analysis, these
hybrid clients were classified as online counseling clients. Due to the small
sample size, the different types of counseling could not be compared, how-
ever this would be an important next step for future research. The results of
the counseling modalities utilized as represented by the sample are found
in Table 1.
The sample consisted of 22% (11) male participants and 78% (39)
female participants. The age range of the sample included participants from
18 to 56 þ years, specifically, 15% (8) of participants between 18–25 years of
age, 52% (26) of participants in the 26–40 age group, 22% (11) in the 41–55
age group, and 10% (5) in the 56 þ age group. Additionally, 90% (45) of part-
icipants identified themselves as Caucasian, 2% (1) identified as Asian; 2% (1)
identified as Native American=Alaskan Native; and 6% (3) identified as two or
more categories.
Participants disclosed the presenting problem(s) for which they initiated
therapy. The representation of presenting issues appears to be proportional
between the mediums of counseling, as well as similar to the ratio of issues
found at mental health agencies. The options provided are found in Table 2.

TABLE 1 Utilized types of counseling (N ¼ 50; percentages may add to more than 100)

N % (Type of counseling) % (Total N)

Online individual counseling (N ¼ 13)

Telephone 5 38.4 10.0
Synchronous chat 5 38.4 10.0
E-mail=asynchronous chat 2 15.3 4.0
Videoconferencing 4 30.7 8.0
Face-to-face counseling (N ¼ 37)
Individual 36 97.2 72.0
Group 4 10.8 8.0
Couples=family 5 13.5 10.0
 Online and F2F Counseling 19

TABLE 2 Client Presenting Problem (N ¼ 50; will add to more than 100 percent)

Face-to-Face#=% Online#=% Total#=%

Depression 20=54 6=46 26=52

Anger management 1=3 1=8 2=4
Addiction 2=5 4=31 6=12
Loss=Grief 6=16 2=15 8=16
Relationship issues 18=49 5=39 23=46
Anxiety 22=59 4=31 26=52
Legal mandate 0=0 1=8 1=2
Self improvement 13=35 7=54 21=42
Sexual identity 0=0 3=23 3=6
Other mental health concern 5=14 0=0 5=10

Materials
DEMOGRAPHIC QUESTIONNAIRE
A questionnaire was developed to obtain participant demographic information
including age, number of sessions in counseling, race=ethnicity, gender, level
of education, and presenting problem for counseling.

WORKING ALLIANCE INVENTORY—SHORT FORM

The Working Alliance Inventory-Short Form (WAI-S; Kokotovic & Tracey,
1989) measures clients’ and counselors’ perception of the quality of the
therapeutic alliance, based on Bordin’s pantheoretical conceptualization of
the working alliance (Horvath & Greenberg, 1989). The pantheoretical
concept pertains to the idea that regardless of the theoretical orientation, a
working alliance between the client and counselor is a crucial part of the
therapeutic process (Horvath & Greenberg, 1989).
Horvath and Greenburg (1989) published the original Working Alliance
Inventory. To create a shorter instrument for faster and easier distribution,
Kokotovic and Tracey (1989) developed the WAI-S shortly after the develop-
ment of the original. The WAI-S is a 12-item self-report form that maintains
the same seven-item Likert scale format, ranging from never to always.
The WAI-S contains three subscales named Bonds, Goals, and Tasks, with
four items per subscale. Raw scores range from twelve to 84, with higher
scores reflecting increasingly positive perceptions of the alliance (Horvath
& Greenberg, 1989).
Although the technical adequacy of the WAI has been well-documented
by over 100 published research reports, the WAI-S has little information
published on its psychometric quality, aside from the initial findings reported
by the developers (Busseri & Tyler, 2003). However, the WAI and WAI-S
have been used interchangeably to measure the working alliance bet-
ween client and counselor since its publication. In order to validate the
20 C. Holmes and V. Foster

interchangeability of the two instruments Busseri and Tyler (2003) compared

the WAI and the WAI-S directly and showed support for the internal consist-
ency of each subscale as well as the composite score as assessed twice during
therapy. The internal consistency of the total scale composite score at the
fourth session and final session, respectively, ranged from .91 and .92.
Carmel and Friedlander (2009) found internal consistency for the composite
score to be .93. These findings are similar to the technical adequacy reports
of the WAI and further support the interchangeability of the instruments. In
the current study, a total scale Chronbach’s alpha of .94 was found for the
WAI-S, indicating strong internal consistency. Similarly, internal consistency
was computed for each of the WAI-S’s subscales; for the Task, Bond, and
Goal subscales, Chronbach’s alpha coefficients of .86, .85, and .86 were
obtained, respectively.

THE NETWORKED MINDS SOCIAL PRESENCE MEASURE

The Networked Minds Social Presence Measure (NMSPM; Harms & Biocca,
2004) was developed as a method to assess the concept of social presence
in computer-mediated communication. The NMSPM is composed of 36 items
that are answered on a 7-point Likert-type scale; with anchors ranging from
strongly agree to strongly disagree. The instrument takes between 10 and
15 min to complete. A higher score on the instrument or its respective sub-
scales indicate a stronger perception of social presence. The six subscales
include: (a) Co-Presence, (b) Attentional Allocation, (c) Perceived Message
Understanding, (d) Perceived Emotional Understanding, (e) Perceived
Behavioral Interdependence, and (f) Perceived Emotional Interdependence
(Harms & Biocca, 2004).
Support for the technical adequacy of the instrument is limited, yet the
few available studies validate the use of the instrument for research. Concur-
rent validity investigations substantiate the ability of the instrument to detect
differences in levels of social presence across different types of media
(Biocca et al., 2001). In a study with 240 students at a Midwestern college=
university campus, confirmatory factor analysis supported six distinct factors
or subscales and the retention of 36 items out of the originally proposed 100.
Confirmatory factor analysis showed each of the 36 items with factor loadings
ranging from .52 to .81 on their respective scales. This confirmatory factor
analysis confirmed the hypothesis that six distinct factors would collate from
the data. The six subscales (individual factors) maintained internal consist-
ency between .81 to .87. Initial studies completed by the developers show
initial support of the reliability and validity of the instrument with concurrent
validity and internal consistency as evidenced by alpha levels of the sub-
scales between .81 to .87. In the current study, a total scale Chronbach’s
alpha of .89 was found for the Networked Minds Social Presence Measure,
indicating strong internal consistency.
 Online and F2F Counseling 21

GENERAL HEALTH QUESTIONNAIRE-12

The General Health Questionnaire (GHQ) was ‘‘designed to be a self-
administered screening test aimed at detecting psychiatric disorders among
respondents in community settings and non-psychiatric clinical settings’’
(Goldberg & Williams, 1988, p. 1). The GHQ requests that the participant fill
out the survey with regard to psychological functioning over the last two
weeks. The original GHQ, which consisted of sixty items, was revised and
restructured into three shorter instruments, the GHQ-30, GHQ-28, and the
GHQ-12. The GHQ-12 is primarily used as a screening tool to indicate poten-
tial psychological disturbance or mental illness (LoBello, 1995).
The GHQ-12 consists of twelve questions that the participant is asked to
answer based on feelings over the last two weeks. Each question is evaluated
on a 4-point scale described as follows: ‘‘less than usual,’’ ‘‘no more than
usual,’’ ‘‘rather more than usual,’’ ‘‘much more than usual.’’ The measure
can be scored with a binary coding using only ‘‘0’’ and ‘‘1,’’ as well as with tra-
ditional Likert-type scoring with each answer earning a ‘‘0,’’ ‘‘1,’’ ‘‘2,’’ or ‘‘3.’’ A
higher total score on the GHQ-12 describes a more troubled and less healthy
description of mental health status.
According to Mathers and Shipton (1993), ‘‘the GHQ is one of the most
thoroughly tested mental health questionnaires, with high reliability and
validity . . . ’’ (p. 311). They found significant decreases in pre- and post-GHQ-
12 scores using a sample of 45 university counseling clients. These decreases
between pretreatment and posttreatment indicate both positive counseling
outcomes as well as the ability of the GHQ-12 to be sensitive to the change
in mental health quality over time. The GHQ-12 has a split-half reliability of
.83 and a test–retest reliability of .73 (Goldman & Williams, 1988). In the cur-
rent study, a total scale Chronbach’s alpha of .72 was found for the GHQ-12,
indicating a moderate internal consistency for this screening measure.

Procedures
All data were gathered through online recruitment from October 2010 to
February 2011. A website outlined the study and contained an embedded link
that housed the informed consent form, assessments, and a password to
protect the survey. The website link was advertised through Yahoo# online
support groups, the Online Therapy Institute website and blog, independent
online therapy practitioner sites, Facebook, and a counselor education blog
targeting those who have participated in either type of counseling. Addition-
ally, as people were contacted about the study, a snowball effect occurred in
which those contacts passed the study onto family and friends who were either
providing or enrolled in either type of therapy. The volunteer, convenience
sample provides limitation in the representativeness of the current study; future
research should continue to procure more representative samples.
22 C. Holmes and V. Foster

Upon deciding to participate, each volunteer clicked on the embedded

link and completed the informed consent to begin the battery of assessments.
This study employed Qualtrics to house the survey. Assessments were admi-
nistered in the order as followed: (a) demographic questionnaire, (b) the
WAI-S, (c) the NMSPM, and (d) the GHQ-12. All of the information obtained
was captured in an online database that maintains the confidentiality and
anonymity of the participants. The online database of information was
password protected and only accessible by the researcher. Additionally,
participants did not enter identifying information protecting all data as anony-
mous and confidential. The battery of measures took between 20 and 25 min
for respondents to complete.
Dependent variables included the WAI-S, the NMSPM, and the GHQ-12.
Independent variables included the type of counseling in which the client
was involved face-to-face and online. Research questions included how the
demographic characteristics of online counseling participants compare with
face-to-face counseling participants, if total scale means on the dependent
measures differed between the two counseling groups, and the strength of
the relationships between the dependent measures.

Analysis
As a descriptive study, analyses included frequency data, independent
samples t-tests, and zero order correlations. Frequencies were computed to
ascertain and describe the demographics of the volunteer sample. Inde-
pendent samples t-tests were computed to investigate differences between
online and face-to-face counseling clients on all three dependent measures,
the WAI-S, NMSPM, and the GHQ-12. Cohen’s d was computed to measure
the effect size for the significant differences found in the analyses, followed
by a post hoc power analysis. Correlations were computed to discover the
relationship between the three dependent measures. Finally, internal consist-
ency was computed for each of the dependent measures. Post hoc power
analyses were conducted to examine the power (i.e., sample size) to achieve
statistical significance.

RESULTS

This study was designed to gain an understanding of the similarities and dif-
ferences between the demographic characteristics of the clients involved in
both online and face-to-face counseling. Chi Square analyses were completed
to test the null hypothesis that each type of counseling had equal distributions
of participants. A Chi Square analysis showed that the percentage of parti-
cipants that were involved in each type of counseling did not differ signifi-
cantly by age, x2(3, N ¼ 50) ¼ 4.20, p > .05, gender, x2 (1, N ¼ 50) ¼ .787,
 Online and F2F Counseling 23

p > .05, race=ethnicity, x2 (3, N ¼ 50) ¼ .797, p > .05, or education level, x2 (3,
N ¼ 50) ¼ 5.63, p > .05, supporting the null hypothesis of equal proportions
for each variable.
T-tests for independent means were conducted to test the differences
between face-to-face and online counseling and social presence, mental
health status and working alliance. Levene’s test for equality of variances
was not significant for each dependent measure, ensuring that the data met
this assumption for each analysis. The total NMSPM mean scores were com-
pared between the independent groups of online and face-to-face counseling;
means and standard deviations are found in Table 3. A two-tailed t-test for
independent means was nonsignificant between the two counseling groups
[t(48) ¼ .515, n.s. p > .05]. This analysis indicated that both counseling groups
rated the social presence of the counseling relationship similarly. Composite
GHQ-12 scores were compared between online and face-to-face counseling
participants, means and standard deviations can be found in Table 3. A
two-tailed t-test for independent means was nonsignificant [t(48) ¼
1.205,
n.s. p > .05]. These results indicate that the two groups had similar levels of
general mental health functioning at the time of the study.
A t-test for independent means was conducted to compare total scores
on the WAI-S between online and face-to-face counseling participants. This
analysis indicated a significant difference between the two groups [t(48) ¼

2.017, p < .05]. These data revealed that the online counseling participants
rated their working alliance significantly stronger than did the face-to-face
counseling participants. A Cohen’s d statistic was used to examine the effect
size of this outcome. An effect size of .63 was found for this analysis, indicat-
ing a small result in the difference between the two groups. With a sample size
of 13 and 37, a d coefficient of .63, and a probability level of .05, the power is
.48. Power should be around .80 to avoid a significant chance of error. It is
probable that the small sample size led to the low power statistic.

TABLE 3 Results for all t-tests for independent means (N ¼ 50)

Online Face-to-Face

M SD M SD t

Networked Minds Social Presence Measure 185.69 22.00 189.43 22.72 .515
General Health Questionnaire-12 22.92 6.36 20.68 5.58
1.205
Working Alliance Inventory-Short
Composite score 73.85 8.72 67.62 9.84
2.017
Task subscale 23.85 3.74 22.32 2.97
1.484
Bond subscale 25.46 2.88 23.41 3.66
1.834
Goal subscale 24.54 3.41 21.89 3.84
2.196

Note: Higher scores reflect higher self-reports of social presence and working alliance; Lower scores reflect
a higher quality of mental health Online (n ¼ 13); Face-to-face (n ¼ 37).

p < .05.
24 C. Holmes and V. Foster

Due to the significant difference found in the composite score, separate

t-tests for independent means were computed for Task, Bond, and Goal sub-
scales between the two counseling groups to further explore the cause of the
total test mean score differences. Mean and standard deviations are found in
Table 3. All three subscales indicated nonsignificant Levene’s tests for variance
equality, thereby indicating equal variances across samples. The Task sub-
scale analysis indicated no significant difference between the two groups
[t(48) ¼
1.484, n.s. p > .05]. The Bond subscale analysis also indicated no
significant difference between the two groups [t(48) ¼
1.834, n.s. p > .05];
however, the Goal subscale analysis indicated a significant difference
between the two groups [t(48) ¼
2.196, p < .05]. These data indicate that
online counseling clients perceived a significantly stronger alliance through
the Goal subscale than the face-to-face counseling participants. The effect
size, as indicated by Cohen’s d statistic, showed an effect size of .68 indicating
a small effect. The online counseling group indicated a stronger perception
that the goals for treatment were agreed upon and shared between the client
and counselor, and that the goals are mutually known and understood. With a
sample size of 13 and 37, a d coefficient of .68, and a probability level of .05,
the power is .54 for the Goal subscale. Again, it is likely that a small sample
size led to low power. Given a similar effect size of the two statistically signifi-
cant results and a similar participation allocation ratio, the minimum sample
size for finding statistical significance on the Bond subscale, Task subscale,
the NMSPM, and the GHQ-12 is 106 participants. A larger effect size could
be reached through increasing the sample size.
T-tests for independent means were conducted on the data when the two
hybrid participants were deleted. These analyses were computed to compare
the online and face-to-face only participants on the WAI-S total score and the
Goal subscale to determine if the results would remain significant without the
hybrid clients. The analyses indicated nonsignificant Levene’s tests for vari-
ance equality, thereby indicating equal variances across samples. The WAI-S
analysis indicated a significant difference between the online and face-to-face
groups [t(46) ¼
2.237, p < .05]. These results are located in Table 4. The
Cohen’s d statistic showed a small effect size of .66. The Goal subscale analysis

TABLE 4 Results for t-tests for independent means on the WAI-S and Goal Subscale (N ¼ 48)

Online Face-to-Face

M SD M SD t

Working Alliance Inventory-Short

Composite score 75.00 8.69 67.62 9.84
2.237
Goal subscale 24.73 3.69 21.89 3.84
2.167

Note: Higher scores reflect higher self-reports of working alliance Online (n ¼ 11); Face-to-face (n ¼ 37).

p < .05.
 Online and F2F Counseling 25

indicated a significant difference between the two groups [t(46) ¼
2.167,

p < .05]. The Cohen’s d statistic showed a small effect size of .64. With a
sample size of 11 and 37, a d coefficient of .66 and .64, respectively, and a
probability level of .05, the power is .47 and .45 for the two analyses. Power
analyses also showed ideal sample sizes to reach statistical significance to be
96 participants. Future studies might be able to find robust results if sample
sizes reached this number.
Pearson correlations were used to analyze the relationships between the
dependent measures for the participants in online counseling, face-to-face
counseling, and as a total sample. All correlational analyses were nonsignifi-
cant except for a Pearson correlation comparing the total NMSPM and WAI-S
scores within the face-to-face counseling group and in the total sample. A sig-
nificant positive correlation was found between the total scores of the NMSPM
and the WAI-S in the total sample (r ¼ .43, p < .01), as well as between the
total scores of the NMSPM and the WAI-S in the face-to-face counseling group
(r ¼ .53, p < .01). This analysis shows that the working alliance and social
presence constructs share similar qualities both for the total sample and the
face-to-face counseling group. More research is needed to understand the
relationship and meaning behind the significant overlap between the two
constructs in the total sample. Additional inquiry into why a significant
relationship existed within the face-to-face setting and was not shown within
the online counseling group is also warranted. This distinction may have been
highlighted due to the small sample size in the online counseling group.
Future research with a larger sample might clarify this finding.

DISCUSSION

Counseling, from a Rogerian perspective, is a profession built upon the rela-

tional aspect of human nature and on the assumption that if a facilitative,
‘‘growth-promoting climate’’ can be fostered within the therapeutic relation-
ship, a client can utilize innate resources for self-understanding and personal
growth to produce personal change (Rogers, 1986). However, the importance
of this perceived human-to-human relationship creates a theoretical rift when
conceptualizing the utility of online counseling. According to Langas (2005),
‘‘One of the biggest challenges facing the field of e-therapy today is how the
relationship that develops between therapist and client over the internet can
fulfill its important function in the process of therapy’’ (p. 3). To that end, the
current study examined the concepts of social presence and working alliance
within the context of an ongoing therapeutic relationship, including how
these concepts are similar or different when compared between online coun-
seling and face-to-face counseling. However, due to the preliminary nature of
the study and the small, convenience sample, results should be interpreted
with caution.
26 C. Holmes and V. Foster

Discussion of Findings
Chi square analyses found no significant differences between the two client
groups based on demographic factors such as age, gender, race=ethnicity,
and education level. These results indicate that for this sample, significant cli-
ent differences in these areas did not exist. Some literature touts the ability for
online counseling to reach underserved populations (Layne & Hohenshil,
2005; Ragusea & VandeCreek, 2003), although certain studies have shown
the overrepresentation of White, educated women as those most often parti-
cipating in online counseling studies (Cook & Doyle, 2002; Leibert et al.,
2006). The current data show that while, indeed, the general sample was
mostly White females between the ages of 26 and 40, there was no significant
different between the clients in each type of counseling, indicating that there
may be little demographic difference between online and face-to-face coun-
seling clients in the current study. However, research volunteers in general are
more likely to be female and educated, which may have played a role in the
obtained sample (Gall, Gall, & Borg, 2007).
As little to no research exists in the literature examining social presence
within the context of online counseling, this study sought to establish a base-
line understanding of the role social presence may play in online counseling
relationships when compared to face-to-face counseling relationships. The
results indicate that both groups rated the perceived social presence of their
counselors similarly. As social presence is defined by, ‘‘ . . . the moment-by-
moment awareness of the co-presence of another sentient being accompanied
by a sense of engagement with the other’’ (Biocca et al., 2001, p. 2), the current
data suggest that clients who developed relationships with their counselor
through computer-mediated communication perceived a similar sense of
awareness and engagement of their counselor at the time of the study.
These results coincide with Walther (1995), who found that users of
computer-mediated communication felt a similar sense of social presence as
users of face-to-face communication. Biocca et al. (2001) assert that psycho-
logical, emotional, and behavioral components are involved in computer-
mediated communications, even though they lack a face-to-face component,
as was the report of clients in the current study who expressed a similar level
of perceived social presence regardless of the counseling modality used. Such
findings illustrate the need for a reevaluation of the ‘‘essential attributes that
are needed to establish connection with others’’ (Biocca, Harms, & Burgoon,
2003, p. 459), as many of these attributes appear to be supported through
computer-mediated communication.
The working alliance encompasses an operationally defined and
measured piece of the overall therapeutic relationship, one that speaks to
the collaboration between the client and counselor, working together in ther-
apy based on the acceptance of mutual goals and trust (Horvath & Bedi, 2002).
When the working alliance scores were compared between the online and
 Online and F2F Counseling 27

face-to-face counseling groups, the online counseling participants reported a

significantly stronger working alliance, supporting the notion that a working
alliance can be formed whether or not the client and counselor share the same
physical space. These data are supplementary to findings reported in previous
research contending that online counseling clients perceived the working
alliance equally as strong as or stronger than face-to-face clients (Cook &
Doyle, 2002; Knaevelsrud & Maercker, 2006; Reynolds, Stiles, & Grohol, 2006).
The current findings further indicate a shift in the perception of connec-
tion and interpersonal awareness formed between two people through elec-
tronic means. The data do not render the face-to-face therapeutic relationship
obsolete or conclude that online counseling can offer the identical nuance of a
traditional relationship (Langas, 2005). However, these findings clearly sug-
gest that online formats can supplement or be useful in lieu of face-to-face
counseling. After the total working alliance scores were found to be signifi-
cantly different, comparisons were run between the groups and each of the
three subscales, Task, Bond, and Goal. The scores of the Goal subscale were
significantly different, indicating that online counseling users perceived a
stronger alliance with regard to mutual understanding of and work toward
specified counseling goals; similar results were also found by Cook and Doyle
(2002). The current findings, although preliminary, add support to the conten-
tion that online counseling provides a space for therapeutic relationships to
form. Online therapy, particularly synchronous chat and e-mail, may be parti-
cularly effective in setting agreed upon and clear goals between the counselor
and client due to the written nature of the sessions. The concrete nature of set-
ting goals in writing may be helpful in both client and counselor forming
mutual understanding. A larger sample comparing various types of online
therapy would allow future research to analyze potential differences.

Limitations
The current study was designed as a survey, asking participants to answer
only once about their experiences and thus did not study the measures long-
itudinally to see how or if the working alliance, general mental health status,
and social presence changed over time through therapy. The results collected
are indicative of the overall feelings of participants only at the specific time
they took the study. Furthermore, the study did not directly link to any parti-
cular counseling strategies. The sample size of this study was small, which
may potentially confound the results. The post hoc power analyses showed
the significant findings to have relatively low power, indicating the low num-
ber of participants may have impacted the current study’s findings. This limi-
tation also highlights the challenges of online recruitment, as it proved difficult
to reach potential participants via only online means. Additionally, two of the
13 online counseling clients had been involved in both online and face-to-face
counseling and these clients were identified as online counseling clients in
28 C. Holmes and V. Foster

the analyses. This distinction could also have confounded the results. Finally,
the authors could not control for many individual differences such as parti-
cular counselor qualities or type of diagnosis.

Recommendations for Further Study

Social presence theory offers a new way of understanding nuanced similari-
ties and differences of the two modalities of counseling regarding relational
aspects based on the type of the communication. The current study’s results
found no significant differences between the two types of counseling, indicat-
ing that a similar level of social presence can be perceived when comparing
the different modalities of counseling. Studies recruiting more participants
from the various modalities of online counseling would allow a valid compari-
son of the social presence experienced by participants in the different online
counseling modalities (i.e., e-mail, synchronous chat, videoconferencing, and
telephone). Extended data collection with more systematic, targeted recruit-
ing procedures regarding client or counselor characteristics may help to attain
more representative samples. Furthermore, studies targeting online counsel-
ing practitioners could yield a wealth of information, including comparisons
of theoretical orientation and other particular counselor characteristics.
Additionally, extensive validation of the Networked Minds Social Pres-
ence Measure is necessary to refine a reliable and valid instrument that mea-
sures the particular construct of social presence in counseling. Supplementary
analyses to understand the relationship between mediated social presence
and the working alliance is warranted. The current study found a significant
relationship between the two constructs both for the total sample and for
the face-to-face counseling group indicating a potential new area for under-
standing the utility of online counseling. Although the online counseling
group did not show a significant correlation between the working alliance
and social presence constructs, a larger sample would be useful in gathering
more data. Another potential investigation is the relationship between social
presence and overall satisfaction with computer-mediated communication. As
it has been shown in previous literature (Gunawardena & Zittle, 1997), per-
ceived social presence may play a role in how consumers perceive and feel
about their overall satisfaction with the relationships built through online
counseling.
Finally, future studies should work to increase sample size and hetero-
geneity of participants. Small samples and homogenous self-selected volun-
teers limit past studies, including the current study. Statistical power was
significantly impacted by the small number of participants making it imposs-
ible to depict more than preliminary information with the current project.
Larger sample sizes would support further analyses and increase the statistical
power for investigating current and future research questions. Additionally,
including a covariate and performing ANCOVA (Analysis of Covariance) with
 Online and F2F Counseling 29

a larger sample size will provide more insight into the differences and simila-
rities in counseling modalities. If possible, this research should attempt
experiential designs to control for specific variables and the ability to produce
robust results further impacting the field beyond preliminary data.

CONCLUSION

Counseling is currently in an era of change as computer-mediated services

are evolving and developing within the field, even ahead of a firm research
base. Regarding the field works to match research with practice, Lundberg
(2001) aptly notes:

Counselors are still groping for an effective merger between an increas-

ingly technological world and a profession that is practiced through very
personal contact. More fundamentally, human service professionals are
still trying to understand just how human relationships change when they
are mediated by computers. (p. 142)

The results of this study indicated that a working alliance can be formed
through computer-mediated communication, and that clients perceive a simi-
lar social presence of their counselor regardless of the modality of counseling.
The data clearly imply that a therapeutic bond can be formed if a client and
counselor do not share the same physical space and thus adding preliminary
data to support the use of online counseling, while offering a useful ground-
ing for further study.

REFERENCES

Ainsworth, M. (2002). My life as an e-patient. In R. Hsuing (Ed.), E-Therapy: Case

studies, guiding principles, and the clinical potential of the internet (pp. 194–
216). New York, NY: W.W. Norton and Company, Inc.
Aragon, S. R. (2003). Creating social presence in online environments. In S. Ismel &
J. Ross-Gordon (Eds.), New directions for adult and continuing education 100
(pp. 57–68). San Francisco, CA: Jossey-Bass.
Biocca, F., Harms, C., & Burgoon, J. (2003). Towards a more robust theory and mea-
sure of social presence: Review and suggested criteria. Presence: Teleoperators
and Virtual Environments, 12(5), 456–480.
Biocca, F., Harms, C., & Gregg, J. (2001). The networked minds measure of social
presence: Pilot test of the factor structure and concurrent validity. Retrieved
from www.mindlab.org
Borgmann, A. (1984). Technology and the character of contemporary life: A
philosophical inquiry. Chicago, IL: University of Chicago Press.
Busseri, M., & Tyler, J. (2003). Interchangeability of the working alliance inventory
and working alliance inventory, short form. Psychological Assessment, 15(2),
193–197.
30 C. Holmes and V. Foster

Carmel, M. S., & Friedlander, M. L. (2009). The relation of secondary traumatization

to therapists’ perceptions of the working alliance with clients who commit
sexual abuse. Journal of Counseling Psychology, 56(3), 461–467.
Chester, A., & Glass, C. (2006). Online counselling: A descriptive analysis of
therapy services on the internet. British Journal of Guidance & Counselling,
34(2), 145–160. doi:10.1090=03069880600583170.
Cook, J., & Doyle, C. (2002). Working alliance in online therapy as compared
to face-to-face therapy: Preliminary results. Cyberpsychology & Behavior, 5(2),
95–105.
Duncan, B., Miller, S., Wampold, B., & Hubble, B. (2010). The heart and soul of
change: Delivering what works in therapy. Washington, DC: American Psycho-
logical Association.
Gall, J. P., Gall, M. D., & Borg, W. R. (2007). Educational research: An Introduction,
(8th ed.). Needham Heights, MA: Allyn & Bacon.
Goldberg, D., & Williams, P. (1988). A user’s guide to the General Health Question-
naire. Windsor, UK: NFER-Nelson.
Grohol, J. (1998). Future clinical directions: Professional development, pathology, and
psychotherapy on-line. In J. Gackenback (Ed.), Psychology and the internet: Intra-
personal, interpersonal, and transpersonal implications (pp. 111–140). San Diego,
CA: Academic Press.
Gunawardena, C. N., & Zittle, F. J. (1997). Social presence as a predictor of satisfac-
tion within a computer-mediated conferencing environment. The American
Journal of Distance Education, 11(3), 8–26.
Hanley, T., & Reynolds, D. (2009). Counselling Psychology and the internet: A
review of the quantitative research into online outcomes and alliances within
text-based therapy. Counselling Psychology Review, 24(2), 4–13.
Harms, C., & Biocca, A. (2004). Internal consistency and reliability of the networked
minds social presence measure. In M. Alcaniz & B. Rey (Eds.), Seventh Annual
International Workshop: Presence 2004. Valencia, Spain: Universidad Politenc-
nica de Valencia.
Horvath, A. O., & Bedi, R. P. (2002). The alliance. In J. Norcross (Ed.), Psychotherapy
relationships that work: Therapist contributions and responsiveness to patients.
New York, NY: Oxford University Press.
Horvath, A. O., & Greenberg, L. S. (1989). Development and validation of the Working
Alliance Inventory. Journal of Counseling Psychology, 36, 223–233.
Hubble, M., Duncan, B., & Miller, S. (1999). The heart and soul of change: What
works in therapy. Washington, DC: American Psychological Association.
Knaevelsrud, C., & Maercker, A. (2006). Does the quality of the working alliance
predict treatment outcome in online psychotherapy for traumatized patients?
Journal of Internet Medical Research, 8(4), 1–13. doi:10.2196=jmir.8.4.e31
Kokotovic, A. M., & Tracey, T. J. (1989). Factor structure of the Working Alliance
Inventory. Psychological Assessment. A Journal of Consulting and Clinical
Psychology, 1, 207–210.
Langas, E. (2005). Online therapy: Conceptualizing the therapeutic relationship via
e-therapy. Paper presented at the 40th APS Annual Conference, Melbourne,
Australia.
 Online and F2F Counseling 31

Layne, C. M., & Hohenshil, T. H. (2005). High tech counseling: Revisited. Journal of
Counseling and Development, 83, 222–226.
Leibert, T., Archer, J., Munson, J., & York, G. (2006). An exploratory study of client
perceptions of internet counseling and the therapeutic alliance. Journal of
Mental Health Counseling, 28(1), 69–83.
LoBello, S. (1995). Review of the General Health Questionnaire. The Mental Measure-
ments Yearbook. Lincoln, NE: The Buros Institute.
Lundberg, D. J. (2001). Integrating on-line technology into counseling curricula:
Emerging humanistic factors. Journal of Humanistic Counseling, Education &
Development, 38(3), 142–155.
Mallen, M., Vogel, D., Rochlen, A., & Day, S. (2005). Online counseling: Reviewing
the literature from a counseling psychology framework. The Counseling
Psychologist, 33(6), 819–871.
Mathers, N., & Shipton, G. (1993). The impact of short-term counseling on general
health questionnaire scores. British Journal of Guidance & Counselling,
21(3), 310–321.
Pelling, N. (2009). The use of Email and the Internet in counselling and psychologi-
cal service: What practitioners need to know. Counselling, Psychotherapy, and
Health (The Use of Technology in Mental Health Special Issue), 5(1), 1–25.
Ragusea, A. S., & VandeCreek, L. (2003). Suggestions for the ethical practice of
online psychotherapy. Psychotherapy: Theory, Research, Practice, Training,
40, 94–102.
Reynolds, D., Stiles, W., & Grohol, J. (2006). An investigation of session impact and
alliance in internet based psychotherapy: Preliminary results. Counseling and
Psychotherapy Research, 6(3), 164–168.
Rogers, C. R. (1957). The necessary and sufficient conditions of therapeutic person-
ality change. Journal of Counseling Psychology, 22, 95–103.
Rogers, C. R. (1986). Client-Centered Therapy. In I. Kutash & A. Wolf (Eds.),
Psychotherapist’s Casebook (pp. 197–208). San Francisco, CA: Jossey-Bass Inc.
Short, J., Williams, E., & Christie, B. (1976). The social psychology of telecommunica-
tions. London, UK: John Wiley & Sons, Ltd.
Walther, J. B. (1995). Relational aspects of computer-mediated communications:
Experimental observations over time. Organization Science, 6(2), 182–203.
Copyright of Journal of Technology in Human Services is the property of Taylor & Francis Ltd and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express
written permission. However, users may print, download, or email articles for individual use.