Computers in Human Behavior 39 (2014) 197–203
Contents lists available at ScienceDirect
Computers in Human Behavior
journal homepage: www.elsevier.com/locate/comphumbeh
Text-based communication influences self-esteem more than
face-to-face or cellphone communication
Amy L. Gonzales ⇑
University of Pennsylvania, Robert Wood Johnson Foundation Health and Society Scholars, 3641 Locust Walk, Philadelphia, PA 19104, United States
a r t i c l e
i n f o
Article history:
Keywords:
Text-based communication
Internet
Self-esteem
Hyperpersonal Model
Interpersonal self-disclosure hypothesis
Ecological momentary assessment
a b s t r a c t
Meaningful social interactions are positively associated with improvements in self-esteem, but this
phenomenon has largely been unexplored in digital media despite the prevalence of new, text-based
communication (e.g. Facebook, texting, email, etc.). To address this gap in the literature the frequency
and quality, or meaningfulness, of communication was measured in mediated and non-mediated channels
across a random sample of 3649 social interactions using Experience Sampling Methods. Results revealed
that most communication took place face-to-face (62%), with less text-based (about 22%) and cell phone
voice (14%) communication. Meaningful face-to-face and text-based communication were associated
with changes in self-esteem according to a marginally significant and significant finding, respectively.
Text-based communication was more important for self-esteem than face-to-face or phone communication, which is consistent with research on the magnifying effect of text-based communication on interpersonal processes. According to the Internet enhanced self-disclosure hypothesis, the psychological
benefits of text-based communication stems from enhanced self-disclosure, which is also supported in
the data. Additional work is needed to better understand the mechanisms underlying the positive
relationship between meaningful text-based interactions and self-esteem, but findings point to the
important role of digital communication for psychological health.
Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction
Since the inception of the social sciences, researchers have been
trying to understand the effect of social interaction on mental and
physical well-being (Durkheim, 1897). Over the last century we
have learned that having strong support networks is associated
with better cardiovascular, neuroendocrine (i.e. stress), and
immune system functioning (for review see Uchino, Cacioppo, &
Kiecolt-Glaser, 1996). Also, socially isolated people are at a greater
risk of being ill and they die sooner than socially connected people
(House, Landis, & Umberson, 1988). In other words, having a
supportive social network is a key part of maintaining good psychological and physical health. This investigation takes on new
meaning, however, in light of recent advances in communication
technology. With 85% of the world subscribed to a mobile number,
and 33% connected to the Internet, access to friends and loved ones
is nearly constant for much of the world’s population
(International Telecommunication Union (ITU), 2011).
⇑ Address: 1229 E. 7th St, Department of Telecommunications, Indiana University, Bloomington, IN 47401, United States. Tel.: +1 812 856 9051.
E-mail address: [email protected]
http://dx.doi.org/10.1016/j.chb.2014.07.026
0747-5632/Ó 2014 Elsevier Ltd. All rights reserved.
Growth in the number of ways to connect has simultaneously
given rise to fears that new technologies undermine our capacity
for meaningful social connections (e.g. Turkle, 2011). But as
anxieties about the negative effects of technology mount so does
evidence of potential benefits to mental and physical health (e.g.
Byrne et al., 2012; Hampton, Sessions-Goulet, Rainie, & Purcell,
2011; Rains & Young, 2009; Valkenburg & Peter, 2009b). At the
center of this debate persist the following questions: Do digital
communication systems allow for meaningful social interactions
that are critical for maintaining quality of life? If so, how do they
compare to the benefits of face-to-face interactions?
One important factor associated with well-being and quality of
life is self-esteem. Self-esteem is a determinant of mental and
physical health (Ford & Collins, 2010; Trzesniewski et al., 2006),
and is improved when people have supportive social ties (Cohen,
1988). Although previous research has explored the effect of the
frequency of Internet use on self-esteem, no studies have
examined the relationship between the quality or meaningfulness
of digital interactions and changes in self-esteem. Moreover, most
studies to-date on the digital media effects of self-esteem have
been descriptive, do not compare channels (e.g. cell phone voice,
text, face-to-face), and rely primarily on survey data. However,
channel comparisons, or comparisons in the effects of the mode
198
A.L. Gonzales / Computers in Human Behavior 39 (2014) 197–203
of communication, are important because they help to refine computer-mediated communication theory (see Walther, 2012), and
surveys are an effective tool for collecting large generalizable
samples, but they are also biased towards a respondent’s most
prominent and recent experiences (Kahneman, 1999). To address
these concerns, I used ecological momentary assessment (EMA)
to capture everyday social interactions as they occured.
EMA is a methodology that involves the ‘‘repeated collection of
real-time data on subjects’ behavior and experience in their natural
environments’’ (Shiffman, Stone, & Hufford, 2008, p. 3). EMA is
another term used to describe diary methods that have been used
for decades, the accuracy of which may be improved with the use
of handheld digital recording devices. EMA provides enhanced
ecological validity, minimizes risks of recall biases, and enables
assessment of within-person processes. Individuals act as their
own controls by prompting participants to report on social interactions as they occur (Almeida, 2005; Bolger, Davis, & Rafaeli, 2003).
Findings from this dataset thus enhance the internal validity of
research on self-esteem effects of new media. They also update
earlier work examining the frequency of communication across
different channels, and variations in the quality of that communication (Baym, Zhang, & Lin, 2004).
1.1. Social connection and self-esteem
Self-esteem, self-worth, and other forms of self-regard are
essential components of overall quality of life (Cohen, 1988).
Self-esteem in this study is defined as the ‘‘value people place on
themselves’’, and is presumed to vary as a function of different factors, including relationship quality (Baumeister, Campbell,
Krueger, & Vohs, 2003). Greater self-esteem has been associated
with financial successes, greater overall happiness, and better
physical health (Baumeister et al., 2003; Martens et al., 2010;
Trzesniewski et al., 2006). In other words, self-esteem is a central
component of overall psychological health and well-being.
A substantial body of research suggests that the quality of one’s
relationships can alter levels of self-esteem, suggesting that selfesteem can be thought of both as a trait, but also as a state
(Blackheart, Nelson, Knowles, & Baumeister, 2009; Wills, 1985).
Sociometer theory, for example, argues that changes in self-esteem
are designed to gauge whether or not social relationships are functioning properly (Leary, Tambor, Terdal, & Downs, 1995). If relationships are working self-esteem will improve or remain stable.
If social relationships are strained or an individual is isolated
self-esteem will drop. Specific interpersonal processes that have
been shown to improve self-esteem include supportive listening,
mutual sharing, and an absence of critiquing (Wills, 1985). A
meta-analysis of 72 sociometer articles found that positive interpersonal experiences, more than exclusion, have a notable effect
on self-esteem (Blackheart et al., 2009), and a recent diary study
supporting sociometer theory found that high quality interactions
throughout the day predict corresponding daily changes in selfesteem (Denissen, Penke, Schmitt, & van Aken, 2008). This association between fluctuations in interpersonal dynamic and selfesteem is consistent with other research that finds a link between
the quality of daily social interaction and changes in outcomes
such as health, relationship satisfaction, and daily well-being
(Emmers-Sommer, 2004; Reis, Sheldon, Gable, Roscoe, & Ryan,
2000; Reis, Wheeler, Kernis, Spiegel, & Nezlek, 1985).
1.2. Digital technology and self-esteem
Despite consistent findings of the benefits of meaningful
social interaction for self-esteem offline, research on the relationship between digital communication and self-esteem has
been mixed. Some research finds that people with low self-
esteem have a preference for Internet communication (Joinson,
2004; Niemz, Griffiths, & Banyard, 2005), though experimental
work has found that online communication can be beneficial
for self-esteem (Gross, 2009; Shaw & Gant, 2002). A subset of
this work has focused on online social networking, also with
mixed results (Gonzales & Hancock, 2011; Kalpidou, Costin, &
Morris, 2011). Variation in these findings is likely due to variation in individual differences and the quality of communication
that takes place online, as quality of communication has rarely
been measured in these studies (see Valkenburg, Peter, &
Schouten, 2006).
Most of the research above has explored the frequency of Internet use and self-esteem, but has not examined the relationship
between the meaningfulness of digital social interactions and
changes in self-esteem. Although the nature of social interaction
has long been known to have an effect on changes in self-esteem,
such that positive social interactions are good for self-esteem while
negative social interactions may be bad for self-esteem (Leary
et al., 1995; Blackheart et al., 2009), this relationship has been largely unexplored online. Yet there is reason to believe that these
processes may be different online.
First, the Hyperpersonal Model claims that interpersonal processes may lead to enhanced or hyperpersonal impression formation in text because users are able to carefully construct
messages, and attributions caused by those messages may be exaggerated in the absence of additional real-time visual and audio
cues (Walther, 1996). Indeed, studies have demonstrated that
online content has a more substantial effect on interpersonal
impressions (Epley & Kruger, 2005; Jiang, Bazarova, & Hancock,
2011), and may also influence impressions of the self (Gonzales
& Hancock, 2011). From this perspective then, the enactment of
supportive listening and mutual sharing that has been shown to
improve self-esteem offline may be perceived as even more supportive and intimate when it takes place in text, which could have
an enhanced effect on self-esteem.
Another communication theory that may inform differences in
the relationship between meaningful communication and selfesteem in text-based contexts is the Internet-Enhanced Self-Disclosure (IESD) hypothesis. The IESD hypothesis proposes that adolescents often disclose more in text-based communication than
face-to-face communication, which contributes to greater wellbeing (Valkenburg & Peter, 2009a). Indeed a number of studies
have confirmed that people often disclose more online compared
to offline (Jiang et al., 2011; Valkenburg & Peter, 2009b; Tidwell
& Walther, 2002), and greater disclosure is associated with more
supportive relationships, which are good for self-esteem (Reis
et al., 2000; Uchino et al., 1996; Wills, 1985). If people disclose
more in text-based communication there is reason to suspect that
text-based communication may play an even greater role in shaping self-esteem than offline communication.
Another contribution of these data is that they update and
extend previous work by Baym et al. (2004) on the frequency
and quality of communication across different channels, or modes
of communication. That study, conducted in 2004, found that faceto-face communication comprises over 60% of social interactions,
and, along with telephone communication, is perceived as slightly
better quality than Internet communication. It is one of few studies
to appreciate the use of multiple media channels by individuals. In
that study, the authors noted that little research had taken advantage of diary methods to do comparative, within-person analyses
of media use. Baym et al. (2004) established important groundwork for understanding the frequency and relative quality of communication across channels. I build on these findings by again
collecting diary data using a sample that is more economically,
racially, and age diverse, and by adding dependent measures of
self-esteem.
A.L. Gonzales / Computers in Human Behavior 39 (2014) 197–203
In sum, according to previous research on differences in textbased interpersonal communication, participants should disclose
more online than on the cell phone or face-to-face, and text-based
communication should be more valuable for self-esteem than faceto-face or cellphone communication. Also, if findings are consistent
with previous diary findings, face-to-face communication should
still constitute the majority of social interactions, and, along with
cell phone communication, should be rated as slightly more meaningful than text-based communication. These findings would
extend the Hyperpersonal Model and the IESD hypothesis, as well
as build on research from over a decade ago. To test this, I pose the
following hypotheses:
H1. Most communication will take place face-to-face.
H2. People will disclose more in text-based communication than
in face-to-face or cellphone communication.
H3. People will find face-to-face and cell phone communication
more meaningful than text-based communication.
H4. The relationship between meaningful communication and
self-esteem will be stronger in text-based communication than in
face-to-face or cellphone communication.
2. Methods
2.1. Participants
In this study, 98 people between the ages of 18–38 years participated for $90. Age was capped at 38 to minimize age effects. Participants were recruited using flyers across upper-, middle-, and
low-income neighborhoods throughout a northeastern urban area.
According to time stamped survey logs, any participant that completed 5 or more real-time surveys less than an hour apart was
excluded from analysis for not following instructions before any
hypothesis testing was performed. The final analytic sample was
comprised of 76 participants who completed over 3649 surveys.
Demographic details of the data are described further in Table 1.
2.2. Procedure
Upon arriving at the lab, participants completed baseline
measures of self-esteem, along with additional psychological measures not included in this analysis. Participants were then provided
Table 1
Independent and dependent descriptive variables.
Gender (F)
Age
Race
African–American/black
Asian
Latino
White
Education
6Highschool
Community college/training school
PBachelors degree
Knew another participant
Recorded interactions per person
Baseline self-esteem
Follow-up self-esteem
Overall interaction meaningfulness
Mean/%
SD
Range
54%
25
6.20
18–38
20.8
1.20
1.16
0.70
20–110
1–7
1–7
1–5
38%
19%
4%
39%
22%
16%
62%
19%
49
5.36
5.36
2.96
Note: Education: participants are enrolled in or have completed as their highest
level of education. Text-based meaningfulness is comprised of texts, emails, and
Facebook messages.
199
with a personal information booklet describing the purpose and
procedures of the study. Participants were told that the study
was designed to examine, ‘‘who people talk to and how often are
they using technology?’’ Each participant was given a Palm Pilot
that was pre-programmed with 6–10 random alarms for 6 days
within the participants’ unique waking hours. According to Stone
and Shiffman (2002), random EMA sampling, ‘‘is the best way to
obtain a representative sample’’ (p. 238). They argue that the frequency of daily sampling and diary study length can vary by study
(Stone & Shiffman, 2002).
Each time a pre-programmed Palm Pilot alarm rang participants
were asked to complete a short survey on the Palm Pilot about her
or his two most recent social interactions, including the channel in
which it took place and the quality of the social interaction. If participants had not had a social interaction since last completing the
survey they did not complete the full survey. Social interactions
were defined broadly as ‘‘anything that is social,’’ though it was
made explicit that this included asynchronous digital interactions,
or ‘‘half of an interaction (e.g. writing a text, reading a Facebook
post, etc.).’’ Participants were told that the minimum requirement
for participation in the study was 20 completed surveys during the
6 days. The Palm Pilot was chosen as a data collection tool because
it time-stamped each survey without enabling new forms of digital
communication.
Participants left with the Palm Pilot and instruction booklet and
were asked to return to the lab in 7 days for follow-up surveys.
Upon returning participants completed the same battery of psychological surveys, including a follow-up measure of self-esteem
that was identical to the baseline measure of self-esteem, which
was used as the primary dependent variable in the analysis. Participants were then compensated for their participation and
debriefed.
2.3. Measures
2.3.1. Self-esteem
The ten-item Rosenberg self-esteem scale was used to evaluate
self-esteem at the beginning and end of the study (Rosenberg,
1965). Each item was assessed on a 7-point Likert scale and averaged to create a single measure of self-esteem for each person.
Items were reverse coded and the final index reflects a high score
for high self-esteem. This same procedure was followed for baseline measures of self-esteem (a = .93) and follow-up measures of
self-esteem (a = .92). Baseline measures of self-esteem were used
as a control variable and follow-up measures of self-esteem were
used as a dependent measure of self-esteem.
2.3.2. Communication channel
In each survey participants were first asked questions about
their most recent social interaction and then asked the same series
of questions about their second most recent social interaction. For
each social interaction, participants were asked, ‘‘how were you
interacting with that person?’’ Participants were given the following options from which to choose: face-to-face, landline telephone,
cellphone, email, texting, Facebook, video game, instant messaging,
online discussion board/forum, twitter, and other. If participants
chose other they were able to describe the channel in an openended response. Only the channels that were used at least 100
times across all people in the study were included in the analyses.
This choice was made before any analyses were performed, and
was made to better represent broadly used technologies. These
included face-to-face communication, cellphone voice communication, texts, emails, and Facebook communication. Inclusion of
other interactions in the analyses does not meaningful change
the results.
200
A.L. Gonzales / Computers in Human Behavior 39 (2014) 197–203
2.3.3. Meaningful social interactions & self-disclosure
Measurement of the quality of the social interaction was
adopted from the Rochester Interaction Record measure of meaningfulness which has been used in previous diary studies of the
positive effects of interpersonal communication (Reis et al.,
2000). Participants were asked to rate every social interaction
along the following dimensions: Intimacy, I shared, They Shared,
Quality, Satisfaction. Due to software constraints each item was
evaluated on a 1–5 scale, rather than a 1–7 scale (1 = Low,
5 = High). The 5 items were averaged to create a single measure
of meaningful communication for each person for each exchange
(a = .89). Items were then averaged across the week to create an
average interaction quality measure for each person in each channel. Finally, I shared was used as a single item to test for channel
differences in self-disclosure.
2.3.4. Participant demographic variables
Sex is a dichotomous variable, Female = 1 and Male = 0. Race is a
categorical variable comprised of three levels: White, Black or
Latino, and Asian. In all analyses, White is the reference category.
Education is a categorical variable comprised of three categories:
6high school, community college, or PBachelor’s degree. Participants were assigned a category if they were enrolled in or had
completed a given level of education. In all analyses, 6high school
education is the reference category.
3. Analytic approach
Stata 11 was used to perform these analyses, as it is well
equipped for analyzing mixed-models. To remain methodologically consistent, regression analysis was used to test all hypotheses
that required the use of inferential statistics (H2–H4).
To test the first hypothesis, the frequency of communication
was categorized and tallied by channel. To test the second
hypothesis, a multi-level regression model tested for differences
in self-disclosure across channels. In this analysis, self-disclosure
was the dependent variable regressed on a single categorical channel variable and within-person differences in self-disclosure were
accounted for. To test the third hypothesis, a multi-level regression
model tested for differences in meaningfulness across channels. In
this analysis, meaningfulness was the dependent variable
regressed on a single categorical variable and within-person differences in communication meaningfulness were accounted for.
Finally, because multi-level analysis does not allow for individual-level variables to predict variation in repeated measures,
meaningful communication was averaged for each person across
all channels and a series of OLS regression analyses were conducted to test the fourth hypothesis. The data were collapsed so
that the repeated measure of meaningful communication could
be analyzed as a predictor of the individual-level measure of selfesteem.
First, self-esteem at follow-up was regressed on communication
meaningfulness across all channels. This single variable, overall
meaningfulness, was then stratified into three different channel
variables: face-to-face meaningfulness, cellphone voice interaction
meaningfulness, and text-based interaction meaningfulness (SMS
texts, emails, and Facebook interactions). Next, self-esteem at
follow-up was regressed on face-to-face meaningfulness, call
phone meaningfulness, and text-based communication meaningfulness in three separate regression models. Finally all three channel variables were included in the same regression model to
compare the effect of communication meaningfulness in each
channel on self-esteem.
4. Results
The first hypothesis was an exploration of the frequency of
communication in each channel, as was first investigated by
Baym et al. (2004). In that study, 64% of communication took place
face-to-face, 18% of communication was by telephone, and 16% of
communication was online. In this study the majority of conversations took place face-to-face (62%); telephone communication
(SMS text, cell phone voice, landline voice) constituted 26% of
social interactions, with 98% of those exchanges utilizing a cell
phone; Internet communication constituted about 12% of social
interactions (Table 2). Face-to-face communication was the dominant mode of communication, which supports H1.
The second hypothesis explored whether an enhanced level of
self-disclosure takes place in digital contexts, per predictions of
the Internet Enhanced self-disclosure hypothesis. This hypothesis,
H2, was partially supported. A random effects regression model
reveals that participants reported disclosing less in face-to-face
communication (M = 2.80, SE = 0.08) than they did in text-based
communication (M = 2.96, SE = 0.09, b = 0.16, p = .003) or cellphone
voice communication (M = 3.10, SE = 0.08, b = 0.30, p < .001), after
accounting for all control variables.
To test the third hypothesis a random effects regression was
performed to conduct comparisons of channel meaningfulness,
with text-based communication as the comparison group and
demographic variables included as controls. Previous research
found that telephone and face-to-face communication were perceived as being better quality on average than Internet communication (Baym et al., 2004). Here there was no statistical difference
in the quality of text-based communication (M = 2.90, SE = 0.08,
range 1–5) and face-to-face communication (M = 2.96, SE = 0.08;
b = .05, p = 20). However, cellphone communication (M = 3.28,
SE = 0.08) was more meaningful than text-based communication
Table 2
Number of interactions reported in each communication channel.
Face-to-face
Cellphone voice
Texting
Email
Facebook
Chat
Landline
Videogame
Online group
Twitter
Other
Total
2273
521
405
184
147
50
23
11
4
2
29
3649
62%
14%
11%
5%
4%
1%
1%
<1%
<1%
<1%
1%
Note: Taken from 76 participants over 6 day intervals.
Table 3
Pearson correlation between self-esteem and meaningful communication by channel.
Overall meaningfulness (OM)
Face-to-face meaningfulness (FM)
Cellphone meaningfulness (CM)
Text-based meaningfulness (TM)
Self-esteem
OM
0.23
0.04
76
0.23
0.04
76
0.15
0.22
73
0.31
0.01
71
1
76
1
0
76
0.75
0
73
0.85
0
71
FM
CM
TM
1
76
0.75
0
73
0.85
0
71
1
73
0.65
0
68
1
71
Note: Below each Pearson correlation is the p value; below the p value is the sample
size for that analysis, n.
201
A.L. Gonzales / Computers in Human Behavior 39 (2014) 197–203
(b = .38, p < .001). These findings demonstrate mixed support for
H3.
Finally, to test the fourth hypothesis, a series of OLS regression analyses were conducted. Before conducting regression
analyses, zero-order correlations were calculated (Table 3).
Demographic covariates were then entered into the first
regression model (Table 4, Model 1). None of these variables
predict self-esteem at the end of the week and together make
for a poor fitting model, Adj. R2 = .04. Given model fit subsequent analyses omit demographic variables. The next step in
the hierarchical regression involved the inclusion of three
covariates: knowledge of others in the study, frequency of
social interactions, and baseline measures of self-esteem. Frequency of social interactions and knowing another participant
did not influence self-esteem, but were maintained because
of their conceptual relevance. Not surprisingly baseline selfesteem was highly correlated with self-esteem assessed at
the end of the week.
Before comparing the channel effects on self-esteem, four separate OLS regression analyses were conducted: an overall model and
a model for each individual channel. Findings reveal that people
who had more meaningful communication during the week across
all communication channels had better self-esteem at the end of
the week, after accounting for control variables (Table 4, Model
2). Also, meaningful face-to-face interactions during the week
had a marginally significant effect on self-esteem (Table 4, Model
3); cellphone interactions during the week did not have an effect
on self-esteem (Table 4, Model 4); and text-based interactions
(i.e. texts, emails, Facebook interactions) had a positive effect on
self-esteem after accounting for control variables, though the effect
size is small (Table 4, Model 5). Finally, to test H4, all three channel
variables were entered into the model simultaneously (Table 4,
Model 6). Only text-based interactions had a significant positive
effect on self-esteem. The effect size of meaningful text-based
communication on self-esteem is larger after accounting for meaningful communication in other channels. These findings support
H4.
5. Discussion
Previous research in psychology has demonstrated the importance of social interaction for self-esteem (Denissen et al., 2008;
Wills, 1985). Research on new media has found that frequency of
digital technology use is associated with both negative and positive
effects on self-esteem, depending on individual differences and
type of feedback (Joinson, 2004; Valkenburg et al., 2006). The
majority of this work, however, has not tested the relationship
between the quality of digital interactions and changes in selfesteem, despite communication theory that suggests that this process may be magnified online. Also, previous research has primarily
employed survey methods which have not enabled important
within person channel comparisons (Walther, 2012 for discussion),
and which may systematically bias results (Kahneman, 1999). To
address this, I employed repeated, real-time measures to capture
variability in the meaningfulness of social interaction across different channels as it is associated with self-esteem. Doing so also
allowed for a replication of previous descriptive research that
examined the frequency and quality of communication across multiple channels using within-person ratings (Baym et al., 2004).
Findings reveal that face-to-face was still the dominant form of
communication in this sample, despite the popularity of digital
communication. That is, in this study people still relied primarily
on face-to-face interactions to communicate. Yet, despite the dominance of face-to-face communication, participants disclosed more
in text-based channels than in face-to-face communication. Also,
text-based communication was more beneficial for self-esteem
than face-to-face or cellphone communication. This is consistent
with IESD predictions that computer-mediated communication,
which reduces in-person social pressures and puts greater emphasis on message content, can thus be more valuable for self-esteem
(Valkenburg & Peter, 2009a). These findings are also consistent
with Hyperpersonal Model predictions that impression formation
may be exaggerated in text (Walther, 1996), and as a result extend
the applicability of the Hyperpersonal Model to a new body of
research. That is, whereas most Hyperpersonal research only uses
Table 4
Relationship of communication meaningfulness by channel and self-esteem.
Model 1 control
Gender (F)
Age
Race
African–American/black
Asian
Latino
Education
6Highschool
Community/training
Baseline self-esteem
Knew another participant (1 = Y)
Frequency of all interactions
Freq. of face-to-face
Freq. of cellphone
Freq. of text-based
Overall meaningful communication
Meaningful FtF comm.
Meaningful cell comm.
Meaningful text comm.
0.06
0.03
0.90***
0.00
0.04
R2
0.81
Model 2 overall
Model 3 FtF
Model 4 cell
Model 5 text
Model 6 comparison
0.05
0.08
0.20
0.11
0.12
0.89***
0.00
0.03
0.89***
0.00
0.90***
0.00
0.87***
0.01
0.87***
0.02
0.06
0.02
0.04
0.03
0.09*
0.10*
0.15***
0.04
0.03
0.22**
0.84
0.84
0.07
0.82
0.82
0.82
Note: All coefficients are standardized. White is the comparison category for Race; Bachelors+ is the comparison category for Education. Community/Training refers to
community college or training colleges, such as beauty school. In Model 1, demographic variables were first tested alone, but not included in subsequent models. The second
set of covariates was then tested; R2 reflects these variables.
*
p 6 .10.
**
p 6 .05.
***
p 6 .01.
202
A.L. Gonzales / Computers in Human Behavior 39 (2014) 197–203
outcome measures related to interpersonal impression formation
this study uses that framework to understand the possible implications of interpersonal processes for the self. As a result the findings
underscore the value of text-based systems, particularly for selfesteem.
One benefit of text-based communication, in addition to being
psychologically beneficial, is that it is also incredibly accessible
for a growing number of people worldwide. It is often easier to text
loved ones than it is to talk face-to-face. Although modest, these
changes in self-esteem are a snapshot of possible long-term effects
of having positive (or negative) interactions with loved ones,
though additional research would be needed to empirically test
this. This possibility will only grow as more people use digital
technology to communicate more frequently. At minimum, these
findings indicate that digital technology is not an impoverished
substitute for meaningful face-to-face communication. Instead
these data suggest that digital, text-based technologies may be
one of the best ways to reap the psychological benefits of interpersonal communication for self-esteem.
Finally, in previous research Baym et al. (2004) found no statistical difference between the quality of cell phone and face-to-face
communication, and Internet communication was somewhat less
beneficial. In contrast, here there is no statistical difference
between the meaningfulness of Internet and face-to-face communication, and cell phone voice communication is the most meaningful. Researchers have demonstrated the interpersonal benefits
of cell phone communication (Christensen, 2009; Wei & Lo,
2006); perhaps the fact that telephone voice communication was
used much more infrequently in this sample relative to other forms
of communication means that telephone voice conversations are
increasingly reserved for more intimate or serious matters. That
is, these findings may reflect a normalization of text-based
communication and a new premium on cell phone voice communication, though additional work is needed to better understand
reasons behind differences in the quality of communication across
channels. For now, these findings suggest that mediated communication is comparable, or sometimes better, than face-to-face
communication in perceived quality.
5.1. Limitations
Despite enhancing validity, EMA methodologies are not without
limitations. First, in contrast to representative survey datasets, this
sample is not generalizable to a larger population. Also, EMA
allows for micro-level analyses of interpersonal behavior over
time, but the cost of data collection limits the size of the sample.
Second, these data cannot demonstrate causality. However, sociometer theory has demonstrated that changes in personal relationships influence self-esteem, and the Hyperpersonal Model and
IESD argue that computer-mediated interactions can intensify
communication. Together with these data, these theories suggest
that the quality of the interaction influences self-esteem, rather
than the other way around. On the other hand, if variation in
self-esteem affects the quality of digital communication more than
other forms of communication, this would also indicate an important role of digital media for interpersonal communication. In
either case, these findings suggest that researchers interested in
the link between social support and well-being, including
self-esteem, need to be mindful of how digital communication is
supporting and perhaps augmenting this process.
Third, the structure of this data precludes a test of self-disclosure as a mediator of channel effects on self-esteem. In order to
regress the level-two variable of self-esteem on the level-one
variable of meaningful communication, the data had to be collapsed across levels. Doing so, however, produces a self-disclosure
rating for each channel, making a test for mediation impossible.
The finding that people disclose more in text-based environments
supports an IESD hypothesis, and the results are consistent with
other work in this area (Valkenburg & Peter, 2009a).
Finally, it is unclear why the same pattern of self-esteem
improvement was not seen for cellphone conversations, especially
given that participants disclosed more in cell phone conversations
and cell phone communication was on average rated as more
meaningful than text-based communication. One possibility is that
a ceiling effect is constraining statistically significance due to the
high average meaningfulness of cell phone communication. Future
research should continue to explore the novel features of cellphone
communication as a dominant and evolving channel of accessing
social support, including as it relates to self-esteem.
6. Conclusion
Over the last two decades news media and researchers alike
have noted possible negative consequences of increased reliance
on digital technology for supportive communication (e.g. Turkle,
2011). And although it is valuable to question the implications of
these societal changes, results from this study indicate that digital
communication, though common, is not replacing face-to-face
communication. Moreover, text-based communication such as
email, SMS texts, and Facebook exchanges, can be equally if not
sometimes more important for well-being outcomes such as individual self-esteem. In other words, text-based communication may
be good for us, but it is probably not eliminating the need for
face-to-face contact. Hopefully this study will foster more research
on the mental and physical benefits of computer-mediated communication, as additional theoretical research is needed in this
area. Also, for people interested in designing technologies, these
findings speak to the potential of new text-based technologies to
improve quality of life.
Acknowledgements
This research was funded by a grant from the Robert Wood
Johnson Foundation.
References
Almeida, D. M. (2005). Resilience and vulnerability to daily stressors assessed via
diary methods. Current Directions in Psychological Science, 14, 64–68. http://
dx.doi.org/10.1111/j.0963-7214.2005.00336.x.
Baumeister, R. F., Campbell, J. D., Krueger, J. I., & Vohs, K. D. (2003). Does high selfesteem cause better performance, interpersonal success, happiness, or healthier
lifestyles? Psychological Science in the Public Interest, 4, 1–44. http://dx.doi.org/
10.1111/1529-1006.01431.
Baym, N. K., Zhang, Y. B., & Lin, M. (2004). Social Interactions across media:
Interpersonal communication on the internet, face-to-face, and the telephone.
New
Media
&
Society,
6(3),
299–318.
http://dx.doi.org/10.1177/
1461444804041438.
Blackheart, G. C., Nelson, B., Knowles, M. L., & Baumeister, R. F. (2009). Rejection
elicits emotional reactions but neither causes immediate distress nor lowers
self-esteem: A meta-analytic review of 192 studies on social exclusion.
Personality and Social Psychology Review, 13, 269–308. http://dx.doi.org/
10.1177/1088868310368523.
Bolger, N., Davis, A., & Rafaeli, E. (2003). Diary methods: Capturing life as it is lived.
Annual Review of Psychology, 54, 579–616. http://dx.doi.org/10.1146/
annurev.psych.54.101601.145030.
Byrne, S., Gay, G., Pollak, J. P., Retelny, D., Gonzales, A. L., Lee, T., et al. (2012). Caring
for mobile phone based avatars can influence youth eating behavior. Journal of
Children
and
Media,
6,
83–89.
http://dx.doi.org/10.1080/
17482798.2011.633410.
Christensen, T. H. (2009). ‘Connected presence’ in distributed family life. New Media
& Society, 11, 433–451. http://dx.doi.org/10.1177/1461444808101620.
Cohen, S. (1988). Psychosocial models of the role of social support in the etiology of
physical disease. Health Psychology, 7, 269–297. http://dx.doi.org/10.1177/
0265407591081005.
Denissen, J. J. A., Penke, L., Schmitt, D. P., & van Aken, M. A. G. (2008). Self-esteem
reactions to social interactions: Evidence for sociometer mechanisms across
days, people, and nations. Journal of Personality and Social Psychology, 95,
181–196. http://dx.doi.org/10.1037/0022-3514.95.1.181.
A.L. Gonzales / Computers in Human Behavior 39 (2014) 197–203
Durkheim, E. (1897). Suicide. New York: Simon & Schuster.
Emmers-Sommer, T. M. (2004). The effect of communication quality and quantity
indicators on intimacy and relational satisfaction. Journal of Social & Personal
Relationships, 21, 399–411. http://dx.doi.org/10.1177/0265407504042839.
Epley, N., & Kruger, J. (2005). When what you type isn’t what they read: The
perseverance of stereotypes and expectancies over e-mail. Journal of
Experimental Social Psychology, 41, 414–422.
Ford, M. B., & Collins, N. L. (2010). Self-esteem moderates neuroendocrine and
psychological responses to interpersonal rejection. Journal of Personality and
Social Psychology, 98, 405–419. http://dx.doi.org/10.1037/a0017345.
Gonzales, A. L., & Hancock, J. T. (2011). Mirror, mirror on my Facebook wall: Effects
of Facebook exposure on self-esteem. CyberPsychology, Behavior & Social
Networking, 14, 79–83. http://dx.doi.org/10.1089/cyber.2009.0411.
Gross, E. F. (2009). Logging on, bouncing back: An experimental investigation of
online communication following social exclusion. Developmental Psychology, 45,
1787–1793. http://dx.doi.org/10.1037/a0016541.
Hampton, K. N., Sessions-Goulet, L. F., Rainie, L., & Purcell, K. (2011). Social
networking sites and our lives. Washington, DC: Pew Internet & American Life
Project.
House, J. S., Landis, K. R., & Umberson, D. (1988). Social relationships and health.
Science, 241, 540–545. http://dx.doi.org/10.1126/science.3399889.
International Telecommunication Union (ITU) 2011, ICT Data and Statistics (nd)
Global ICT Trends Retrieved November 27, 2012, from <http://www.ituint/ITUD/ict/statistics/>.
Jiang, C. L., Bazarova, N. N., & Hancock, J. T. (2011). From perception to behavior:
Disclosure reciprocity and the intensification of intimacy in computer-mediated
communication. Communication Research, 37, 58–77. http://dx.doi.org/10.1177/
0093650211405313.
Joinson, A. N. (2004). Self-esteem, interpersonal risk, and preference for e-mail to
face-to-face communication. CyberPsychology & Behavior, 7, 472–478. http://
dx.doi.org/10.1089/cpb.2004.7.472.
Kahneman, D. (1999). Objective happiness. In D. Kahneman, E. Diener, & N. Schwarz
(Eds.), Well-being: The foundations of hedonic psychology (pp. 3–25). New York:
Russel Sage.
Kalpidou, M., Costin, D., & Morris, J. (2011). The relationship between Facebook and
the well-being of undergraduate college students. CyberPsychology, Behavior,
and Social Networking, 14, 183–189. http://dx.doi.org/10.1089/cyber.2010.0061.
Leary, M. R., Tambor, E. S., Terdal, S. K., & Downs, D. L. (1995). Self-esteem as an
interpersonal monitor: The sociometer hypothesis. Journal of Personality and
Social Psychology, 68, 518–530. http://dx.doi.org/10.1037/0022-3514.68.3.518.
Martens, A., Greenberg, J., Allen, J. J. B., Hayes, J., Schimel, J., & Johns, M. (2010). Selfesteem and autonomic physiology: Self-esteem levels predict cardiac vagal
tone. Journal of Research in Personality, 44, 573–584. http://dx.doi.org/10.1016/
j.jrp.2010.07.001.
Niemz, K., Griffiths, M., & Banyard, P. (2005). Prevalence of pathological internet use
among university students and correlations with self-esteem, the general
health questionnaire (GHQ), and disinhibition. CyberPsychology & Behavior, 8,
562–570. http://dx.doi.org/10.1089/cpb.2005.8.562.
Rains, S. A., & Young, V. (2009). A meta-analysis of research on formal computermediated support groups: Examining group characteristics and health
outcomes. Human Communication Research, 35, 309–336. http://dx.doi.org/
10.1111/j.1468-2958.2009.01353.x.
Reis, H. T., Sheldon, K. M., Gable, S. L., Roscoe, J., & Ryan, R. M. (2000). Daily wellbeing: The role of autonomy, competence, and relatedness. Personality and
203
Social
Psychology
Bulletin,
26,
419–435.
http://dx.doi.org/10.1177/
0146167200266002.
Reis, H. T., Wheeler, L., Kernis, M. H., Spiegel, N., & Nezlek, J. B. (1985). On specificity
in the impact of social participation on physical and psychological health.
Journal of Personality and Social Psychology, 48, 456–471. http://dx.doi.org/
10.1037/0022-3514.48.2.456.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton
University Press.
Shaw, L. H., & Gant, L. M. (2002). In defense of the Internet: The relationship
between internet communication and depression, loneliness, self-esteem, and
perceived social support. CyberPsychology & Behavior, 5, 157–171. http://
dx.doi.org/10.1089/109493102753770552.
Shiffman, S., Stone, A. A., & Hufford, M. R. (2008). Ecological momentary assessment.
Annual Review of Clinical Psychology, 4, 1–32. http://dx.doi.org/10.1146/
annurev.clinpsy.3.022806.091415.
Stone, A. A., & Shiffman, S. (2002). Capturing momentary, self-report data: A
proposal for reporting guidelines. Annals of Behavioral Medicine, 24, 236–243.
Tidwell, L. C., & Walther, J. B. (2002). Computer-mediated communication effects on
disclosure, impressions, and interpersonal evaluations. Human Communication
Research, 28, 317–348. http://dx.doi.org/10.1111/j.1468-2958.2002.tb00811.x.
Trzesniewski, K. H., Donnellan, M. B., Moffitt, T. E., Robins, R. W., Poulton, R., & Caspi,
A. (2006). Low self-esteem during adolescence predicts poor health, criminal
behavior, and limited economic prospects during adulthood. Developmental
Psychology, 42, 381–390. http://dx.doi.org/10.1037/0012-1649.42.2.381.
Turkle, S. (2011). Alone together: Why we expect more from technology and less from
each other. New York, NY: Basic Books.
Uchino, B. N., Cacioppo, J. T., & Kiecolt-Glaser, J. K. (1996). The relationship between
social support and physiological processes: A review with emphasis on
underlying mechanisms and implications for health. Journal of Personality and
Social
Psychology,
119,
488–531.
http://dx.doi.org/10.1037/00332909.119.3.488.
Valkenburg, P. M., & Peter, J. (2009a). Social consequences of the Internet for
adolescents: Decade of research. Current Directions in Psychological Science, 18,
1–5. http://dx.doi.org/10.1111/j.1467-8721.2009.01595.x.
Valkenburg, P. M., & Peter, J. (2009b). The effects of instant messaging on the quality
of adolescents’ existing friendships: A longitudinal study. Journal of
Communication,
59,
79–97.
http://dx.doi.org/10.1111/j.14602466.2008.01405.x.
Valkenburg, P. M., Peter, J., & Schouten, A. P. (2006). Friend networking sites and
their relationship to adolescents’ well-being and social self-esteem.
Cyberpsychology & Behavior, 9, 584–590. http://dx.doi.org/10.1089/
cpb.2006.9.584.
Walther, J. B. (1996). Computer-mediated communication: Impersonal,
interpersonal, and hyperpersonal interaction. Communication Research, 23,
3–44. http://dx.doi.org/10.1177/009365096023001001.
Walther, J. B. (2012). Affordances, effects, and technology errors. In C. T. Salmon
(Ed.), Communication yearbook 36 (pp. 190–193). New York, NY: Routledge.
Wei, R., & Lo, V.-H. (2006). Staying connected while on the move: Cell phone use
and social connectedness. New Media & Society, 8, 53–72. http://dx.doi.org/
10.1177/1461444806059870.
Wills, T. A. (1985). Supportive functions of interpersonal relationships. In S. Cohen &
L. S. Syme (Eds.), Social support and health (pp. 61–82). San Diego, CA, US:
Academic Press.