The cycle of concealment model

JSPR
Article
The cycle of
concealment model
Journal of Social and
Personal Relationships
000(00) 1–16
ª The Author(s) 2010
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0265407510378301
spr.sagepub.com
Tamara D. Afifi1 and Kellie Steuber2
Abstract
This study tested a cycle of concealment model in families. Confirming the model, a
family member’s past verbally aggressive reactions to revelations influenced the extent
to which the family member was estimated to react negatively to revealing a secret.
These factors also attenuated closeness toward family members. In addition, expectations of negative reactions to a revelation and the extent to which those expectations
were violated after the secret was revealed influenced closeness with the family member
and future willingness to reveal secrets. Counter to the model, however, secret disclosure did not affect closeness. Feelings of closeness toward the family member also did
not predict revelation of the secret.
Keywords
avoidance, closeness, concealment, family, privacy, satisfaction, secrets
Families in which members feel the need to bury information out of shame or fear of
judgment probably communicate in vastly different ways compared with families where
those fears are minimized (Vangelisti & Caughlin, 1997). For example, families with
rigid conformity rules that foster high expectations for appropriate behavior (Fitzpatrick
& Ritchie, 1994; Koerner & Fitzpatrick, 1997, 2002) can create a pressure to conceal
secrets (Afifi & Olson, 2005). Unfortunately, continually feeling the need to guard
disclosures, because of fear of family members’ negative reactions, may stifle important
identity components. These fears can also negatively affect relationships (Afifi & Olson,
1
2
University of California, Santa Barbara, USA
University of Iowa, USA
Corresponding author:
Tamara D. Afifi, Department of Communication, 4119 Social Sciences and Media Studies Building, University of
California, Santa Barbara, Santa Barbara, CA 93106-4020
Email: [email protected]
1
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
2
Journal of Social and Personal Relationships 000(00)
2005), especially if individuals become entrenched in predictable patterns of concealment
with family members.
Although research demonstrates that secret keeping (e.g., Vangelisti & Caughlin,
1997) and topic avoidance (e.g., Afifi, Coho, McManus, & Steuber, 2009; Caughlin &
Golish, 2002; Dailey & Palomares, 2004) are negatively related to relationship closeness
and satisfaction, this research is primarily cross-sectional in nature. Little research has
examined how concealing secrets influences relationships over time, particularly if the
concealment is motivated by fear of family members’ responses to revelations. Chilling
effect research (Cloven & Roloff, 1993; Roloff & Cloven, 1990) focuses on when people
refrain from voicing opinions, complaints or concerns for fear of how others will react.
Applying the chilling effect to secrets, Afifi and Olson (2005) found that people often
refrain from revealing secrets when they fear aggressive responses. An important next
step, then, is to examine how withholding (and revealing) information is associated with
changes in relational quality.
Understanding patterns of concealment in families and their impact on these relationships is essential because if individuals consistently fear negative reactions to the
revelation from family members, it may create a ‘‘cycle of concealment’’ (Afifi & Olson,
2005). Secrets may be concealed because an individual has been reprimanded for previously revealing sensitive information. This fear of retribution, due to past negative,
disconfirming, and/or verbally aggressive responses, may propel individuals to suppress
information. This suppression might widen feelings of disconnect between the secret
keeper and some family members. These outcomes are likely dependent upon the degree
to which the potential recipient of the secret violated their expectations for certain types
of responses. While research on secret keeping generally focuses on people speculating
what they think would happen if they revealed a secret (e.g., Afifi, Olson, & Armstrong,
2005; Vangelisti, Caughlin, & Timmerman, 2001), we assessed what actually happens to
closeness and willingness to reveal future secrets with a family member after actually
revealing a secret. Therefore, the goal of this study is to test the extent to which a cycle of
concealment exists in families.
Components of the cycle of concealment model
Expectations of negative responses to revelations
According to Communication Privacy Management (CPM) theory (Petronio, 1991,
2000, 2002), revealing secrets or private information leaves people vulnerable, leading
them to erect metaphorical boundaries around themselves that ebb and flow with the
degree of risk associated with disclosing information. The more comfortable people feel
disclosing, the more permeable their privacy boundaries become and the more likely
they are to reveal sensitive information. Privacy boundaries become more impermeable
when negative reactions to disclosures are expected. Boundaries are likely more
permeable when others are trusted or when the confidant is anticipated to be accepting,
supportive, or open-minded (Harber & Pennebaker, 1992; Pennebaker, 1990; Vangelisti
& Caughlin, 1997; Vangelisti et al., 2001). When responses to revelations are consistently disconfirming, privacy boundaries may become increasingly rigid or impermeable
2
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
Afifi and Steuber
3
over time, perpetuating a cycle of concealment (Afifi & Olson, 2005). Past negative,
disconfirming, or verbally aggressive responses may encourage consistent secret keeping from specific family members. Although originally proposed by Afifi and Olson, this
study is an initial test of the cycle of concealment.
Although the chilling effect (Cloven & Roloff, 1993; Roloff & Cloven, 1990) has
been applied to the suppression of complaints in romantic relationships, these power
struggles are also relevant to secret keeping in families (Afifi & Olson, 2005; Afifi
et al., 2005). The chilling effect can stem from different types of power and different
relationship forms (Afifi & Olson, 2005). Dependence power, or the control a less
reliant partner has in the relationship over the more dependent partner (Lawler &
Bacharach, 1987; Solomon & Samp, 1998), is often considered part of the chilling
effect (Cloven & Roloff, 1993). If parents disapprove of their children, they have the
dependence power to restrict a variety of outcomes (e.g., activities, emotional connections, or money). Secrets may be kept from a family member because that person
has dependence power over them. Another power source, punitive power, stems from a
fear of physical or psychological aggression. The chilling effect is typically viewed as
punitive power or the suppression of ideas as a result of fear of verbal and/or physically
aggressive responses (see Cloven & Roloff, 1993; Roloff & Cloven, 1990; Solomon &
Samp, 1998).
The chilling effect can also stem from the anticipation of an aggressive response.
Although no actual aggression has to occur, this anticipation is most likely a result of
past communicative responses. By keeping topics secret, individuals attempt to control
the risk associated with potential responses. Aggression is only one type of response to
the revelation of secrets (see Afifi, Caughlin, & Afifi, 2007). Anticipated aggressive
responses are more likely than other responses (e.g., disappointment, sadness) to elicit
concealment because of the severe consequences of revealing the secret. In the current
study, the chilling effect is operationalized as the suppression of information
(i.e., secrets) generated by past verbally aggressive responses to revelations and/or
expectation of future negative responses to revelations.
Family members’ past negative reactions, whether they are generally negative or
aggressive, to secret revelations establish expectations for future negative reactions to
revelations. People keep sensitive information private to protect themselves and other
family members from getting hurt, ridiculed, or shamed (Afifi et al., 2005; Petronio,
2002) as well as to preserve their relationship. In fact, Afifi et al. (2005) found that the
desire for self and other protection mediated the association between family members’
past aggressive response to negative secret revelation and decisions to continue keeping
secrets.
Habitual secret keeping may result from patterns of disconfirming family messages.
Parental confirmation generally plays a substantial role in children’s identity and selfworth (e.g. Ellis, 2002). Disconfirming environments often include more and stronger
hostile messages than validating environments (Dailey, 2006). Disconfirming environments – including parental responses such as indifference, imperviousness, or disqualification – interfere with childhood development, resulting in children distrusting their own
abilities (Dailey, 2006). Confirming climates (i.e., recognizing, acknowledging, and endorsing responses) help children work through their emotions (Gottman, Katz, & Hooven,
3
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
4
Journal of Social and Personal Relationships 000(00)
Verbally
aggressive
responses to
prior revelations
H1
H3a
Expectation for
negative reaction to
potential revelation
H3b
Violation of
expectations of
actual reaction
H2
H4
H3c
H6
H3f
Closeness Time 1
H3d
H5
Change in closeness
Time 2
Willingness to
reveal secrets
in the future
H3e
Revealment or
continued
concealment of
secret
Figure 1. Cycle of Concealment Model
1996) and enhance communicative and argumentation skills (Dailey, 2006). Along
similar lines, person-centered messages (i.e., empathy-oriented and rationale-driven)
from parents promote more positive communicative environments for children than
unidirectional, dogmatic messages (e.g., Applegate, Burleson, & Delia, 1992). Thus,
when people feel that their ideas and opinions are confirmed rather than rejected, they
are more likely to reveal sensitive information to family members. In general, then,
our cycle of concealment model (see Figure 1) predicts that past verbally aggressive
responses to secret revelations should predict expectations for negative reactions to
other revelations (H1). These expectations for negative reactions to revelations should
be negatively related to actual secret revelation (H2).
Closeness, expectations for negative responses, and concealment
Afifi and Olson (2005) examined closeness as a mediator of the association between
family members’ past aggression to secret revelation and continued concealment. They
hypothesized that aggression weakens bonds between family members and, as a result,
reduces the likelihood of confiding in those family members. They found, however, a
direct relationship between aggression and continued concealment, and that aggression
to past revelations was associated with less closeness between family members. Thus,
verbally aggressive responses to prior revelations of secrets should directly affect family
members’ closeness (H3a). Similarly, the perceptions of potential adverse reactions to
the revelation of secrets should be associated with less closeness before the revelation
(H3b) and even less closeness afterward (H3c).
Although Afifi and Olson (2005) is an exception, closeness is frequently a positive
predictor of the disclosure of secrets (e.g., Vangelisti et al., 2001). In CPM terms, as
closeness increases, the more permeable privacy boundaries become because the
disclosure risk is minimized (Petronio, 2000). Therefore, it is hypothesized that, as
closeness increases, the disclosure of secrets should increase (H3d).
4
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
Afifi and Steuber
5
Revealing secrets may increase relational closeness. While research typically finds
that withholding information is associated with less close and satisfied relationships
(e.g., Afifi et al., 2009; Caughlin, Huston, & Houts, 2000; Vangelisti & Caughlin, 1997),
research has not examined relational consequences of the revelation of secrets.
Revealing a secret could damage family relationships if it is not received well (Afifi &
Olson, 2005) or could bring members closer if bonding and catharsis ensues (Vangelisti
& Caughlin, 1997). Although disclosing secrets improves personal health (e.g.,
Pennebaker, Kiecolt-Glaser, & Glaser, 1988; Pennebaker, 1990), less research has
focused on how it affects relational health over time. One notable exception is Frijns and
Finkenauer’s (2009) large longitudinal study of adolescents. Adolescents who revealed a
secret to no one reported diminished well-being (e.g., depressed mood, lower
self-esteem, loneliness) and poorer parental relationships six months later, when
compared with those who disclosed their secret to someone. The perception that a
partner is keeping a secret may inadvertently cause relationship harm. Finkenauer,
Frijns, Engels, and Kerkhoff (2005) found that parents’ perception that their adolescent
was concealing a secret was associated with poorer parenting practices, regardless of
whether there was actual secrecy or not. Although the direction of the relationship is
unclear, the revelation of a secret should influence family relationships (H3e). Increases
in closeness after a revelation should predict greater willingness to subsequently reveal
secrets (H3f). Likewise, expectations of a negative reaction to the disclosure of a secret
should predict the likelihood of revealing future secrets (H4).
Expectancy violations
The link between family bonds and future revelation decisions is also likely based upon
the extent to which the target met or violated expectations to past revelations. From a
CPM (Petronio, 1991, 2002) perspective, when faced with consistently disconfirming
responses to revelations, privacy boundaries may become increasingly impermeable or
rigid over time. This could potentially widen the emotional disconnect between family
members. These rigid boundaries and weakened relationships likely will continue, unless
the receiver positively violates expectations for responses to secrets. When positive
expectancy violations for reactions occur, it may slowly build trust, closeness, and
willingness to disclose future secrets. People reveal secrets, however, for a host of
reasons (e.g., receiver has the right to the information, internal or external pressure,
catharsis), despite the anticipated negative responses (Afifi & Steuber, 2009). There is a
tendency, however, to conceal secrets when aggressive responses are anticipated (Afifi
et al., 2007; Kelly & McKillop, 1996; Petronio, 2000).
The ramifications of revealing secrets are real, but potential negative reactions may be
exaggerated through rumination (see Caughlin, Afifi, Carpenter-Theune, & Miller,
2005; Dunn & Laham, 2006; Lane & Wegner, 1995; Roloff & Ifert, 2000). Reactions to
secret revelations tend to be more positive than expected, which can enhance self-esteem
(Afifi & Caughlin, 2006). Thus, expectations for negative reactions, and whether those
expectations are met or violated, likely influence closeness and willingness to reveal
future secrets. Even if negative reactions to secret revelations occurred in the past, a
more positive reaction than expected may improve the relationship and increase the
5
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
6
Journal of Social and Personal Relationships 000(00)
likelihood of future revelations. The opposite is also true – consistently negative reactions to revelations, or worse reactions than expected, likely diminish closeness and foster continued concealment. Our cycle of concealment model (see Figure 1) hypothesizes
that responses to secret revelations that positively violate expectations should be positively associated with closeness (H5) and the willingness to reveal future secrets (H6).
Method
Participants
Participants included 594 female (n ¼ 383, 65%) and male (n ¼ 207, 35%) college
students (age range from 17 to 28, M ¼ 19.65), who were predominantly sophomores
(n ¼ 268, 45%) and juniors (n ¼ 163, 27%). Most participants were White (n ¼ 452,
76%), Hispanic (n ¼ 21, 4%) or Asian (n ¼ 58, 10%).
Procedures
Participants were recruited from communication courses at a large Western and a large
Eastern University in the United States and were offered nominal extra credit for their
participation. They completed two surveys, two months apart. The first survey gathered
information about a secret the participants were keeping from another family member,
their relationship with that family member, and their expectations of how that person
would react if s/he revealed the secret. The second survey determined whether participants revealed the secret to the target and how it affected their relationship. Surveys were
completed in a laboratory on campus. To reduce attrition, all participants received email
reminders, resulting in a 77% compliance rate. Only data from students who completed
both surveys are reported.
Of the 594 participants who completed both surveys, 122 (21%) revealed their secret
within the two-month period. In the second survey, participants who indicated that the
target person learned the secret since the first survey were asked how the target learned
the secret. We needed to differentiate voluntary revelation and the target accidently discovering the secret through another means (Caughlin et al., 2005). As Afifi and Steuber
(2009) reported, however, people often use indirect strategies for disclosing secrets (e.g.,
telling a third party whom they know would tell the target or providing evidence of the
secret). To retain participants who used indirect disclosure strategies, we eliminated
those who reported not using any of these strategies to reveal the secret. Five participants
had no intent of revealing the secret to the target and three participants who reported
keeping no secrets from a family member were removed from the final sample. These
deletions, along with eliminating 66 participants with positive secrets, resulted in a final
sample size of 520 participants, 100 of whom had their secret revealed. Of these participants, 300 (58%) indicated they were primarily keeping the secret from their mother,
137 (27%) from their father, and 54 (18%) from a sibling.
Participants were initially informed that secrets could be positive or negative and that
they were a conscious choice to withhold information from a particular person. They
were asked to think of someone in their immediate family from whom they were keeping
a secret and to consider this person throughout the study. In the second data collection,
6
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
Afifi and Steuber
7
participants were reminded of their secret and the name of the family member from
whom they were keeping the secret. All participants who initially had a secret claimed to
remember it. The secrets in the current manuscript are consistent with those reported in
earlier studies (Vangelisti, 1994; Vangelisti, Caughlin, & Timmerman, 2000).
Measures
Past aggressive responses. The target family member’s past aggression in response to
secret revelation was measured at time 1 with modified verbal aggression items from the
Conflict Tactics Scales (CTS) (Mason & Blankenship, 1987; Straus, 1990). Participants
reported whether the target family member responded with various verbally aggressive
behaviors in response to past secret revelations. Seven items measured symbolically
aggressive responses (e.g., verbally or psychologically abusive). These items were
augmented with eight additional items that measured the psychological nature of
aggression (e.g., ‘‘gave me a disappointing look’’). Participants indicated whether the
target performed each behavior in response to prior secret revelations. Summed scores
ranged from zero (no aggression) to 15 (high aggression).
Expected negative reaction. At time 1, the target’s expected negative reaction to secret
revelation was measured with six items, each accompanied by a seven point, Likert-type
scale (1 ¼ ‘‘very unlikely’’, 7 ¼ ‘‘very likely’’). Items tapped disappointment, judgment,
and withdrawal. One item assessed the overall negativity of the target’s reaction. Higher
scores reflect increasingly negative reactions. Confirmatory factor analysis revealed
good fit and unidimensional structure of the items (w2 [5, N ¼ 493] ¼ 16.45, p ¼ .06,
NFI ¼ .99, CFI ¼ .99, RMSEA ¼ .07, where NFI is the normal fit index, CFI is the
comparative fit index and RMSEA is the root mean square error of approximation). The
coefficient alpha for this scale was .80.
Expectation violation. At time 2, the extent to which the target’s reaction violated
expectations was measured with two, 7-point semantic differential items (Caughlin et al.,
2005) (e.g., ‘‘this person reacted much more negatively (to much more positively) than
I expected’’). Higher scores reflect more positive violations. Given a strong correlation
between items (r ¼ .75), the items were averaged.
Closeness. Participants’ closeness with the target family member was measured with eight
of the original 10 items from Buchanan, Maccoby, and Dornbusch’s (1991) closeness
scale at time 1 and time 2. Two items were removed as they focused on openness and
were strongly correlated with revealment. The five-point Likert-type scale was coded
such that larger scores reflect closer relationships. The coefficient alphas were .89 at time
1 and .86 at time 2.
Secret valence. Only participants reporting neutral or negative secrets were analyzed
because responses are unlikely to be negative or aggressive to positive secrets. Two
items (Vangelisti & Caughlin, 1997) measured secret valence. Seven-point semantic
differential items (extremely good - extremely bad and extremely positive - extremely
7
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
8
Journal of Social and Personal Relationships 000(00)
Table 1. Correlations, means, and standard deviations for the variables in the cycle of concealment
model
Variables
1.
2.
3.
4.
5.
Past verbal aggression (T1)
Expected neg. reaction (T1)
T1 closeness
T2 closeness
Violation of expectations
(T2)
6. Willingness to reveal in
future (T2)
M
SD
1
2
3
4
5
6.51
3.42
4.18
4.12
1.20
4.14
1.24
.62
.67
5.22
——
.51***
–.48***
–.43***
.10
——
–.30**
–.35***
–.07
——
.60***
.22*
——
.20*
——
5.27
1.68
–.14
–.30*
.11
.23*
.21*
6
——
Note. * p < .05; ** p < .01; *** p < .001.
negative) were strongly correlated (r ¼ .70). Participants were retained if they scored at
or above 4.0 on negativity on this scale (M ¼ 4.71; SD ¼ 1.41).
Secret revelation. Whether participants revealed the secret to the target in the period
between times 1 and 2 was dummy coded (did or did not reveal it).
Willingness to reveal future secrets. Participants’ willingness to reveal future secrets to the
target was measured with one item at time 2 (Vangelisti & Caughlin, 1997; Vangelisti et.
al., 2001). Participants who revealed their secret responded to a seven-point Likert-type
scale (1 ¼ ‘‘not at all likely,’’ 7 ¼ ‘‘very likely’’).
Results
Preliminary data analyses
Means, standard deviations, and correlations among all cycle of concealment model
variables are provided in Table 1. Target family members reacted slightly more positively to the secret’s revelation than participants expected and participants, on average,
reported relatively and consistently high closeness with the target.
Structural equation modeling (using AMOS 18.0) with maximum likelihood estimation was used to test the cycle of concealment model (Figure 1) and account for
missing data. This version of the software allows for estimation of categorical outcome
variables using Bayesian estimation. The cycle of concealment model was tested with
two structural equation models. The first model used all participants to test H1, H2, and
H3a-e (see Figure 2). The second model tested all of the hypotheses, except those involving the actual revelation and included data only from participants who revealed their
secret (see Figure 3). One path was added to the second model, simply to account for the
fact that reactions anticipated before the revelation should be associated with the degree
to which those expectations were violated after the revelation.
All confirmatory factor analyses (CFAs) indicated good unidimensional fit. After the
CFAs were calculated, preliminary measurement models were formed by parceling each
of the latent constructs. The latent constructs for expectation for negative responses and
8
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
Afifi and Steuber
9
.37
Expectation of
neg reaction
.61***
Past
aggression
-.03
-.18**
-.26***
.13**
.16
.42
.65***
Closeness T1
Closeness T2
.00
.02
.02
Revealment or
continued
concealment
Figure 2. Hypothesized Model for Full Sample for Revelation of Secret and Association with
Closeness
Note. The number to the upper right of the latent construct is the multiple R2. All of the parameters
in the model are standardized.** ¼ p < .01; *** ¼ p < .001.
.33
Expectation
of negative
reaction
.57***
Violation of
expectation
Past verbal
aggression
-.09
-.21*
-.31*
.22*
.27*
-.47***
.28
-.20*
Closeness T1
.17
.61
.05
.63***
Closeness T2
Willingness
to reveal
Figure 3. Hypothesized Model with Those Who Revealed Their Secret
Note. The number to the upper right of the latent construct is the multiple R2. All of the parameters
in the model are standardized.* ¼ p < .05; ** ¼ p < .01; *** ¼ p < .001.
closeness (separately for time 1 and time 2) were parceled into three parcels each. Items
were formed into the parcels by averaging similarly operationalized items together (see
Matsunaga, 2008). The measurement model for the first model with all of the participants was a good fit to the data (w2 [40, N ¼ 493] ¼ 93.46, p ¼ 09, NFI ¼ .94; CFI ¼
.95, RMSEA ¼ .08), as was the second revealment model (w2 [48, N ¼ 100] ¼ 76.18,
9
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
10
Journal of Social and Personal Relationships 000(00)
p ¼ .06, NFI ¼ .97, CFI ¼ .98, RMSEA ¼ .07). However, to further improve the fit of
the model because of the limited sample size for model two, we formed single
composite variables or indicators (with individual items, subscales or scales) for the
latent variables instead of parceling the items for the structural models. To control for
measurement error, the error variance for observed indices of latent constructs was
fixed to (1-a) multiplied by the variance of the indicator (see Bollen, 1989; Stephenson
& Holbert, 2003).
Testing the cycle of concealment model
The first model (see Figure 2) tested H1, H2, and H3a-e and was an excellent fit to the data,
w2 (2, N ¼ 493) ¼ 4.99, p ¼ .09, NFI ¼ .99, CFI ¼ .99, RMSEA ¼ .06. Past verbal aggression was strongly and positively associated with expectations that the target would respond
negatively upon revelation (H1). Both verbal aggression to prior secrets (H3a) and the
expectation for aggression to the revelation (H3b) were significantly associated with less
closeness at time 1. The expectation of a negative response was significantly and positively
associated with the continued secret concealment (H2). The path between past aggressive
reactions and continued concealment was also tested; however, this association only
approached significance (.12, p < .06) and was not included in the model. Inconsistent with
hypotheses, participants’ closeness with the target was not associated with secret
revelation (H3d). Moreover, the secret revelation did not significantly affect participants’
closeness with their target at time 2 (H3e).
The second model (see Figure 3) was tested using data only from participants who
revealed their secret. The model was an excellent fit to the data w2 (5, N ¼ 100) ¼ 3.00,
p ¼ .85, NFI ¼ .99, CFI ¼ .99, RMSEA ¼ .01. Past verbal aggression was strongly and
positively related to the expectation that the target would respond negatively upon secret
revelation (H1) as well as strongly and negatively related to closeness at time 1 (H3a).
The expectation for negative reactions was not related to time 1 closeness (H3b).
Closeness at time 1 was strongly related to closeness at time 2. Time 2 closeness,
however, was not significantly related to willingness to reveal future secrets to the
target (H3f). The more positively targets violated secret-keepers’ expectations,
the greater time 2 closeness (H5) and the more willing secret-keepers were to reveal
future secrets (H6).
Discussion
Most research on secrets has examined factors likely to propel people to reveal or
conceal secrets (e.g., Derlega, Winstead, & Folk-Barron, 2000; Petronio, Reeder,
Hecht, Ros-Medoza, 1996; Vangelisti et al., 2001) and indicates that secret keeping
tends to be associated with relational dissatisfaction (e.g., Caughlin et al., 2000;
Vangelisti & Caughlin, 1997). It is less clear is how secrecy patterns are created and
maintained in families and the relational costs associated with these patterns
over time. The current study makes important contributions to this literature by
examining how secrets and their revelations change relational closeness with a
family member.
10
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
Afifi and Steuber
11
The cycle of concealment in families
These data provide initial support for a ‘‘cycle of concealment’’ in families. As hypothesized, a target family member’s verbally aggressive reactions to prior revelations
influenced individuals’ estimations of that target family member’s negative reaction to
present revelations. When family members were expected to react negatively to revelations, continued concealment resulted. In addition, both past aggression and the
expectation of negative reactions were related to decreases in closeness with the target.
Consistent concealment of secrets out of fear of how the other will respond could
slowly tear at family relationships (Afifi & Olson, 2005). However, unlike what prior
research suggests (e.g., Vangelisti et al., 2001), revealing a secret to a target was not
related to closeness. Participants’ expectations for the target person’s reaction to
revelations and the extent to which that perception was violated seemed to make the
biggest difference in decisions to reveal. Fear of how others will react to revelations
will likely lead to continued concealment (Kelly & McKillop, 1996), perhaps
regardless of closeness with that person.
Willingness to reveal sensitive information in families (with their long communicative history) is likely based on disclosure and response patterns that emerge over
time. Perception of a target’s reaction to secret revelation depends both on the teller’s
expectations for the response and whether the target’s action response meets or violates
that expectation. Consistent with Caughlin et al. (2005), targets responded more positively to secret revelations than was expected, perhaps because the expected negative
reaction is exacerbated in the secret-keeper’s mind. The secret’s emotional burden may
be greater on the secret keeper than the target or relationship. This also suggests that,
when expectations are positively violated (i.e., family members respond to the revelation
more favorably than was thought), future revelations to that target are more likely.
Continued concealment is likely to follow consistently negative responses to
revelations from family members. Over time, these responses to revelations may
influence family bonds and, perhaps, the teller’s attachments with the respondent. If
family members positively violate expectations in their responses to revelations, it may
create confidence and a sense that future revelations can occur, perhaps also fostering
more secure attachments in the teller. Continued disconfirmation of revelations may
produce insecure attachments in general and inconsistent responses to family secret
revelations may produce a fearful avoidant attachment, specifically (see Bartholomew
& Horowitz, 1991), where family members fear revealing information because they
cannot predict the response.
An interesting practical, and theoretical, avenue of future research would be to
consider how to break family cycles of concealment. Are there trigger events in families
that change disclosure and concealment norms? For example, significant life events
(e.g., marriage, divorce, children, or illness) may create openness and acceptance. A new
spouse, for example, could help a partner communicate more directly with family
members. These same events, however, could initiate conflict and establish taboo topics
in families. Personality traits or communication characteristics may also facilitate or
inhibit cycles of concealment. For example, communication efficacy may ease the
communication of secrets (Afifi et al., 2005). Similarly, the sociometer hypothesis
11
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
12
Journal of Social and Personal Relationships 000(00)
(Leary, 1999; Leary, Haupt, Strausser, & Chokel, 1998; Leary, Tambor, Terdal, &
Downs, 1995) suggests that people want to feel included or accepted and will avoid
situations that prompt feelings of exclusion. High self-esteem, then, may make it easier
to handle negative family reactions to revelations and enable future revelation of negative information. Individuals with low self-esteem tend to closely monitor people’s reactions for signs of rejection and avoid situations where they could be rejected. Therefore,
low self-esteem may facilitate the censoring of negatively valenced disclosures from
those with whom one expects a negative response. Findings from the present study suggest that one way of breaking a cycle of concealment may be revelation.
Finally, scholars should investigate how cycles of concealment in families influence
identities. More specifically, how do family members’ continued negative reactions to
disclosures, over time, influence communication of one’s identity? People may hide
aspects of their identity from certain family members because revealing it produced
previous negative reactions. For instance, failed attempts at revealing one’s sexual
identity may perpetuate the hiding of that part of one’s identity. The hierarchically
structured nature of identity is a normal part of how people communicate their identities
to others, but if it is done out of fear of others’ negative reactions, over time it might have
a stifling effect on one’s sense of self. As secrets examined were, on average, moderately
important to identities (a mean of 4.89 on a 7-point scale), future work could examine
how identity salience influences concealment patterns. Does identity salience of the
secret and/or history of negative reactions to similar secrets influence likelihood of revelation? Highly salient secrets may be more likely to be concealed because of the negative
consequences associated with revelation. However, some topics are so important to identity (e.g., being gay), that direct revelation, regardless of ramifications, may be a way of
remaining true to the self (Denes & Afifi, 2010).
While this study examined secrets, defined as information purposefully withheld from
a family member, the cycle of concealment likely applies to topic avoidance where the
target knows of the topic, but it is not discussed. In this study, a topic may have been
introduced in the past, but dropped when repeated negative responses resulted in frustration. It is also possible that targets suspect a secret, but individuals conceal it anyway,
transitioning to topic avoidance. These important distinctions require clarification in
future research. Repeated failures to reveal a secret, especially if it is important to one’s
identity, may be particularly harmful to the concealer’s self and relationship because it
has a stifling effect. Future research should assess whether initial responses to secrets or
discussions of topics commonly avoided perpetuate cycles of concealment over time and
the role of identity in this process.
Concluding remarks
This article’s contributions must be considered within its limitations. This study traced
the cycles of concealment over a two month period. Secrets tracked over a longer time
likely would have generated more revelations. In addition, secrets revealed within such a
short period may be less negatively valenced than longer-term secrets. Identity threatening, stigmatized secrets, according to societal standards, are especially likely to be concealed (Vangelisti, 1994). Even given the longitudinal nature of this study, model paths
12
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
Afifi and Steuber
13
may not be causal. For instance, it is possible that the expectation for negative reactions
predicts actual negative reactions for the respondent (i.e., self-fulfilling prophesy).
Moreover, secrets are concealed for many reasons, such as relationship protection and
privacy (see Petronio, 2000; Afifi & Stueber, 2009; Caughlin et al., 2005; Vangelisti
et al., 2001). Finally, these results are limited to secrets in families, specifically, one
family member (typically a parent). To fully test a cycle of concealment model in families, researchers should consider multiple family members and a longer time period.
Processes underlying secret keeping in families likely apply to other relational contexts
(e.g., romantic relationships). The model should probably be restricted to contexts with a
chilling effect, which requires dependence power or power discrepancy. Children, for
example, may be more likely to engage in the cycle of concealment rather than parents
because children typically have no dependence power over parents. Parents also are less
likely to fear an aggressive response from children.
Even with limitations, the study highlights patterns of concealment in families and
how they affect family relationships. Results provide initial support for a cycle of
concealment such that past aggressive responses, expectations for present negative
responses to secret disclosures, and expectancy violations influence concealment and
family relationships. Counter to our expectations, however, revealing or concealing a secret
did not influence closeness. Additional research is necessary to confirm or disconfirm this
part of the model.
Acknowledgements
Duncan Cramer served as Action Editor for this article.
Conflict of interest statement
The author(s) declared no conflicts of interest with respect to the authorship and/or publication
of this article.
Funding
This research received no specific grant from any funding agency in the public, commercial, or
not-for-profit sectors.
References
Afifi, T. D., Caughlin, J., & Afifi, W. A. (2007). Exploring the dark side (and light side) of
avoidance and secrets. B. Spitzberg and B. Cupach (Eds.), The darkside of interpersonal
relationships (2nd ed., pp. 61–92). Mahwah, NJ: Erlbaum.
Afifi, W. A., & Caughlin, J. (2006). A close look at revealing secrets and some consequences that
follow. Communication Research, 33, 467–488.
Afifi, T. D., & Olson, L. N. (2005). The chilling effect in families and the pressure to conceal
secrets. Communication Monographs, 72, 192–216.
Afifi, T. D., Olson, L. N., & Armstrong, C. (2005). The chilling effect and family secrets. Human
Communication Research, 31, 564–598.
Afifi, T. D., McManus, T., Steuber, K., & Coho, A. (2009). Verbal avoidance and dissatisfaction in
intimate conflict situations. Human Communication Research, 35, 357–383.
13
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
14
Journal of Social and Personal Relationships 000(00)
Afifi, T. D., & Steuber, K. (2009). The Risk Revelation Model (RRM) and strategies used to reveal
secrets. Communication Monographs, 76, 144–176.
Applegate, J., Burleson, B., & Delia, J. (1992). Reflection-enhancing parenting as an antecedent
to children’s social-cognitive and communicative development. In Sigel, Irving E.;
McGillicuddy-DeLisi, Ann V.; Goodnow, Jacqueline J (Eds.). Parental belief systems: The
psychological consequences for children (2nd ed.). (pp. 3–39). Hillsdale, NJ, England: Lawrence
Erlbaum Associates, Inc.
Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles among young adults: A test of a
four-category model. Journal of Personality and Social Psychology, 61, 226–244.
Ben-Air, A. (1995). Coming out: A dialectic of intimacy and privacy. Families in Society: The
Journal of Contemporary Human Services, 76, 306–314.
Bollen, K. A. (1989). Structural equations with latent variables. Oxford, England: John Wiley &
Sons.
Buchanan, C. M., Maccoby, E. E., & Dornbusch, S. M. (1991). Caught between parents:
Adolescents’ experience in divorced homes. Child Development, 62, 1008–1029.
Caughlin, J. P., Afifi, W.A., Carpenter-Theune, K. E., & Miller, L. E. (2005). Reasons for, and
consequences of, revealing personal secrets in close relationships: A longitudinal study.
Personal Relationships, 12, 43–49.
Caughlin, J. P, & Golish, T. D. (2002). An analysis of the association between topic avoidance and
dissatisfaction: Comparing perceptual and interpersonal explanations. Communication
Monographs, 69, 275–295.
Caughlin, J. P., Huston, T. L., & Houts, R. M. (2000). How does personality matter in marriage?
An examination of trait anxiety, interpersonal negativity, and marital satisfaction. Journal of
Personality and Social Psychology, 78, 326–336.
Cloven, D. H., & Roloff, M. E. (1993). The chilling effect of aggressive potential on the expression
of complaints in intimate relationships. Communication Monographs,60, 198–219.
Dailey, R. (2006). Confirmation in parent–adolescent relationships and adolescent openness:
Toward extending confirmation theory. Communication Monographs, 73, 434–458.
Dailey, R. M., & Palomares, N. A. (2004). Strategic topic avoidance: An investigation of topic
avoidance frequency, strategies used, and relational correlates. Communication Monographs,
71, 471–496.
Denes, A., & Afifi, T. (2010). The second coming out: LGBTQ individuals’ re-disclosures of their
sexual identity to their parents. Manuscript submitted for publication.
Derlega, V. J., Winstead, B. A., & Folk-Barron, L. (2000). Reasons for and against disclosing HIVseropositive test results to an intimate partner: A functional perspective. In S. Peteronio (Ed.).,
Balancing the secrets of private disclosures (pp. 53–71). Mahwah, NJ: Lawrence Erlbaum.
Dunn, E. W., & Laham, S. A. (2006). Affective forecasting: A user’s guide to emotional time
travel. In J. Forgas (Ed.), Hearts and minds: Affective influences on social cognition and behavior. New York: Psychology Press.
Ellis, K. (2002). Perceived parental confirmation: Development and validation of an instrument.
Southern Communication Journal, 67, 319–334.
Finkenauer, C., Frijns, T., Engels, R., & Kerkof, P. (2005). Perceiving concealment in relationships between parents and adolescents: Links with parental behavior. Personal Relationships,
12, 387–406.
14
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
Afifi and Steuber
15
Fitzpatrick, M. A., & Ritchie, D. L. (1994). Communication schemata within the family: Multiple
perspectives on family interaction. Human Communication Research, 20, 275–301.
Frijns, T., & Finkenour, C. (2009). Longitudinal associations between keeping a secret and
psycho-social adjustment in adolescence. International Journal of Behavioral Development,
33, 145–154.
Gottman, J. M., Katz, L. F., & Hooven, C. (1996). Parental meta-emotion philosophy and the emotional life of families: Theoretical models and preliminary data. Journal of Family Psychology,
10, 243–268.
Harber, K. D., & Pennebaker, J. W. (1992). Overcoming traumatic memories. In S. Petronio (Ed.),
Balancing the secrets of private disclosure (pp. 53–71). Mahwah, NJ: Lawrence Erlbaum.
Kelly, A., & McKillop, K. (2006). Consequences of revealing personal secrets. Pychological Bulletin, 120, 450–165.
Koerner, A. F., & Fitzpatrick, M. A. (1997). You never leave your family in a fight: The impact of
family of origin on conflict-behavior in romantic relationships. Communication Studies, 48, 59–75.
Koerner, A. F., & Fitzpatrick, M. A. (2002). Understanding family communication patterns
and family functioning: The roles of conversation orientation and conformity orientation. In
W. Gudykunst (Ed.), Communication yearbook 26, (pp. 36–68). Beverly Hills, CA: Sage.
Lane, D. J., & Wegner, D. M. (1995). The cognitive consequences of secrecy. Journal of Personality and Social Psychology, 69, 237–253.
Lawler, E.J., & Bacharach, S. B. (1987). Comparison of dependence and punitive forms of power.
Social Forces, 66, 446–462.
Leary, M. R. (1999). Making sense of self-esteem. Current Directions in Psychological Science,
8, 32–35.
Leary, M. R., Haupt, A. L., Strausser, K. S., & Chokel, J. T. (1998). Calibrating the sociometer:
The relationship between interpersonal appraisals and state self-esteem. Journal of Personality
and Social Psychology, 74, 1290–1299.
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.
Mason, A., & Blankenship, V. (1987). Power and affiliation motivation, stress, and abuse in intimate relationships. Journal of Personality and Social Psychology, 52, 203–210.
Matsunaga, M. (2008). Item parceling in structural equation modeling: A primer. Communication
Methods and Measures, 2, 260–293.
Pennebaker, J. W. (1990). Opening up: The healing power of confiding in others. New York:
Morrow.
Pennebaker, J. W., Kiecolt-Glaser, J. K., & Glaser, R. (1988). Disclosure of traumas and immune
function: Health implications for psychotherapy. Journal of Consulting and Clinical Psychology, 56, 239–245.
Petronio, S. (1991). Communication boundary management: A theoretical model of managing disclosure of private information between marital couples. Communication Theory, 1, 311–335.
Petronio, S. (2000). The boundaries of privacy: Praxis of everyday life. In S. Petronio (Ed.),
Balancing the secrets of privacy disclosures (pp. 37–49). Mahwah, NJ: Lawrence Erlbaum.
Petronio, S. (2002). Boundaries of privacy: Dialectics of disclosure. Albany, NY: SUNY Press.
Petronio, S., Reeder, H. M., Hecht, M. L., & Ros-Mendoza, T. M. (1996). Disclosure of sexual
abuse by children and adolescents. Journal of Applied Communication Research, 24, 181–199.
15
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016
16
Journal of Social and Personal Relationships 000(00)
Roloff, M. E., & Cloven, D. H. (1990). The chilling effect in interpersonal relationships: The reluctance to speak one’s mind. In D. D. Cahn (Ed.), Intimates in conflict: A communication perspective (pp. 49–76). Hillsdale, NJ: Lawrence Erlbaum.
Roloff, M. E., & Ifert, D. E. (2000). Conflict management through avoidance: Withholding complaints, suppressing arguments, and declaring topics taboo. In S. Petronio (Ed.), Balancing the
secrets of private disclosures (pp.151–163). Mahwah, NJ: Lawrence Erlbaum.
Solomon, D. H., & Samp, J. A. (1998). Power and power appraisal: Perceptual foundations of the
chilling effect in dating relationships. Journal of Social and Personal Relationships, 15, 191–209.
Stephenson, M. T., & Holbert, R. L. (2003). A Monte Carlo simulation of observable versus latent
variable structural equation modeling techniques. Communication Research, 30, 332–354.
Straus, M. A. (1990). The conflict tactics scale and its critics: An evaluation and new data on
validity and reliability. In M. A. Straus & R. J. Gelles (Eds)., Physical violence in American families: Risk factors and adaptations to violence in 8 145 families (pp. 49–73). New Brunswick, NJ:
Transaction.
Vangelisti, A. L. (1994). Family secrets: Forms, functions and correlates. Journal of Social and
Personal Relationships, 11, 113–135.
Vangelisti, A. L., & Caughlin, J. P. (1997). Revealing family secrets: The influence of topic, function, and relationship. Journal of Social and Personal Relationships, 14, 679–705.
Vangelisti, A. L., Caughlin, J. P., & Timmerman, L. (2001). Criteria for revealing family secrets.
Communication Monographs, 68, 1–17.
16
Downloaded from spr.sagepub.com at PENNSYLVANIA STATE UNIV on September 19, 2016