Seeking interpersonal information over the Internet

Abstract

Online information-seeking strategies are well understood in the context of health information acquisition but less is known about the information management of interpersonal information gathered over the Internet. One way to understand this nascent method for information seeking is to examine the motives that underlie one’s intent to seek online information about others. The present study applies theory of motivated information management (TMIM) to online information seeking. The first goal of the study is to examine whether the processes predictive of offline interpersonal information seeking, articulated in TMIM, apply to mediated contexts. The second goal is to investigate whether predictions of TMIM are consistent across various relationship types. A survey design was employed to collect data from college students and nonstudent adults. The results of the study demonstrate strong support for the application of TMIM to information seeking over the Internet. The findings also indicate support for the model across various relationships except best friendships. Implications to research and theory on interpersonal information management over the Internet are discussed.

Keywords

Information-seeking strategies Internet interpersonal theory of motivated information management

Information-seeking behaviors of one’s surrounding environments have evolutionary roots (Shoemaker, 1996). In primitive times, it was necessary to scan the environment for signs of potential harm; the absence of a threat reduced uncertainty about the unknown. This evolutionary carryover now motivates information seeking on a large bandwidth of topics, including political figures, sporting events, prospective purchases, and community activities, from a wide range of information sources (Aspray & Hayes, 2011). The desire to seek information can also originate from circumstance, such as an equivocal event or referent. Information-seeking strategies can be initiated, for example, by the diagnosis of a medical condition (Brashers, Goldsmith, & Hsieh, 2002) or initial contact with a new acquaintance (Douglas, 1990). In the latter case of interpersonal information seeking, individuals have available myriad avenues through which they are able to obtain information on others (Gibbs, Ellison, & Lai, 2011).

The acquisition of information about others can be facilitated by engaging in several information-seeking behaviors. Historically, the options available for seeking this information were relatively limited. People could gather information on others through in-person observations of the target, an innately passive information-seeking strategy, or they could actively or interactively pursue personal information by communicating directly with the target or asking third parties about the target individual (Berger, 1979). These passive and active strategies relied on personal or third-party observations of the target; few other readily accessible means were available for collecting information. This narrow view on seeking interpersonal information, however, is something of an anachronism in the age of digital technologies. The Internet, through its various platforms, makes it possible to perform asymmetrical or transactive information exchange without considerable effort (Hancock, Toma, & Fenner, 2008; Pratt, Wiseman, Cody, & Wendt, 1999; Stefanone, Hurley, & Yang, 2013).

Traditional information sources, such as off-line friends or mutual acquaintances, remain viable options for seeking information. Nevertheless, individuals can now expand their repertoire of information sources to Weblogs and social network sites, considered nontraditional sources of information, to find information on others (Antheunis, Valkenburg, & Peter, 2010; Westerman, Van Der Heidi, Klein, & Walther, 2008). People commonly consider these nontraditional communication channels tenable alternatives to conventional information seeking (Tokunaga, 2011; Utz & Beukeboom, 2011). Some may even seek information from multiple sources or enlist multiple strategies (i.e., use online information to supplement off-line information) to find a broad range of information about those in whom they are interested (Ramirez, Walther, Burgoon, & Sunnafrank, 2002).

The idea of nontraditional modes of information seeking has received growing empirical attention. Nuanced variations of this information-seeking concept, including partner monitoring (Darvell, Walsh, & White, 2011), partner “stalking” (Lyndon, Bonds-Raacke, & Cratty, 2011), and interpersonal electronic surveillance (Tokunaga, 2011), have begun to appear in the literature. Given that over 2 billion people are now using Internet technologies worldwide (Internet World Stats, 2011), the interest in these alternate forms of interpersonal information seeking comes as no surprise. Yet, relatively little is known about the evaluative process Internet users undergo in their decision to employ interpersonal information-seeking strategies uniquely over the Internet. The present investigation endeavors to understand whether motivations and related cognitions involved with online interpersonal information seeking parallel factors that influence one’s decision to seek information through off-line interpersonal channels. Theory of motivated information management (TMIM; Afifi & Weiner, 2004) is applied to information seeking over communication technologies to examine whether its tenets are supported in the online context. Deviations in predictions of TMIM among different relationship types are also considered in this investigation. The results have theoretical implications for testing boundary conditions of TMIM and introducing to the model meaningful variables indigenous to online information seeking (e.g., Internet self-efficacy).

Theory of motivation information management

TMIM (Afifi & Weiner, 2004) theorizes about events in which individuals assume active strategies in pursuit of discovering desired information on others. TMIM distinguishes itself from other information-seeking theories (e.g., uncertainty reduction theory; Berger & Calabrese, 1975) in that it does not attribute the motivation of information seeking to a desire for uncertainty reduction; rather, the catalyst of information management in TMIM lies in the discrepancy between actual and desired uncertainty. Although this theory is primarily interested in information management through interpersonal channels (Afifi & Weiner, 2004), it may also be appropriately employed to explain decisions regarding information management in communication technology channels (Afifi & Lucas, 2008). Moreover, TMIM has been mainly applied in contexts wherein information acquisition occurs through a transactive communication process with the target of interest; however, this theory is extended to include decisions about passive information-seeking strategies as well. According to TMIM, the information management process is comprised of sequential interpretation, evaluation, and decision stages (Afifi & Weiner, 2004). In the following sections, the stages are described and discussed in relation to information seeking over the Internet.

Interpretation stage

The process of information management in the interpretation stage begins with a perceived disparity between the amount of information possessed and the amount of information desired about a particular issue (Afifi & Weiner, 2004). This perceptual deviation based on comparison levels of desired and enacted uncertainty, called uncertainty discrepancy, inspires emotional arousal specific to the discrepancy (Afifi & Morse, 2009; Fowler & Afifi, 2011). Although various emotions may also be exhibited in response to the discrepancy (Afifi & Morse, 2009), a position consistent with emotional appraisal theories, the primary emotion that has received most attention is anxiety (e.g., Afifi & Afifi, 2009; Afifi, Dillow, & Morse, 2004; Afifi & Weiner, 2006). Emotions such as jealousy and distrust in a romantic partner, possibly from an event eliciting uncertainty discrepancy (e.g., relational transgression), are also commonly discussed in relation to active, conscious online interpersonal information searches (Muise, Christofides, & Desmarais, 2009). Decisions about what information-management strategy to enact are guided not by the desire to reduce uncertainty discrepancy but by the often emotional arousal emerging from the discrepancy (Afifi & Afifi, 2009; Fowler & Afifi, 2011).

Evaluation stage

The next sequence in the information-management process is the evaluation phase. This stage is comprised of specific appraisals that involve the possible outcomes of seeking information and the capacity of the individual to obtain and cope with the information sought (Afifi & Weiner, 2004). Outcome assessments are evaluated along three lines: expectancies, importance, and probability. Outcome expectations are broadly conceptualized as “individuals’ assessments of the benefits and costs of a particular information-seeking strategy” (Afifi & Weiner, 2004, p. 176). Social cognitive theory (SCT; Bandura, 1986, 1989) explains that outcome expectations are based on the probable incentives of performing a behavior. Perceived incentives follow from realized gratifications of performing a behavior or vicarious observation of others. The most relevant of these expected outcomes to online information seeking is novel sensory incentives described as the likelihood of a behavior providing access to a wealth of information that cannot be found elsewhere. Individuals incentivized by novel sensory information obtained through the Internet, provided that the information is seen as important, reliable, and readily accessible, are more likely to use the Internet to seek information (LaRose, Lin, & Eastin, 2003). Outcome expectancies of information seeking are concomitant estimations of information quality (i.e., anticipated reliability of the information) and information valence (i.e., anticipated positive or negative information) (Afifi & Weiner, 2006).

Outcome expectations accompany and exert influence on internal evaluations of the individual performing the behavior. Efficacy appraisals, or the belief that a behavior lies within the capacity of an individual to perform and cope with its outcomes, are important constituents in any goal-oriented or need-satisfaction endeavor (Bandura, 1997). Applied specifically to information seeking, efficacy in the evaluation stage of TMIM is comprised of coping efficacy and Internet self-efficacy (Afifi & Weiner, 2004). Coping efficacy refers to a self-evaluation of one’s ability to manage the emotional and psychological burden associated with the information sought. The belief that individuals are unable to marshal the possible ill effects of pursuing sensitive information about their partners affects the decision to engage in or avoid information seeking. However, it does not have to lie within the scope of the individual to cope with the information in isolation; strong social support networks may also amplify perceptions of coping efficacy (Afifi & Weiner, 2004). Internet self-efficacy, called communication efficacy in TMIM, reflects a sense of confidence that individuals can successfully gather the information of interest through an Internet-based technology. Internet self-efficacy deviates from what has been labeled “technological efficacy,” or the belief that technologies can help information seekers acquire desired communication (Ramirez et al., 2002). For behaviors not innately falling under one’s volitional control, people must hold the belief that they can perform the behavior for them to enact the behavior (Ajzen, 1985, 1987).

Coping efficacy and Internet self-efficacy are germane to a model of online information seeking. LaRose and Eastin (2004) explain that “as Internet users become more self-efficacious, their expectations that they will obtain specific outcomes (e.g., finding useful information) also increase, and that encourages more usage” (p. 362). Before individuals decide to engage in online information seeking, they must believe that they can effectively use the Internet to yield the desired information and cope with the information sought. The decision to engage in or avoid online information seeking or to reassess their informational needs altogether is largely influenced by these two independent efficacy evaluations (Afifi & Weiner, 2004).

TMIM also maintains that expected outcomes of a particular information-management strategy influence efficacy perceptions (Afifi & Weiner, 2004). Anticipating novel sensory incentives from future Internet use elicits imagery involving how individuals would access the information, thereby reinforcing perceptions that they can effectively accomplish the task. Moreover, the visualization also allows people to think about how they would respond to undesirable information retrieved through information seeking, thus fortifying their coping efficacy. Efficacy assessments thus partially mediate the relationship between outcome assessments and the decision to enact an information-management strategy (Afifi & Weiner, 2004).

Decision stage

TMIM proposes three general courses of action for information management: seeking relevant information, avoiding relevant information, or reappraising the situation (Afifi & Weiner, 2004). The decision to engage in any of these general information-management strategies is guided by a hedonic heuristic; that is, the strategy chosen is the one that best mitigates the uncomfortable arousal state caused by the discrepancy perceived between the amount of desired information and information presently possessed (Fowler & Afifi, 2011). In this way, people are able to move away from aversive emotional states toward pleasurable ones. Decisions of information management also rely on favorable perceptions of the information-management strategies established through outcome and efficacy assessments.

Several passive, active, and interactive strategies for information seeking can be employed through the Internet. Passive strategies involve distanced observation of the target. Obtaining information by browsing one’s social networking site or Weblog represents passive online information seeking (Afifi & Lucas, 2008; Antheunis et al., 2010). Internet users can seek information, using active strategies, with search engine queries or by posting content that elicits responses from other users (Gibbs et al., 2011). Interactive strategies involve direct contact and information sharing between at least two parties over the Internet (Antheunis, Schouten, Valkenburg, & Peter, 2012; Tidwell & Walther, 2002). Instant messaging, e-mail, and video chat can be used, through transactive exchanges, to seek information online. Asynchronicity, anonymity, and accessibility are some online affordances that can motivate individuals to pursue information through communication technologies (Valkenburg & Peter, 2011). Although the anticipated valence of social information affects one’s information-seeking behaviors, the type of medium through which information is communicated, whether rich or lean, seems to play a nominal role in information-management decisions (Ramirez & Burgoon, 2004).

Application of TMIM to online information seeking

TMIM provides a framework for explaining information management in the initial (Afifi & Lucas, 2008) and developed stages (Afifi & Weiner, 2006) of a relationship. The fundamental difference between previous information-seeking theories and TMIM is the etiological foundations of the enacted information-management strategy (Afifi & Weiner, 2004). In the context of online information seeking, uncertainty discrepancy begins the causal chain that ultimately motivates one’s decision to seek information through the Internet. The incongruity between how much information is desired and how much information is possessed originates from various conditions specific to the relationship, such as the need for attraction-related or general information, and events outside the relationship (e.g., an ambiguous message from a third party to the partner). Uncertainty discrepancy from factors internal or external to the relationship initiates anxiety related to the uncertainty discrepancy. Although Ramirez, Walther, Burgoon, and Sunnafrank (2002) acknowledge the relative unimportance of uncertainty as a motivator of online information-seeking behaviors, uncertainty discrepancy explains the locomotion of information seeking online that happens across different relationships at various stages, not only in the initial stages of new computer-mediated relationships. The motivation to reduce unpleasant arousal states, such as anxiety, prompts individuals to evaluate the properties of different information-management strategies.

In the evaluation stage, individuals consider the anticipated information they may discover about the target. The quality of information is also considered when appraising outcome expectations. More reliable and positive information from online sources improves the likelihood that people employ online information seeking. Given that information obtained through the Internet is commonly perceived as reliable and credible (Flanagin & Metzger, 2000), people increasingly rely on the Internet to seek information on various topics, including others. People must also believe that they have a strong enough command over the Internet to obtain information about their target, either actively or passively, through online information seeking and cope with whatever information is gathered. Online information seeking occurs to the extent that favorable outcome assessment and efficacy conditions are met.

Finally, individuals must make the rational decision to engage in an information-management strategy based on their outcome assessment and efficacy. Because many communication technologies may be consulted to seek information on others, the information seeker takes into account technological affordances (e.g., sychronicity, accessibility, anonymity), quantity and quality of information, goal types, temporal considerations, and one’s preferences and skills (see Ramirez et al., 2002, for discussion). This is also not to say that online information seeking is the only means by which information on others can be gathered. Indeed, online information seeking may be used in collaboration with traditional methods of information seeking, and there are also some who would never consider using the Internet to seek interpersonal information (Ramirez & Zhang, 2007). Nonetheless, the central focus of this investigation is to examine factors that influence one’s decision to seek information on others over the Internet. The proposed conceptual model of the application of TMIM to online information seeking is provided in Figure 1. This model also indicates seven testable pathways through which uncertainty discrepancy indirectly influences information seeking in interpersonal relationships.

Figure 1.

Model applying the theory of motivated information management to online information seeking.

Internet technologies are often used with success to locate information about well-known and lesser known others. E-mails, instant messengers, and social networking sites are among some of the Internet applications often employed in online information seeking. These technologies, however, are not exercised arbitrarily in information seeking; their use depends on the nature of the relationship between the target and information seeker. Active and interactive information-management strategies are instituted when the person is well known to the seeker (Westerman et al., 2008). For lesser known targets, passive information-management strategies, such as consulting Weblogs and social networking sites, are used. The disparity in technology use in online information seeking gives rise to many important questions about the generalizability of TMIM to various relationship types. Online information seeking, for instance, may not occur regularly in serious dating relationships and spousal relationships except when motivated by a relational event (e.g., transgressions, third-party rival). Friends and casual daters, in contrast, might commonly engage in online information seeking keeping in mind the favorable affordances of the Internet. Afifi and Lucas (2008) suggest that TMIM may be more suitable in explaining information management for certain relationship types than others. The second goal of this investigation is to examine whether relationship type moderates the associations of TMIM when applied to Internet use.

Method

Procedures

Two recruitment methods were employed to collect participants for the present study. Undergraduate students from a university in the US were asked to participate in the study in exchange for research credit. In an effort to expand diversity of the sample in age, education, Internet exposure, and Internet experience, students were asked to refer nonstudent adults who were willing to participate. The proportion of the sample that came from the nonstudent adult population was 31% (n = 197) and the remaining 69% (n = 440) were undergraduate students. Participants were asked to remotely complete an Internet-based questionnaire comprising seven scales and five demographic questions. A series of t tests conducted between the student and nonstudent adult subsamples, using enacted online information seeking, outcome evaluations, efficacy, and anxiety as dependent variables, were nonsignificant. These null difference tests provided evidence that the subsamples could be treated as a larger sample.

In the instructions of the questionnaire, participants were asked to think about a specific interpersonal relationship in which they are currently involved. When asked about this relationship, they were told to consider their current romantic partner. If they were not presently involved in a romantic relationship, they were told to think about a close friend, causal dating partner, or family member with whom they have frequent contact. Although most participants responded to the questionnaire with a serious dating partner in mind (27.6%; n = 176), some reported on close friendships (24.7%; n = 157), best friendships (17.9%; n = 114), casual dating relationships (14.9%; n = 95), spouses (10.7%; n = 68), and their relationships with family members (4.2%; n = 27).

Participants

In total, 637 (308 males and 329 females) participants completed the questionnaire in full. Eleven cases were removed from the data set because the respondents failed to complete entire scales. The average age of the participants was 24.64 years (standard deviation (SD) = 10.07; range = 17–65). The ethnic makeup of this sample was as follows: 60.7% Asian, 22.4% White/Caucasian, 11.1% Hawaiian/Pacific Islander, 3.3% Latino/a, 2.2% African American/Black, and 0.3% Native American. Most relationships (78.3%; n = 499) reported on were considered geographically close (i.e., both partners resided in the same geographical region), although 21.7% (n = 138) of the respondents were in long-distance relationships. The large majority reported on cross-sex relationships (72.4%; n = 461), but some of the relationships were with individuals of the same sex (27.6%; n = 176).

Instruments

Uncertainty discrepancy about a partner

The majority of scales employed in this project were adapted from previous measures used to test TMIM (Afifi & Weiner, 2006). Four items of the uncertainty discrepancy scale evaluated the amount of information, on various aspects of life (e.g., general interests, interpersonal relationships), the respondent had on a partner. The additional four statements related to the amount of information desired on the various aspects of the partner. An index of uncertainty discrepancy, for a given aspect of a partner’s life, was determined by subtracting scores for the corresponding items. For example, to determine uncertainty discrepancy of a partner’s life, the score from the item “How much information do you know about your partner’s life” would be subtracted from the score for the item “How much information do you want to know about your partner’s life.” The items were rated on a 7-point Likert scale from 1 (nothing) to 7 (everything). Higher difference scores on the scale reflect greater uncertainty discrepancy.

Anxiety about uncertainty discrepancy

The anxiety measure, developed by Afifi and Weiner (2006), evaluates the uncomfortable state created by the uncertainty discrepancy. The four items, measured on a 7-point Likert scale, include statements such as “The size of the similarity/difference between how much I know and how much I’d like to know about my partner is … [extremely comforting/anxiety producing] and [soothing/nerve wracking].” An additional item that probed the general anxiety experienced in the relationship (“My relationship with my partner makes me feel anxious”) was included in this measure. This item was measured on a 7-point Likert scale, with anchors at 1 (strongly disagree) and 7 (strongly agree).

Outcome expectations

Anticipated outcomes of online information seeking were measured using two subscales created by Afifi and Weiner (2006). The first subscale, comprising three items, evaluated the prospective valence of the information; that is, whether participants expected the information search will yield good or bad news (e.g., “An online search for information on my partner would produce [extremely negative/extremely positive] information”). Items of the information valence scale were measured on a 7-point Likert scale from extremely negative (1) to extremely positive (7). The second subscale measured the expected quality and reliability of the sought information (e.g., “An online search for information on my partner would yield [extremely bad quality/extremely good quality] information”). The four-item information quality subscale was measured using a 7-point Likert scale, with anchors at extremely bad quality (1) and extremely good quality (7).

Efficacy

Two types of efficacy were measured in the present study: coping efficacy and Internet self-efficacy. The measure for coping efficacy was adapted from a scale developed by Afifi and Weiner (2006). The four-item coping efficacy measure probed participants’ level of agreement with statements such as “I feel I can manage discovering information about my partner I find online” and “I have a strong support system that would help me to manage the information I discover about my partner online.” Higher scores on the coping efficacy scale, measured on a scale from 1 (strongly disagree) to 7 (strongly agree), represented greater self-confidence in one’s ability to cope with the information discovered through information seeking.

The self-efficacy measure used in this investigation was modified items from Eastin and LaRose’s (2000) Internet Self-Efficacy Scale. The four-item scale included statements such as “I feel confident using Internet software to search for information about my partner” and “I have advanced skills for finding information on my partner using Internet programs.” Items were measured on a 7-point Likert scale, with scores of 1 representing strongly disagree and 7 representing strong agree. Larger values on this scale indicate stronger confidence in one’s online information-seeking skills.

Online information-seeking strategy

Four items were used to evaluate the extent to which individuals actively sought information about their partners through the Internet. The general awareness of partners’ presence (i.e., digital footprint) on the Internet was also probed. The focal interest of this study was online information seeking, so other measures of information management (i.e., information avoidance, reevaluation of uncertainty discrepancy) were not included in the questionnaire. The items included “I used the Internet to find information about my partner” and “I look at comments, pictures, and/or messages written about my partner.” Higher scores on items, which were measured on a 7-point Likert scale from strongly disagree (1) to strongly agree (7), suggest that individuals have used the Internet in the past to seek information on their partners.

Results

Seven cases had data that were assumed to be missing at random. Missing data were estimated using a multiple imputation (MI) procedure, which approximates plausible values for the absent responses. A parametric regression method was used in the MI procedure because the pattern of missing data followed a monotoned distribution. Only one imputation was requested in the procedure given the nominal percentage of missing data. The imputed data set was used to test the proposed model. Additional information on scale reliability coefficients (i.e., Cronbach’s α), a complete zero-order correlation matrix, means, and SDs of all variables in the model are provided in Table 1.

Table 1.

Cronbach’s α reliabilities, means, standard deviations, and zero-order correlation matrix for variables in the study.

Variable	α	M	SD	1	2	3	4	5	6
1. Uncertainty discrepancy	.76	0.03	0.93	–
2. Anxiety	.88	2.62	1.13	.28	–
3. OE: information valence	.94	4.90	1.29	−.06	−.40	–
4. OE: information quality	.96	4.82	1.30	−.05	−.33	.70	–
5. Internet self-efficacy	.93	5.48	1.26	−.05	−.10	.24	.25	–
6. Coping efficacy	.94	4.23	1.65	−.11	−.45	.43	.38	.28	–
7. Online information seeking	.89	2.89	1.40	.17	.27	−.06	−.05	.27	−.05

Note. OE: outcome expectation; SD: standard deviation; M: mean.

The eight-factor measurement model was initially evaluated, by means of confirmatory factor analysis (CFA), to see whether the items of each factor demonstrated internal consistency and parallelism. Seven factors exerted independent influence on their respective manifest variables; anticipated information quality and information valence were modeled as a first-order factors of a generalized outcome expectancy second-order factor. The fit of the measurement and structural models were assessed using Hu and Bentler’s (1999) dual criteria. A good fitting model under these criteria should have a comparative fit index (CFI) greater than or equal to .96 and a standardized root mean square residual (SRMR) value less than or equal to .10. The CFA finds that the measurement model fits the data well, χ²(490) = 1079.40, p < .001, CFI = .97, root mean square error of approximation (RMSEA) = .044, 90% confidence interval (CI) [.040, .047], SRMR = .05.

Structural equation modeling (SEM) was used to evaluate the proposed relationships in the application of TMIM to online information seeking. SEM tests the global fit of the overall model to the data and also provides parameter estimates for the local relationships. The hypothesized model tested through SEM is illustrated in Figure 1. In this model, uncertainty discrepancy predicts anxiety specific to the uncertainty discrepancy. This anxiety, in turn, leads to outcome expectancy, Internet self-efficacy, coping efficacy, and online information seeking; the efficacy components are also modeled as predictors of online information seeking. Lastly, outcome expectancy exerts influence on Internet self-efficacy and coping efficacy in the proposed model.

The fit indices of the SEM suggest that explaining online information seeking through principles found in TMIM is tenable, χ²(509) = 1,099.44, p < .001, CFI = .97, RMSEA = .040, 90% CI [.040, .047], SRMR = .06. The modification indices (i.e., LaGrange Multiplier or Wald test) did not specify paths that would significantly improve model fit if they were included or excluded. Standardized path coefficients and variance explained in the endogenous variables are presented in Figure 2. All but two paths in the model are significant. Anxiety related to the uncertainty discrepancy does not directly predict Internet self-efficacy (β = .03, SE = 0.049, p = .61); the path from coping efficacy to online information seeking is also nonsignificant (β = −.05, SE = 0.048, p = .32).

Figure 2.

Final model that applies theory of motivated information management to online information seeking. The parameter estimates provided in the model are standardized coefficients. The estimates above each endogenous variable represent the R ² value. *p < .05, **p < .01, ***p < .001.

The model presents seven different pathways in which uncertainty discrepancy indirectly predicts online information seeking. These pathways indicate various indirect effects that can be tested to ascertain the principal evaluative and interpretive processes individuals undergo in their decision to enact online information-seeking behaviors. In the model, Internet self-efficacy and coping efficacy partially mediate the effect of anxiety on online information seeking. Additionally, Internet self-efficacy and coping efficacy partially mediate the effect of outcome expectancy on online information seeking. Anxiety also mediates the broader relationship of uncertainty discrepancy and information seeking.

Indirect effects were probed using the PROCESS macro (Hayes, 2013), a flexible computational tool that can test multiple mediator models with variables operating in parallel or in sequence (i.e., multiple mediators linearly organized). The observed indicators from the SEM analysis were averaged into item parcels for this test. Tests for mediation of variables in sequence (Model 6) were used for this set of analyses. The indirect effect coefficients and the associated bias corrected 95% CIs for the estimates of the 5,000-bootstrap resamples procedure are presented in Table 2. Bootstrapping, an asymptotic resampling procedure, is desirable for testing indirect effects because the method does not rely on normal theory (see Hayes, 2009, for review). Bootstrapping instead empirically derives the sampling distribution of the product of coefficients through an iterative data resampling procedure. Two coefficients are presented for each indirect effect: the ratio of the indirect effect relative to the total effect (P _M) and the completely standardized indirect effect (ab _cs). Preacher and Kelley (2011) recommend the completely standardized indirect effect for its interpretable scale, its accompanying CIs, and independence of sample size. A 95% CI around an indirect effect estimate that does not include zero specifies a statistically significant coefficient.

Table 2.

Indirect effects of uncertainty discrepancy on online information seeking.

	Mediation	Indirect effect coefficient	Boot SE	Bootstrapping
				Bias-corrected 95% CIs
				Lower	Upper
1.	UD → ANX → ISE → OIS	−0.008	0.004	−0.017	−0.001
		−0.047	0.028	−0.114	−0.007
2.	UD → ANX → CE → OIS	−0.012	0.006	−0.026	−0.001
		−0.070	0.046	−0.184	−0.003
3.	UD → ANX → OE₁ → ISE → OIS	−0.009	0.002	−0.014	−0.005
		−0.050	0.022	−0.109	−0.025
4.	UD → ANX → OE₂ → ISE → OIS	−0.008	0.002	−0.012	−0.004
		−0.043	0.018	−0.091	−0.022
5.	UD → ANX → OE₁ → CE → OIS	−0.003	0.002	−0.007	0.0001
		−0.017	0.013	−0.051	0.001
6.	UD → ANX → OE₂ → CE → OIS	−0.002	0.001	−0.006	0.012
		−0.013	0.010	−0.039	0.0003
7.	UD → ANX → OIS	0.069	0.017	0.041	0.108
		0.663	5.661	0.253	3.630

Note. UD: uncertainty discrepancy; ANX: anxiety; OE₁: outcome expectations (information quality); OE₂: outcome expectations (information valence); ISE: Internet self-efficacy; CE: coping efficacy; OIS: online information seeking. The upper row for each mediation test represents the completely standardized indirect effect point estimate, bootstrapped standard error, and bootstrapped 95% CIs. The lower row for each mediation test represents estimates of the ratio of indirect effect to total effect.

The results of the indirect effect tests underscore the sound theoretical structure of TMIM in its application to online information seeking. Although paths that lead from uncertainty discrepancy to information seeking through anxiety, outcome expectancy, and coping efficacy sequentially were nonsignificant (IE estimate [for information quality] = .003, 95% CI [−.007, .0001], SE = 0.002; IE estimate [for information valence] = −.002, 95% CI [−.006, .012], SE = .001), all other routes from uncertainty discrepancy to online information seeking were substantiated.

Differences in online information seeking among relationship types

A multigroup SEM (Jöreskog, 1971) was conducted to examine whether the parameters in the model change as a function of relationship type. Multigroup SEM is useful when the larger sample of data collected is comprised of distinct subsamples (Yuan & Bentler, 2001). Multigroup SEM can be interpreted as a moderator test for overall model fit and the local path coefficients. Cases in the data set were assigned to one of four groups (i.e., friendships, best friendships, casual dating relationships, and serious dating or spousal relationships) determined by the kind of relationship that participants reported. The models tested for each group were equivalent to the model pictured in Figure 1, with the exception that all variables except for the outcome expectancy latent variable were observed. These observed variables were parcels of the items used in the SEM analysis. The indices used in determining goodness of fit of the four SEMs are provided in Table 3. The results indicate that the hypothesized TMIM in the context of online information seeking fits the data well for all types of relationships except best friendships, χ²(9) = 19.02, p = .03, CFI = .91, RMSEA = .10, 90% CI [.03, .16], SRMR = .07, where the CFI falls short of the conventional standards for good model fit (Hu & Bentler, 1999).

Table 3.

Fit indices for the theory of motivated information management of online information seeking by relationship type.

	Model	N	χ²	df	p	CFI	RMSEA	RMSEA 90% CI	SRMR
1.	Friends	157	17.93	9	.04	.97	.079	[.020, .134]	.08
2.	Best friends	114	19.02	9	.03	.91	.099	[.034, .161]	.07
3.	Casual daters	95	8.80	9	.46	1.00	.000	[.000, .114]	.03
4.	Serious daters and spouses	224	18.93	9	.03	.97	.067	[.023, .110]	.05
5.	Combined groups	610^a	64.67	36	.002	.97	.073	[.043, .101]	.06

Note. CFI: comparative fit index; RMSEA: root mean square error of approximation; SRMR: standardized root mean square residual.

^a Participants reporting on relationships with family members were removed from these analyses.

A stepwise procedure was used to estimate the multigroup SEM. The first step of the process is to fit an unconstrained multigroup factor model to determine configural invariance. The second model, which assessed between-group equivalence of factor loadings, constrained all path coefficients across the groups to be equal. Third, a series of models tested between-group equivalence of individual path coefficients, freeing a path in sequence using recommendations of the modification indices. Fit in the third step was determined using model chi-square change, given the nested nature of the models. The chi-square test of the unconstrained multigroup factor model demonstrated reasonable fit, χ²(36) = 64.67, p = .002. The second model, which constrained all path coefficients across the four groups to be equal, showed consistency with the data, χ²(66) = 106.57, p = .001, and the χ² difference test showed no marked difference in model fit, χ² _Δ(df _Δ = 30) = 41.90, p = .07.

The modification indices of the multigroup SEM suggested four paths that could be freed to improve the overall model fit. Releasing the equality constraint for the path between anxiety and information seeking for the best friendship group was initially suggested by the modification indices. The overall model fit with this unconstrained path improved, χ²(65) = 98.90, p = .004, and the model difference test was significant, χ² _Δ(df _Δ = 1) = 7.67, p = .006. The second parameter estimated represented the path between outcome expectancy and coping efficacy for close friends. Freeing this constrained parameter improved the fit of the model demonstrated by the model chi-square, χ²(64) = 91.47, p = .01, and the χ² change test, χ² _Δ(df _Δ = 1) = 7.43, p = .006. The third unconstrained path was between anxiety and coping efficacy for casual dating partners. Model fit was improved by removing this equality constraint; the model chi-square was χ²(63) = 86.13, p = .03, and the model difference test was χ² _Δ(df _Δ = 1) = 5.34, p = .02. The final parameter estimated was the path between Internet self-efficacy and online information seeking for the best friendship group. Freeing the constrained path improved global model fit, χ²(62) = 81.79, p = .05, and the chi-square change was significant, χ² _Δ(df _Δ = 1) = 4.34, p = .04. Freeing other constrained parameters, in accordance with the modification indices, would not have significantly improved the model fit. The final multigroup SEM, illustrated in Figure 3, provides the common path coefficients shared by the groups and unique coefficients that were freed in Step 3 of the stepwise process. Table 4 presents the standardized path coefficients of the model stratified by relationship types.

Figure 3.

Multigroup structural equation model applying theory of motivated information management to online information seeking. G1 = close friendship group; G2 = best friendship group; G3 = casual dating partner group; and G4 = serious dating partner/spousal group. The parameter estimates provided in the model are unstandardized coefficients. Paths share common coefficients across groups unless specified in the diagram. *p < .05, **p < .01, ***p < .001.

Table 4.

Standardized path coefficients for the theory of motivated information management of online information seeking by relationship type.

	Relational type	UD → ANX	ANX → OE	ANX → ISE	OE → ISE	ANX → CE	OE → CE	ANX → OIS	ISE → OIS	CE → OIS
1.	Friends	.21**	−.41***	.05	.26**	−.28***	.45***	.38***	.28***	.01
2.	Best friends	.25**	−.34***	.08	.20*	−.38***	−.02	.12	.13	.05
3.	Casual daters	.30***	−.34***	.04	.28**	−.06	.46***	.44***	.33***	−.09
4.	Serious daters and spouses	.29***	−.37***	.05	.33***	−.38***	.29***	.29***	.35***	−.02

Note. UD: uncertainty discrepancy; ANX: anxiety; OE: outcome expectations; ISE: Internet self-efficacy; CE: coping efficacy; OIS: online information seeking.

*p < .05; **p < .01; ***p < .001.

Discussion

The present investigation applied TMIM (Afifi & Weiner, 2004) to online information seeking in response to the growing use of Internet technologies to seek information on others. Although the general TMIM framework reliably explains the interpretation, evaluation, and decision-making processes involved with information management in a wide range of contexts, such as organ donation (Afifi et al., 2006), sexual health (Afifi & Weiner, 2006), and parental caregiving (Fowler & Afifi, 2011), research on TMIM has been limited to traditional modes of information seeking. From its very inception, TMIM has been described as “a theory framed within interpersonal contexts” (Afifi & Weiner, 2004, p. 168) and used almost exclusively to examine in-person transactive communication. However, in a world that is becoming increasingly dependent on the Internet as a tool for communication, information storage, and archival, the Internet cannot be ignored as viable and often desirable means of interpersonal information acquisition.

The findings indicate that a discrepancy between the amount of information desired and possessed on a target motivates the locomotion of events that in the end determines online information seeking. A large discrepancy in uncertainty inspires feelings of anxiety, which in turn influence outcome expectancies and coping efficacy. Anxiety, however, has little effect on the level of personal confidence in online searching skills, demonstrating the impervious nature of Internet efficacy evaluations from one’s aversive emotional states. Quality and valence expectations of online information seeking affect efficacy perceptions. Internet self-efficacy and anxiety conspire to predict information-seeking behaviors.

The general principles of TMIM are stable across relationship types, with the exception of best friendships. Best friends seem to be the least preoccupied with engagement in online information seeking; neither the interpretation nor the evaluation stage encourages best friends to seek interpersonal information over the Internet. It may be that the nature of the information sought between best friends is more germane to in-person interactive information seeking than for mere friends. Moreover, the wide range of transgressions that might motivate information seeking in romantic relationships may not arise as regularly in best friendships (Casper & Card, 2010), which may account for the deviation in predictions of TMIM for best friends as well.

This investigation makes a significant contribution to the literature on TMIM by broadening the boundaries of the theory and providing a better understanding of interpersonal information management that takes place through the Internet and among different types of relationships. It is no longer necessary to consider TMIM a theory only applicable to traditional modes of information flow or, more importantly, transactive information seeking. Instead, the general tenets of TMIM transcend channels and can be applied to interactive or passive online information seeking as well. The characteristics that trigger information-seeking strategies in offline contexts appear to be the same motivating factors in online information seeking. An important distinction in this application, however, is that expectations about the quality and reliability of the online information, in concert with confidence in one’s Internet abilities, promote online information seeking. Accordingly, people’s beliefs about the Internet and their mastery over it are largely instrumental in their decision to seek interpersonal information online.

Limitations

Findings from the present investigation should be interpreted aside the limitations of the study. First, the study does not discriminate between information seeking that occurs across Internet technologies; instead, various affordances of communication technologies, such as synchronous and asynchronous, interactive and noninteractive, and public and private channels, are collapsed into generalized online information seeking. The shortcoming of this approach is that it does not appreciate the sometimes important differences among these information devices (e.g., Ramirez, 2009). Additionally, given that interpersonal modes of information seeking are excluded from the proposed model, it remains unknown whether the distinction between online and off-line information-seeking strategies made in this project is theoretically meaningful. To reconcile this limitation, future investigations should endeavor to separate different Internet technologies and examine whether the information-management strategies through the different technologies are independently informed by the predictors articulated in TMIM. Further, a comparison between traditional and nontraditional modes of information seeking must be explicated in the future.

Second, the cross-sectional data complicate the causal model tested in the present investigation. Causal conclusions are drawn from the findings of these models without fulfilling necessary prerequisites for making tentative causal statements from cross-sectional data, such as replicating the tests across independent samples and eliminating alternative equivalent explanations (Bullock, Harlow, & Muliak, 1994). Nevertheless, findings of this study are consistent with other investigations that have tested TMIM using longitudinal data, providing confidence in the causal relationships between the variables. Future investigations should replicate these findings with experimental or longitudinal data given some inconsistencies between TMIM and other theories regarding the directional relationship between outcome expectancy and self-efficacy. For instance, although TMIM argues that one’s evaluation of self-efficacy is shaped by outcome expectations, LaRose and Eastin’s (2004) SCT of Internet use contends that self-efficacy appraisals precede outcome expectations. Future investigation should bring clarity to this and other theoretical inconsistencies.

Lastly, future research should consider the differences in motives of online information seeking, which exist between the mixed-mode (offline/online) relationships of interest in the present investigation and Internet-only relationships. Previous research indicates deviations in the personal reasons for seeking information and the strategies used for seeking information on others between these two types of relationships (Antheunis et al., 2012). Disentangling exclusively text-based from audiovisual inclusive Internet-only relationships will further improve the predictive efficacy of TMIM in online information seeking.

Future directions

The support for the application of TMIM to online information seeking in this research project opens many new avenues for future research. TMIM explains how individuals manage anxiety caused by the discrepancy between desired and possessed levels of information. What is noticeably absent from the findings is an understanding of the media selection process that individuals with uncertainty discrepancy and attendant anxiety undergo. Future studies would benefit from a theoretical integration in which traditional and nontraditional modes of information seeking are tested together under a TMIM framework. A project that incorporates multiple avenues of information seeking for an information seeker would yield important information about channel selection. TMIM would then be able to forecast instances in which information seeking in a given medium would be selected over another medium, thereby expanding the theory’s predictive utility.

Beyond future research opportunities on the theoretical level, there are also additional considerations to be made on the individual level. Given that TMIM provides a framework for explaining motivating factors in offline and now online information-seeking strategies, it would be useful to explore the circumstances under which online information seeking is preferred over in-person information seeking through a dual-channel model (i.e., a model that includes both traditional and online information seeking). Specifically, who tends to seek information through interpersonal and online channels, what kind of information is sought, and what influences these choices are all questions germane to empirical investigation. Considering the important role played by Internet self-efficacy in instigating online information seeking, information-seeking preferences could be distinguished across a digital divide, taking into account differences in age, socioeconomic status, or culture (Livingstone & Helsper, 2007). Also, future research could determine what individual and topical factors influence the similarities and differences between online and off-line information-seeking behaviors.

Conclusion

This investigation provides an understanding of the interpretation, evaluation, and decision-making processes involved with online information seeking. The principal contribution of this project is that it expands the boundaries of TMIM and offers new opportunities for investigating information-management strategies in the diverse domain of Internet technology. These data have significant implications for research and theory on relational development and maintenance involving information accessible through the Internet.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Afifi

W. A.

Afifi

T. D.

(2009). Avoidance among adolescents in conversations about their parents’ relationship: Applying the theory of motivated information management. Journal of Social and Personal Relationships, 26, 488–511.

Afifi

W. A.

Dillow

M. R.

Morse

(2004). Examining predictors and consequences of information seeking in close relationships. Personal Relationships, 11, 429–449.

Afifi

W. A.

Lucas

A. A.

(2008). Information seeking in initial stages of relational development. In Sprecher

Wenzel

Harvey

(Eds.), Handbook of relationship initiation (pp. 197–215). New York, NY: Psychology Press.

Afifi

W. A.

Morgan

S. E.

Morse

Stephenson

M. T.

Harrison

T. R.

Reichert

Long

S. D.

(2006). Examining the decision to talk about organ donation: A test of the theory of motivated information management. Communication Monographs, 73, 188–215.

Afifi

W. A.

Morse

C. R.

(2009). Expanding the role of emotion in the theory of motivated information management. In Afifi

T. D.

Afifi

W. A.

(Eds.), Uncertainty and information regulation in interpersonal contexts: Theories and applications (pp. 88–105). New York, NY: Routledge.

Afifi

W. A.

Weiner

J. L.

(2004). Toward a theory of motivated information management. Communication Theory, 14, 167–190.

Afifi

W. A.

Weiner

J. L.

(2006). Seeking information about sexual health: Applying the theory of motivated information management. Human Communication Research, 32, 35–57.

Ajzen

(1985). From intentions to actions: A theory of planned behavior. In Kuhland

Beckman

(Eds.), Action-control: From cognitions to behavior (pp. 11–39). Heidelberg, Germany: Springer.

Ajzen

(1987). Attitudes, traits, and actions: Dispositional prediction of behavior in personality and social psychology. In Berkowitz

(Ed.), Advances in experimental social psychology: Vol. 20. (pp. 1–63). San Diego, CA: Academic Press.

10.

Antheunis

M. L.

Schouten

A. P.

Valkenburg

P. M.

Peter

. (2012). Interactive uncertainty reduction strategies and verbal affection in computer-mediated communication. Communication Research, 39, 757–780.

11.

Antheunis

M. L.

Valkenburg

P. M.

Peter

(2010). Getting acquainted through social network sites: Testing a model of online uncertainty reduction and social attraction. Computers in Human Behavior, 26, 100–109.

12.

Aspray

Hayes

. (2011). Everyday information: The evolution of information seeking in America. Boston, MA: MIT Press

13.

Bandura

(1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice Hall.

14.

Bandura

(1989). Human agency in social cognitive theory. American Psychologist, 44, 1175–1184.

15.

Bandura

(1997). Self-efficacy: The exercise of control. New York, NY: Freeman.

16.

Berger

C. R.

(1979). Beyond initial interaction: Uncertainty, understanding, and the development of interpersonal relationships. In Giles

Clair

R. St.

(Eds.), Language and social psychology (pp. 122–144). Baltimore, MD: University Park Press.

17.

Berger

C. R.

Calabrese

R. J

. (1975). Some explorations in initial interaction and beyond: Toward a developmental theory of interpersonal communication. Human Communication Research, 1, 99–112.

18.

Brashers

D. E.

Goldsmith

D. J.

Hsieh

(2002). Information seeking and avoiding in health contexts. Human Communication Research, 28, 243–258.

19.

Bullock

H. E.

Harlow

L. L.

Mulaik

S. A.

(1994). Causation issues in structural modeling research. Structural Equation Modeling, 1, 253–267.

20.

Casper

D. M.

Card

N. A

. (2010). “We were best friends, but”: Two studies of antipathetic relationships emerging from broken friendships. Journal of Adolescent Research, 25, 499–526.

21.

Darvell

M. J.

Walsh

S. P.

White

K. M.

(2011). Facebook tells me so: Applying the theory of planned behavior to understand partner-monitoring behavior on Facebook. Cyberpsychology, Behavior, and Social Networking, 14, 717–722.

22.

Douglas

(1990). Uncertainty, information-seeking, and liking during initial interaction. Western Journal of Speech Communication, 54, 66–81.

23.

Eastin

M. S.

LaRose

. (2000). Internet self-efficacy and the psychology of the digital divide. Journal of Computer-Mediated Communication, 6. Retrieved December 18, 2013, from https://onlinelibrary-wiley-com.web.bisu.edu.cn/doi/10.1111/j.1083-6101.2000.tb00110.x/full

24.

Flanagin

A. J.

Metzger

M. J.

(2000). Perceptions of Internet information credibility. Journalism and Mass Communication Quarterly, 77, 515–540.

25.

Fowler

Afifi

W. A.

(2011). Applying the theory of motivated information management to adult children’s discussion of caregiving with aging parents. Journal of Social and Personal Relationships, 28, 507–535.

26.

Gibbs

J. L.

Ellison

N. B.

Lai

C-H.

(2011). First comes love, then comes Google: An investigation of uncertainty reduction strategies and self-disclosure in online dating. Communication Research, 38, 70–100.

27.

Hancock

J. T.

Toma

C. L.

Fenner

(2008). I know something you don’t: The use of asymmetric personal information for interpersonal advantage. Proceedings of the ACM 2008 conference on computer supported cooperative work (pp. 413–416). San Diego, CA.

28.

Hayes

A. F.

(2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Communication Monographs, 76, 408–420.

29.

Hayes

A. F

. (2013). Introduction to mediation, moderation, and conditional process analysis. New York, NY: Guilford Press.

30.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55.

31.

Internet World Stats. (2011). Internet world stats: Usage and population statistics. Retrieved March 23, 2012, from http://www.internetworldstats.com/stats.htm

32.

Jöreskog

K. G

. (1971). Statistical analysis of sets of congeneric tests. Psychometrika, 36, 109–133.

33.

LaRose

Eastin

M. S.

(2004). A social cognitive theory of Internet uses and gratifications: Toward a new model of media attendance. Journal of Broadcasting & Electronic Media, 48, 358–377.

34.

LaRose

Lin

C. A.

Eastin

M. S.

(2003). Unregulated Internet usage: Addiction, habit, or deficient self-regulation? Media Psychology, 5, 225–253.

35.

Livingstone

Helsper

(2007). Gradations in digital inclusion: Children, young people, and the digital divide. New Media & Society, 9, 671–696.

36.

Lyndon

Bonds-Raacke

Cratty

A. D.

(2011). College students’ Facebook stalking of ex-partners. Cyberpsychology, Behavior, and Social Networking, 14, 711–716.

37.

Muise

Christofides

Desmarais

(2009). More information than you ever wanted: Does Facebook bring out the green-eyed monster of jealousy? CyberPsychology & Behavior, 12, 441–444.

38.

Pratt

Wiseman

R. L.

Cody

M. J.

Wendt

P. F.

(1999). Interrogating strategies and information exchange in computer-mediated communication. Communication Quarterly, 47, 46–66.

39.

Preacher

K. J.

Kelley

(2011). Effect size measures for mediation models: Quantitative strategies for communicating indirect effects. Psychological Methods, 16, 93–115.

40.

Ramirez

(2009). The effect of interactivity on initial interactions: The influence of informational seeking role on computer-mediated interaction. Western Journal of Communication, 73, 300–325.

41.

Ramirez

Burgoon

J. K.

(2004). The effect of interactivity on initial interactions: The influence of information valence and modality and information richness on computer-mediated interaction. Communication Monographs, 71, 422–447.

42.

Ramirez

Walther

J. B.

Burgoon

J. K.

Sunnafrank

(2002). Information-seeking strategies, uncertainty, and computer-mediated communication: Toward a conceptual model. Human Communication Research, 28, 213–228.

43.

Ramirez

Zhang

(2007). When online meets offline: The effect of modality switching on relational communication. Communication Monographs, 74, 287–310.

44.

Shoemaker

P. J.

(1996) Hardwired for news: Using biological and cultural evolution to explain the surveillance function. Journal of Communication, 46, 32–47.

45.

Stefanone

M. A.

Hurley

C. M.

Yang

Z. J

. (2013). Antecedents of online information seeking. Information, Communication, & Society, 16, 61–81.

46.

Tidwell

L. C.

Walther

J. B.

(2002). Computer-mediated communication effects on disclosure, impressions, and interpersonal evaluations: Getting to know one another a bit at a time. Human Communication Research, 28, 317–348.

47.

Tokunaga

R. S.

(2011). Social networking site or social surveillance site? Understanding the use of interpersonal electronic surveillance in romantic relationships. Computers in Human Behavior, 27, 705–713.

48.

Utz

Beukeboom

C. J.

(2011). The role of social networking sites in romantic relationships: Effects on jealousy and relationship happiness. Journal of Computer-Mediated Communication, 16, 511–527.

49.

Valkenburg

P. M.

Peter

(2011). Online communication among adolescents: An integrated model of its attraction, opportunities, and risks. Journal of Adolescent Health, 48, 121–127.

50.

Westerman

Van Der Heide

Klein

K. A.

Walther

J. B

. (2008). How do people really seek information about others?: Information seeking across internet and traditional communication channels. Journal of Computer-Mediated Communication, 13, 751–767.

51.

Yuan

K. H.

Bentler

P. M

. (2001). A unified approach to multigroup structural equation modeling with nonstandard samples. In Marcoulides

G. A.

Schumacker

R. E.

(Eds.), Advanced structural equation modeling: New developments and techniques (pp. 35–56). Mahwah, NJ: Lawrence Erlbaum Associates.