Abstract
The term “self-disclosure” refers to actions by which individuals reveal information about themselves. The interest in such conduct has resurged with the development of networked participatory technologies, which enable creation, dissemination, analysis, and use of large amounts of personal information, thereby increasingly augmenting the effect of online self-disclosure on disclosers and disclosees. This article reviews 309 empirical studies about online self-disclosure published between 2010 and 2020 and aggregates insights thereof into an overarching model describing the ways in which this socio-technical undertaking unfolds. The review shows that online self-disclosure research overwhelmingly focuses on the individual and de-emphasizes structural elements that influence these practices and their outcomes. Based on these findings, we propose a structurational framework centered on the dialectic relationship between individuals and structures involved in self-disclosure and affected by such behaviors. This approach provides a useful tool to address the recent transformations of self-disclosure and its implications in the wake of computer-mediated technological developments.
Keywords
Introduction
For more than five decades now, researchers have been preoccupied with issues related to self-disclosure. The noun “disclosure” derives from the old French word “desclos,” meaning open, exposed, plain, and explicit. Thus, a person who self-disclose puts personal information out in the open, usually information that was formally secret or unknown to others.
In early studies on self-disclosure, this phenomenon was defined as the telling of a previously unknown so that it becomes shared knowledge—the process of making the self known to others (Jourard and Lasakow, 1958). An extensive number of studies about self-disclosure have nuanced this early conceptualization. Some scholars emphasize the nature of this practice, perceiving self-disclosure as the process of revealing concealed information. Others have focused on the characteristics of the information disclosed. For example, defining self-disclosure as a practice by which individuals reveal personally identifiable and private information (e.g. Krasnova et al., 2010). In some studies, the definition to self-disclosure contains references to the media used to disclose and to receive personal information. For example, “Broadcasting status updates/posts on social media that contain any form of self-information” (Lin and Utz, 2017).
Notwithstanding these differences, there is agreement that self-disclosure is vital to the development of relationships and to the happiness of established relationships as it enables people to create intimacy and strong social bonds (e.g. Boyd, 1995; Hayes et al., 2015; Luo and Hancock, 2020). It plays an important role in fostering the construction of self-identity and identity management, in securing individual autonomy, and in enhancing well-being (e.g. Bazarova and Choi, 2014; boyd and Marwick, 2011; Geusens and Beullens, 2017; Gibbs et al., 2010; Masur, 2018; Mehdizadeh, 2010). Self-disclosure also serve society at large by generating diversity, securing spheres for deliberation, enhancing pluralism, and promoting freedom of expression and equality. When strategically employed in a commercial context, self-disclosure enhances profitability of firms particularly of data-driven entities (e.g. Couldry and Mejias, 2018; Draper and Turow, 2019). Likewise, such undertakings support public institutions by enabling them to detect needs and provide a variety of services to both individuals and groups (e.g. Efroni et al., 2019; Fisher, 2022).
Realizing the importance of self-disclosure for both individuals and society at large, academic investigations on self-disclosure have centered on the effects of self-disclosure on those who disclose personal information (hereafter, disclosers or senders), on those exposed to the information disclosed (hereafter, disclosees or recipients), and on the relations formed between these two sides (e.g. Jourard, 1971; Masur, 2018; Omarzu, 2000; Petronio, 2002). The interest in self-disclosure intensified with the development of networked computers that enable various forms of human communication, both synchronous and asynchronous, involving one-to-one, one-to-many, or many-to-many exchanges of text, audio, and/or video messages. Websites, social network sites (SNS), designated apps, wearables, voice assistants, and a wide range of other tools allow individuals to disseminate large amounts of personally identifiable and private information such as hobbies or sexual preferences, to both individuals and groups of various sizes. Mobile technologies, and particularly smartphones and the tools that they support, also enable senders and recipients to reveal and consume large amounts of personal information at any time and place.
A development that fueled research on online self-disclosure (OSD) is the popularization of technologies based on Big Data (e.g. Mayer-Schönberger and Cukier, 2013), which allow institutions to extract and store large amounts of personal information revealed in various online arenas, analyze it, and use it for a wide variety of purposes, including commercial gain.
A finding that warrants a deeper investigation of OSD has to do with the context in which it is practiced: The degree of familiarity between the discloser and the disclosee has been demonstrated to strongly influence both the quantity and the quality of the personal information disclosed, as well as the effects of such behaviors on both sides (e.g. Cozby, 1973; Petronio, 2002). In this connection, studies have shown that the more extreme the nature of the relationship between a sender and a recipient—either great intimacy or total unfamiliarity—the more likely the sender will be to disclose sensitive information to the recipient and the greater the benefits they will derive therefrom (e.g. Cozby, 1973). Networked (online) media (like SNS) tend to blur the degrees of intimacy between people, since they enable the creation and development of new types of relationships which impact disclosure behaviors and their implications (see, e.g. Turkle, 2011). A related phenomenon, known as “context collapse,” is the flattening of multiple audiences—particularly SNS users—into a single context, which makes the management of social interactions and of the self across varied settings an increasingly complicated task (e.g. Davis and Jurgenson, 2014; Marwick and boyd, 2011; Masur, 2018; Zillich and Müller, 2019).
Given the centrality of self-disclosure practices in online spaces, and their effects on individuals, institutions, and on society at large in almost all its domains, this literature review sets out to examine a growing body of interdisciplinary research on OSD. Building on earlier definitions of “self-disclosure”, we conceive OSD as a process in which the individual agent—intentionally and/or unintentionally—reveals, by means of computer-mediated technologies, personal data and information that was formally secret or unknown to others. The widespread adoption of computer-mediated technologies adds complexity to self-disclosure as “public by default” (boyd and Marwick, 2011) becomes a key characteristic of communication in the digital era, thus affecting individuals’ ability to make decisions about what data or information to disclose and how to disclose it.
By encompassing several main perspectives such as social science, education, health, and management, the review sheds light on the complex nature and ramifications of OSD and offers an analytical framework to address the transformations in self-disclosure in the wake of the development of computer-mediated technologies.
In what follows, we first discuss the methodology and research procedures and then present the empirical findings that emerged from reviewing 309 articles on OSD. The results are comprehensively formulated in a model entitled An Interdisciplinary Online Self-Disclosure Model. The article proceeds to map out directions for future research, based on insights from two research streams: the body of research on OSD which focuses on the individual agent who discloses personal information online, and the critical research on datafication which highlights issues of power that may affect self-disclosure practices and the relationships between the different agents involved therein (e.g. Couldry and Mejias, 2018; Kennedy, 2018; Turow, 2021; Zuboff, 2019). In the last section, we introduce a framework for OSD studies that we term “a practice lens to study OSD.” The framework underscores the dialectics of the relationship between agents and structures that both enable and constrain self-disclosure practices and outcomes. As such, it can be used in future studies to account for the dialectics of self-disclosure at both the individual and collective levels. We submit that such research is essential given the rapid development of a wide range of popular networked technologies that allow individuals to share personal information and disseminate it to the masses and due to the expansion of technologies which enable the information thus revealed to be harvested, stored, analyzed, and used for a wide variety of purposes by powerful entities, both public and commercial.
Methodology: sample, procedures, and limitations
This review discusses empirical articles devoted to OSD and published in academic journals. The following criteria were requisite for an article to be included in the corpus: (a) The discussion in the article centers on OSD, (b) the article was published in a peer-reviewed journal, and (c) the article reports primary empirical findings. In discussing the various articles that meet the aforementioned criteria, we also refer to works that present secondary findings (like literature reviews), but these studies were not included in the coded corpus.
The choice of the articles was based on the three-stage method suggested (among others) by Smith et al. (2011), who reviewed the literature on a related topic—information privacy.
At the first stage, we scanned the Web of Science (WoS) database, which is an interdisciplinary repository of academic articles. We chose WoS because it is the most recognized database for peer-reviewed content and well known to librarians and the research community (e.g. Mikki, 2009). To minimize publication bias, we did not include articles outside this database, unpublished works, or works in languages other than English. Clearly, the decision to eschew multiple databases and opt out of translating works originally published in a foreign language is fraught with limitations, but it was grounded in solid considerations of cost and benefits, the principle whereof is succinctly summarized by Petticrew and Roberts (2008): “[I]n situation where resources are limited, thorough quality assessments should take precedence over extensive literature search and translation of articles” (p. 236). Another reason for relying on WoS database has to do with transparency. Since WoS is the most recognized repository of peer-reviewed articles (e.g. Mikki, 2009), we assumed that its readers would be in a position to assess the exhaustiveness of a review. In addition, a review based on such a prestigious database would benefit other scholars in the field, who would be assured of the validity of the results presented. Finally, WoS includes only the articles published in leading journals, thus minimizing biases related to the quality assessment of our corpus.
The corpus included all studies in the database published from January 2010 to January 2020, a decade marked by the development of techno-social wherewithal that has enabled the creation, dissemination, analysis, and use of large amounts of disclosed personal information. Thus, during these years, the effect of OSD on both disclosers and disclosees was increasingly amplified. The decade targeted was preceded by the promulgation of multiple technologies including Facebook, which was founded in 2004 and opened its membership to people other than students in 2006; Apple’s smartphone, introduced in 2007; and mobile applications and voice recognition devices added after mobile participatory media had been widely popularized. We set our search as of 2010, on the assumption that it would have taken 2–3 years to complete research into self-disclosure enabled by these new-fangled technologies and to publish it in leading peer-reviewed journals listed in the WoS database. We ended the search with articles published in 2020, as in 2018, the EU had introduced The Data Protection Act (General Data Protection Regulation [GDPR]), a legislature that, in principle, has the power to affect OSD and its outcomes.
We searched for articles where either the title or list of key words included at least one of these words: “self-disclosure,” “disclosure,” or “self-presentation.” At the risk of excluding some relevant works, particularly studies on privacy and surveillance, we target other verbs associated with self-disclosing, such as “tagging” and “sharing.” Although self-disclosure practices are tightly linked to privacy—both concepts relate to the control over the flow of personal information—our focus was on self-disclosure, namely on specific practices in which the agent puts information about him or her out in the open (dis-closing), usually personal information that was formally unknown to others. This is in contrast to privacy, which often relates to the circumstances in which a person controls the flow of information—revealing it to other or concealing it (e.g. Marwick, 2023).
Given our focus on OSD, each keyword (above) was cross-checked with the keywords semantically linked to the lexemes “Internet,” “World Wide Web,” “online,” “on-line,” “social network sites” or its accepted abbreviation (SNS), and “mobile applications.”
Next, we analyzed the abstracts of the articles selected and ended up with a total of 369 articles. After removing the articles which did not discuss OSD, did not include empirical studies that presented primary findings, or were published in multiple versions, 309 articles remained. We then identified the 50 most frequently quoted articles by searching for the article title using the Google Scholar search engine. The full text of all 50 articles was read by three researchers (the two leading researchers and a research assistant), in order to set the categories (such as actors involved, OSD, nature of information disclosed, and so on) and subcategories (such as individual discloser, individual disclosee, public institution, commercial institution, benefits and losses associated with OSD, and so on) for further analysis. We assumed that a subset of the most frequently cited articles would be optimal for inductively identifying the relevant categories.
Next, we compiled a coding schema to analyze the entire corpus (309 articles), following some useful guidelines on writing literature reviews (Cooper et al., 2019; Okoli, 2015). To this end, three researchers (the leading researcher and two research assistants) analyzed 40 articles randomly selected from the entire corpus (approximately 10%) and adjusted the coding schema accordingly. Thereupon, each coder independently coded another set of 40 articles randomly selected from the entire corpus (approximately 10%). Based on that second round of coding, the coding schema was adjusted again.
Next, the entire corpus of 309 articles was divided between two of the researchers, who coded them separately. A third researcher (the leading author) analyzed a random set of articles (10% of the corpus) to assess the reliability of the coding. Inter-coder reliability was based on a simple agreement between coders and ranged between .88 and 1.
After the coding of the entire corpus had been completed, we formulated a model that integrates the categories and subcategories of the coding schema. The model, under the title An Interdisciplinary Online Self-Disclosure Model, is presented in Figure 1. It is intended as a tool to explore OSD in the digital age, including incentives, practices, and possible outcomes.
An interdisciplinary online self-disclosure model.
Empirical findings
In the following section, we present the results of 10 years of research on OSD conducted from different disciplinary perspectives, with a view to promoting a more accurate systematization of this phenomenon and a deeper understanding of its complexity. The results are comprehensively formulated in the Interdisciplinary Online Self-Disclosure Model introduced earlier (see Figure 1), beginning with the actors involved in OSD and ending with benefits and losses entailed by such practices.
Actors and roles
What actors are involved in OSD?
The first component of the proposed model is the actors involved in OSD. At least in principle, there are two such actors (or agents)—the discloser and the disclosee. However, in the online arenas, a third, mediating, party operates—the mediator of the disclosure. This mediator can be an agent—individuals or groups that disseminate the disclosed information or create the tool or platform which serves to produce and/or circulate it. In addition, since mediating technologies enable the dissemination of messages to large audiences, self-disclosure in online environments affects institutions (both commercial and public) which develop and/or operate these technologies, as well as the society as a whole, which is either directly exposed to the information or indirectly influenced by it. Accordingly, the main actors involved in OSD were identified in the articles reviewed. In the instances where the same article examined more than one actor, the article was coded more than once.
In 276 articles, the main actor studied was the individual discloser. A core concern was how individuals perceive self-disclosure and the characteristics of their self-disclosure behaviors. That said, in 38 articles, the actor was the individual or group exposed to self-disclosure. Three of the articles referred to the actor as “other,” for example, the research of Ashuri and Halperin (2017) that examined media coverage on OSD and pointed to journalists’ perceptions regarding the benefits and losses incurred by the actors involved in OSD.
Of the aforementioned 276 articles, 125 examined individuals (disclosers or disclosees) belonging to highly differentiated groups: sportsmen and sportswomen (4 studies), patients or people with a medical problem (11 studies), bloggers (12 studies), people seeking relationships (5 studies), teaching staff (6 studies), young adults (7 studies), and children and adolescents (41 studies). An additional 31 studies also discussed a distinct group, specifically as disclosers, but on average, only four each, for example, politicians, celebrities, web personalities, pregnant women, mothers, gamers, journalists, and sex offenders. Another eight studies employed a fictitious Internet user who ostensibly disclosed personal information. In 93 articles, self-disclosure was examined within the general population, for example, by using a representative national sample.
Research focused on the individual discloser essentially continued the path forged by scholars of self-disclosure during the twentieth century (e.g. Cozby, 1973; Jourard, 1971; Omarzu, 2000). Since, in the pre-digital era, self-disclosure usually occurred between individuals, past research primarily analyzed one’s perceptions regarding self-disclosure, one’s reasons for information sharing, and one’s understanding of the repercussions (both positive and negative) of this behavior.
Yet, such a narrow focus ignores the implications of self-disclosure for additional parties who are either directly or indirectly involved in these activities. Among the questions that are bypassed as a result are, “How do the stakeholders, who collect, store, and analyze personal information disseminated online, profit from self-disclosure?” “Does society as a whole gain or lose out from the changes in self-disclosure practices, and if either, how?”
Moreover, the primary focus on the individual discloser and on the implications of OSD for that individual implies that it is the individual discloser alone who bears responsibility for acting in this way. This focus diverts the discussion from the reciprocal relations between the various stakeholders (such as peers and institutions) involved in self-disclosure.
The lack of research on the economic aspects of self-disclosure is particularly striking. For example, “What is the value of personal information revealed online?” “How does this value influence the creation of an ecosystem comprised of a set of tools that enable, encourage, feed off, or indeed are based on self-disclosure?” The focus on the individual also obviates the social context of self-disclosure. For example, “How does online self-disclosure affect the creation of values and norms—social, cultural, political, and legal, including the right to privacy or the ‘right to be forgotten’?”
What are the actors’ positions within OSD?
Conceptually, self-disclosure is anchored in two domains: performance (self-disclosure behaviors) and cognition (perceptions concerning self-disclosure). Accordingly, the articles in the corpus address the positions of the actors involved in OSD. Three main categories to that effect were defined a priori, and a fourth was added during coding. Each of the 309 articles in the corpus was assigned to one of these four categories. The first (26 articles) comprised articles focusing on perceptions regarding self-disclosure activities and the factors influencing these perceptions (e.g. Djafarova and Trofimenko, 2017; Forest and Wood, 2012; Levitt and Ippolito, 2014; Saltes, 2013). Research participants had reported their perceptions in in-depth interviews (Levitt and Ippolito, 2014), focus groups (Bryce and Fraser, 2014), or structured questionnaires (DeAndrea and Walther, 2011). One article examined attitudes regarding self-disclosure, based on both self-reports (questionnaires) and behaviors (Forest and Wood, 2012). One article examined journalists’ perceptions of self-disclosure and its implications using quantitative content analysis of reports on the topic of self-disclosure published in major Israeli newspapers (Ashuri and Halperin, 2017).
The second category was of articles exploring actual self-disclosure behaviors and the factors influencing them (97 articles). The self-disclosure activities studied encompassed a broad range of manifestations, including publication of photographs, clips, posts, tweets, replies, notices, and search queries, on a range of tools and platforms (both organic and ones created for research purposes). (a) In 42 of the 97 articles, questionnaires were distributed in which respondents were asked to reveal personal information, such as demographic details—including age, sex, educational level, and income level (e.g. Knijnenburg et al., 2013); personality-related details, such as character traits and level of self-confidence (e.g. Mehdizadeh, 2010); or details related to patterns of their self-disclosure, for example, how the respondent understands his or her self-disclosure in different environments (e.g. Perentis et al., 2017). (b) A group of articles relied on previously shared content, for example, by analyzing users’ search queries in search engines, individual profiles on SNS (Xu and Armstrong, 2019), self-presentation on dating sites (Toma and Hancock, 2010), personal blogs (Mazur and Kozarian, 2010), and personal information shared in online arenas (Bokek-Cohen, 2015). (c) A group of articles examined self-disclosure behaviors during and for the sake of the study, such as disclosing personal information in an online questionnaire (Chang et al., 2016), in interactions with the researchers (Schober et al., 2015), in a virtual interface developed for the study (Z. Li et al., 2019), and with a human factor unknown to the participant (Bareket-Bojmel and Shahar, 2011). (d) Two articles analyzed disclosure behaviors using algorithms that model disclosure patterns. Thus, Ahmadinejad and Fong (2014) developed an algorithm to gauge the probability of a successful phishing attack to obtain users’ information on social media. With users’ self-disclosure as a starting point, the study examined the behavior of the algorithm in regard to various types of information. The behavior of algorithms that model various amounts of self-disclosure was also examined by Chen et al. (2014), with the aim of finding a “ideal” self-disclosure strategy for SNS.
The third category (179 articles) included studies that focused on the factors that influence self-reported intentions to disclose personal information about oneself in the future. More specifically, participants were asked to report their opinions, perceptions, or behavioral intentions as concerns for an issue examined. Thus, for example, they testified regarding their perceptions of the nature of online self-disclosure (Knoll and Bronstein, 2014); their opinions regarding what is and what is not proper to disclose (Geusens and Beullens, 2017; Mesch and Beker, 2010); the goals they pursued through self-disclosure (Kwak, 2012; P. Li et al., 2019); their self-disclosure patterns in various online arenas (Christofides et al., 2012; Gibbs et al., 2010); and their intentions regarding future disclosure (Bol et al., 2018; Lee and LaRose, 2011; Van Oosten and Vandenbosch, 2017). The self-reports were obtained primarily using questionnaires, but also through additional means such as in-depth interviews (Shumaker et al., 2017).
The fourth category (seven articles) included studies that discussed the efficiency of self-disclosure and/or of exposure to other people’s self-disclosure in a variety of contexts. The topics of these articles did not clearly fit into any one of the three major categories outlined earlier.
From the aforementioned discussion, it is apparent that the focus in the studies reviewed is on the role of individuals in OSD and the ramifications of their acts for themselves. As already elaborated, this research focus diverts the discussion from the reciprocal relations between the various actors involved in self-disclosure and those affected by it, including those who are exposed either directly or indirectly to others’ personal information online. Furthermore, an approach that highlights the role of individuals (their perceptions and actions) downplays the broad social and economic aspects of the phenomenon, including the role society at large plays in OSD (reflected, e.g., in social norms) or the roles of institutions therein (manifested, e.g., in the collection, analysis, and usage of individual data).
Theoretical underpinnings of OSD
The theories utilized in the studies on OSD reviewed differed widely, but several prominent trajectories emerged (See Figure 2—distribution by theories employed).
Distribution of articles by theories employed.
First, as Figure 2 indicates, many studies relied on theories associated with what is known as “identity management.” These studies conceived of self-disclosure as a practice that facilitates the formation of self-identity, as it enables the individual to determine the information to be disclosed in any given context. This preoccupation with the individual discloser continued the path forged by researchers of self-disclosure during the twentieth century (e.g. Cozby, 1973; Jourard, 1971). Inasmuch as, in the pre-digital era, self-disclosure usually occurred between individuals, past research tended to examine the individual and his or her interactions with other individuals. In the digital era, however, such an emphasis is unwarranted, as in online settings, information disclosed by individuals is often accessible to peers, to society at large, and to institutions, both commercial and public (see, e.g. Zuboff, 2019).
Second, most of the theories identified put the spotlight on the cost-benefit evaluation by the individuals who disclose personal information online (an example is the “privacy calculus” framework). These theories set the individual at the center and downplay or even ignore the broad social and economic contexts of the phenomenon (e.g. Bol et al., 2018; Dinev and Hart, 2006). Moreover, the schools of thought that embrace cost-benefit approaches take self-disclosure to be a rational process whereby individuals who share information about themselves figure as players aware of the advantages and dangers of this practice (whether in respect of identity management, economic considerations, political engagement, and so on), and the acts of information sharing as such derive from this awareness. Cost-benefit theories overlook the fact that self-disclosure in online settings is not always rational, and in particular, that in many cases, it occurs without the sharer’s knowledge, with him or her leaving behind digital footprints about which they are unaware. Approaches that focus on the individual also ignore the power wielded by such entities as corporations and state arms, which, in multiple and varied ways (including manipulation), encourage self-disclosure; yet, they ensure that the individuals who disclose personal information are unaware of doing so or of the results of their actions (Draper and Turow, 2019).
To sum up, the theories pre-digital researchers adopted in their analyses of OSD align with their preoccupation with the individual, and studies nowadays continue to rely on these frameworks notwithstanding the technological transformation that the world has since undergone. It is not surprising, therefore, that the topics prevalent in the corpus investigated in the current study differ but little from those addressed in the past, be it the motives propelling one to reveal personal information about oneself, the type of information one shares, and outcomes of the self-disclosure behavior for the discloser. Adopting theories that are less individual-centric could have shed light on the effects of self-disclosure on additional stakeholders directly or indirectly involved in this kind of activity: on people exposed to disclosure, on various institutions (commercial and public) which develop and make use of tools established upon personal information shared online, and on society at large, which may both gain and lose out by the resulting change in disclosure norms, for example, the blurring of the boundaries between public and private arenas.
The media utilized for self-disclosure
A crucial component of the model we propose in this review (See Figure 1, An Interdisciplinary Online Self-Disclosure Model) is the media in which and by which personal information is disclosed and disseminated. Networked participatory media, which enable and encourage self-disclosure and are based on personal information, influence how disclosure is carried out, its patterns, and repercussions. Therefore, the articles in the corpus were examined with reference to the media used for self-disclosure. In the cases where an article examined several types of media, each of these was counted separately.
The largest category (192 observations) comprised studies about self-disclosure on online SNS, both identified and unidentified. As observed in earlier literature reviews on self-disclosure on these platforms (Abramova et al., 2017; Luo and Hancock, 2020), online SNS have changed the quantity and quality of self-disclosure, as well as its character and repercussions—both positive and negative. The current review also revealed that studies into self-disclosure on SNS focused on the relationship between individual disclosers and disclosees (mostly peers). Yet, nowadays, those exposed to the information thus revealed are not only individuals but also institutions that marshal online platforms, tools, and devices and share such personal information with second and third parties with which they partner. Therefore, future studies need to address the sharing of information with these entities.
The second largest category included studies regarding self-disclosure on websites (52 observations). This result is consistent with the rapid development, in both quality and quantity, of websites that enable people to create and share personal information using a wide range of features and disseminate it effectively and quickly to individuals and groups.
A surprising finding is the small number of studies on self-disclosure using smartphones (six articles), wearables such as clothes, wrist fitness trackers, and smart watches (three articles), and using or in the presence of bots or virtual assistants (five articles).
Of the 309 articles in the corpus, only 25 related to self-disclosure in two types of media, and only three articles in three types thereof. The articles that examined self-disclosure in different media yielded several insights. Of especial note are differences found in how the users understand and think about various types of media. One such issue is how much a particular platform protects the information revealed on it. Another is what kind of personal information is acceptable to expose on the platform. The type of media was also found to influence the way in which users actually reveal information about themselves or are prepared to do so– for example, the extent of honesty and authenticity of the information they choose to share or the level of intimacy they are willing to maintain.
It was also found that only 14 articles compared self-disclosure in online and offline settings. This is surprising as past research showed that the arena (virtual or non-virtual) has a significant effect on motivation for self-disclosure, on the nature and quality of disclosed information, and on the effects of self-disclosure on both the discloser and disclosee (See Discussion in Ruppel, 2015).
The findings presented earlier regarding the differences in self-disclosure online and offline and via different media point to a need for more extensive comparative research in these areas. Among the topics that can be investigated in such comparative frameworks are self-disclosure in various media (e.g. in online SNS versus revealing information to voice assistants), self-disclosure on similar media operated by different companies (e.g. SNS operated by Facebook and Twitter), or self-disclosure on media designed for different purposes (e.g. apps for professional needs and dating as opposed to navigation or consumer apps).
The review also revealed a sharp rise in the number of relevant articles during 2013–2015 and a drop as of 2016 (See Figure 3—distribution of articles by year of publication)
Distribution of articles by year of publication.
The sharp rise in the number of studies in 2013 can be explained by the popularity of SNS and an attendant increase in the overall number of disclosers and disclosees. In addition, as they gained popularity, these platforms also became more sophisticated and gradually added an increasing variety of features through which personal information of different quality and quantity could easily be disclosed and disseminated broadly and effectively among network members.
As stated earlier, the rise in the number of studies about self-disclosure between 2013 and 2015 was followed by a drop, from 2016 until 2020. This finding is surprising in light of the steadily growing popularity of SNS and, in particular, the rapid development of mobile technologies, apps, wearables, voice recognition technologies, and other media based on self-disclosed data and information.
The nature of information disclosed
Past research has shown that decisions about disclosing information are strongly influenced by considerations of anonymity (See, e.g. Cozby, 1973). Therefore, a core component of the model (Figure 1) relates to the characteristics of the information disclosed online, like the extent to which such personal information can be linked to a specific identifiable individual. We found that most of the articles (204 out of 309) looked at identified information, meaning information whose publisher could be identified and was regarded as the protagonist of the disclosure narrative. Of the remaining articles, 95 discussed both identified and unidentified information, and only 10 were about anonymous information.
In the Web 1.0 era, information shared by users was primarily anonymous. Correspondingly, the initial studies that investigated online self-disclosure focused on this anonymity and demonstrated that anonymity encourages self-disclosure. Based on past research regarding self-disclosure in offline settings, these studies explained that online anonymity reduces the sense of danger emanating from self-disclosure and contributes to the “stranger-on-the-train” phenomenon, whereby people will reveal large amounts of personal information to other people whom they do not know, since they assume they will never meet the latter again (Hollenbaugh and Everett, 2013). In contrast, over the past decade, with the development of participatory media (most notably SNS), information revealed online has become identified with the person producing it. This can possibly explain why most articles reviewed focused on disclosure of identified information.
Since self-disclosure involves vastly different types of information (e.g. emotional or factual, sensitive or not), the type of information discussed in the corpus articles was also examined. Almost one-third of the articles (95 out of 309) did not provide details regarding the information content addressed. However, the majority of the researchers distinguished between the various types of information being disclosed, and therefore, the analysis included different kinds of information, like medical information, financial information, information regarding location, and many more (see Figure 4—distribution by nature of personal of information disclosed).
Distribution by nature of personal of information disclosed.
As Figure 4 shows, only to a limited extent did the articles in the corpus examine disclosed information about specific or context-related topics. Overall, 66 references were found to information linked to medical issues or emotional or mental health, and 19 references to information on social relationships. The predominance of studies on personal information of general nature (as opposed to information in a specific terrain, such as medicine) can be explained by researchers’ propensity to study self-disclosure in SNS, as opposed to designated online platforms in specific fields.
Only 69 articles discussed visual content. Following the development and increased presence of visual content and its use for self-disclosure (such as self-disclosure using photographs and clips, or self-presentation using emojis and stickers), as well as the increasing popularity of platforms based on visual content (such as Instagram and TikTok), we anticipated that more attention would be accorded to exposure to content of this kind.
Factors affecting self-disclosure
Self-disclosure practices are affected by a variety of factors, and therefore, this aspect is central to our model (See Table 1—factors affecting self-disclosure).
Factors affecting self-disclosure.
As mentioned in the beginning of the empirical section, the corpus was divided into three main categories, comprising articles devoted to perceptions regarding OSD, OSD behaviors, and perceived OSD behaviors (henceforth, Categories 1, 2, and 3), respectively. The results that emerged regarding the factors influential in these categories were heterogeneous, such that a factor influential in one might have no effect in another, or a factor that had a sweeping effect in one influenced another only sporadically. Neither within the categories were the effects of the factors found consistent, such that a factor identified as influential in some of the articles was found to produce no measurable effects in others, or indeed to have exerted opposite effects altogether.
The number of influential factors identified was too large for the salient analysis to encompass the entire corpus. Accordingly, only factors studied in over 20% of the articles in at least one category were taken into account. Furthermore, as some of the articles tested more than one factor, the number of factors exceeds that of the articles examined.
Given the overall focus on the individual discloser, it came as no surprise that most studies in the corpus examined factors related to the characteristics of these actors. The most frequently examined influential factor in all three categories was the respondents’ sex (or gender)—in 35% of Category 2 and in 53% of Category 3. The respondents’ age was the influential factor in 31% of the articles in Category 1, 30% in Category 2, and 50% in Category 3. However, within each of the three categories, different factors emerged as influential. In Category 1, participants’ ages were found to have no effect on disclosure perceptions in six of the eight articles where they were examined, as opposed to two articles where they were found to matter. Participants’ sex/gender was examined in nine articles and found not to affect disclosure perceptions in six of these but to have such effect in three.
The effects of personality and character traits were examined and found to be influential in 41% of the articles in Category 3, 13% in Category 2, and 15% in Category 1.
The nature and characteristics of the disclosed information were examined and found to be influential in 31% of the articles in Category 1, 20% in Category 2, but only 11% in Category 3.
The component of disclosure context (e.g. the medium where the disclosure took place, the environment—online or offline, the degree of closeness to the discloser, and trust in the receiver of the disclosed information) was examined as the influential factor in Category 1 in 31% of the articles, in Category 2 in 16%, and in Category 3 in only 11%.
The length of time the tool/platform, through which the disclosure occurred, was in use was examined as the influential factor in 28% of the articles in Category 3, in 6% in Category 2, and in 4% in Category 1.
The size of the social network was examined in terms of its influence on OSD in seven articles, of which five found it to be influential, while two did not.
The cultural factor was examined in 17 articles and found not to affect perceived disclosure behaviors in five of them, but to do so in 12.
There was a lack of uniformity in the factors linked to user proficiency as well as the time spent utilizing a specific technology for self-disclosure. The time factor was found to influence perceived disclosure behaviors and disclosure intentions in 31 articles, but not in 19 others, and user proficiency (in the sense of mastery of a certain technological wherewithal) was found to affect perceived disclosure behaviors and disclosure intentions in 23 articles, but not in five others.
Expected effects of OSD were tested. Expected benefits derived from self-disclosure were examined in 50% of the articles in Category 3, in 21% in Category 2, and in 23% in Category 1 and were found influential across the board. Expected losses or concerns regarding the repercussions of self-disclosure were examined in 39% of the articles in Category 3 and found to be of influence in 92% of these. In contrast, the effect of this factor was examined only in 7% of the articles in Category 2 (targeting actual self-disclosure behaviors) and was confirmed in 71% of these. In Category 1, concerns were examined in 15% of the articles and found to be influential in all of them.
Related to concerns is the trust factor, in the sense of the discloser’s trust in both the disclosee and technology. Of the four articles that examined the discloser’s trust in the disclosee(s), two found this factor influential.
The corpus articles address a large number of factors (mentioned earlier), which could affect OSD (both perceptions and behaviors), but in most cases, the spotlight is on the individual who discloses identifiable or personal information online. This is evident in the foregrounding of factors related to the characteristics of the individual discloser (personality traits, gender, age, and so on) and to the context in which such an individual acts (size and density of the social network and so on). The factors tested also pertain to the nature of the connection between the individual disclosers and the disclosees (their relationship, the quality of the interaction between them, the similarity and differences between them, and so on). However, most studies concentrate on disclosure in peer communities.
A significant component of the self-disclosure phenomenon, and one that is not sufficiently addressed, is the direct or indirect involvement therein of institutions—large or small, national or global, public or commercial. Such entities encourage disclosure, receive the information shared, and in many cases, are based on the disclosure activities and the resulting personal information. Glaringly absent from the corpus are factors that govern the creation of self-disclosure tools and platforms by institutions and those that affect the array of uses to which the latter put the information gleaned thereby.
Outcomes of self-disclosure: benefits and dangers
A prominent theme that arose from the review is the positive and negative implications of self-disclosure. This issue was also salient in two recently published literature reviews (Abramova et al., 2017; Luo and Hancock, 2020), both of which limit the discussion to self-disclosure on SNS. The term that is commonly used in the literature to refer to this issue is “risk.” We believe the meaning of this term to be excessively broad for the case at hand, as “risk” connotes with uncertainty and thus implies doubt regarding implications of a certain course of action, misgivings about deviation from the expected outcome, and more specifically, from expected earnings (See Discussion in). Therefore, we opted for the terms with a negative denotation: “losses” and “dangers.”
Of the 309 corpus articles, 278 touched upon the topic of the advantages (benefits) and disadvantages (dangers) of online self-disclosure. These articles fall into three distinct groups based on the aspect of self-disclosure discussed: dangers, advantages, and both of these. The dangers of self-disclosure were discussed in 67 articles, as the main theme in 22 of them. The advantages of self-disclosure were addressed in 148 articles, as the main theme in 51 of them. Both the advantages and disadvantages of self-disclosure were discussed in 63 articles. The theme of anticipated dangers and attendant disadvantages of OSD was salient in 130 articles, and that of its advantages and perceived benefits in 211 articles.
Dangers associated with self-disclosure
Four primary categories were identified regarding causes of dangers: (1) technology (such as a lack of option to adjust privacy policies); (2) relations between individuals and tech companies (e.g. the use of personal information by companies which own SNS); (3) commercial bodies (other than the tech companies) which use personal information shared online for a variety of purposes; employers; family and friends; and other active online users; and (4) other individuals.
Almost 80% of the articles that touched upon the possible hazards of OSD highlighted the danger to the individual discloser. Ten percent of the articles were devoted to the danger to the individual disclosee, and another 10% discussed dangers with reference to a specific population sector. Less than 1% of the articles discussed society as a whole and the dangers facing it as a result of online self-disclosure.
Advantages associated with self-disclosure
Motivations for and perceived advantages of self-disclosure were discussed in 211 articles. Based on this subset, the contexts in which OSD can carry advantages fall into several categories, which align with those identified in previous surveys that discussed self-disclosure in virtual spaces (Luo and Hancock, 2020).
The category addressed in the largest number of articles (N = 27) centers on the social context, that is, relationships between the interviewee and other Internet users and the purpose of self-disclosure, for example, self-presentation, strengthening self-image, and improving social connections. The articles in the second-largest category (N = 24) relate to the emotional context. Another set of articles (N = 11) discussed advantages of OSD in the professional and educational contexts, for example, maintaining a professional image at work or improving the effectiveness of learning by providing feedback regarding the material covered. Two articles discussed advantages in a political context, specifically, online self-disclosure by politicians.
The factors that render OSD advantageous fall into four main categories: (1) other users’ activities; (2) relations with commercial bodies; (3) employers; and (4) technologies, including the companies that develop them.
Of the articles discussing the beneficiaries of OSD, 80% focused on disclosers themselves, 13% on society as a whole, and 7% on a particular population sector. Only one article addressed the potential advantages to the individual disclosee, that is, benefits derived by one exposed to self-disclosure by others.
Discussion and a framework for studying OSD
Our results, presented in the Interdisciplinary Online Self-Disclosure Model (See Figure 1), point to a preponderance of individual-centric approach, which de-emphasizes or indeed overlooks other forces involved, like the role of public and commercial institutions in OSD. This tendency is a holdover from the pre-digital age and is fed by the literature of that period that addressed mainly self-disclosure between individuals via direct interpersonal communication, that is, with no technological mediation.
A major finding relates to the role the media play in self-disclosure perceptions and practices (titled in Model 1 “The media utilized for self-disclosure”).
Research about offline self-disclosure dealt primarily with face-to-face interactions between two people—one “actor” is the sender (discloser) and one “actor” is the recipient (disclosee), who in most cases share the information willingly and are cognizant of each other’s existence and of the interaction taking place between them (See, e.g. Cozby, 1973). In other words, self-disclosure in the pre-digital age generally occurred in a distinct context of which the discloser and the disclosee were both aware. In today’s digital world, however, many people disclose a great deal of information about themselves—knowingly or unknowingly, intentionally or unintentionally—to individuals and institutions unknown to them. Importantly, these relationships are often characterized by asymmetry of information and power, with the recipient (peers or institutions) usually wielding more power: Institutions on the receiving end, for example, collect the data, analyze it, and use it for their profit, in ways of which the discloser is often unaware or approves. In these new circumstances, the scholarly focus on the individual discloser who is cognizant of the nature and consequences of his or her behavior limits the understanding of self-disclosure.
Yet another token of past scholarship is the focus on the individual discloser as the “actor” who either gains or loses by disclosing his or her personal information. In the analog age, such an emphasis was a matter of course, as individuals usually disclosed information about themselves to other individuals. However, in the digital age, when self-disclosure by individuals (and disclosure by individuals about others) has a direct or indirect effect on many other individuals, on society at large, and many kinds of institutions, it is important to take count of these various stakeholders as well.
As this review has shown, OSD research focuses on the individual discloser, who is assumed to bear responsibility for the disclosure. This approach downplays power issues involved in OSD. As amply demonstrated in datafication studies, even when individuals incur losses as a result of OSD (like privacy violation or shaming), they cannot always assess or control the outcomes of their conduct because harms (like data breaches) often take place at the collective level (e.g. Draper and Turow, 2019; Zuboff, 2019). For example, with full endorsement on the part of social norms and regulations, platform companies reserve the right to collect personal information, sell it to second and third parties, and change their terms of service at any time (See, e.g. Masur, 2018).
The overwhelming focus on the individual in the corpus investigated could stem from the specificity of the disciplines within which most of the research works were carried out, notably, psychology and media and communication. It is by scholars in these fields that the issue of self-disclosure was addressed in the pre-digital age as well—a circumstance that may account for the continuity in the tenor of the investigations (See Figure 5).
Distributions of articles by academic disciplines.
Given that online self-disclosure has an impact not only on the individual but also on the institutions that receive the information thus revealed and on society at large, which is involved in OSD either directly or indirectly, future studies would do well to employ theories developed in other disciplines, such as economics and law, which focus on other social elements that affect and are affected by self-disclosure, be it regulatory, economic, cultural, or political.
A critique can be leveled also at the research population which dominates the works on OSD included in the corpus. Most of the research relied on student convenience samples (N = 114). Yet, online self-disclosure occurs today among population sectors that are vastly different from students in age, education, socioeconomic circumstances, and other characteristics. Moreover, as data scholars have demonstrated on many occasions, risks entailed by self-disclosure are not uniform: The poor, especially poor people of color, are more likely to suffer loss of employment or reputation, as well as financial harms, due to disclosure of personal data and information. Future studies could shed light on the effects of self-disclosure on individual OSD actors from diverse social groups and on additional stakeholders directly or indirectly involved in this kind of activity.
Capitalizing on the findings presented earlier to advance studies into OSD, we have developed a dialectical framework that underscores local and global structures that influence self-disclosure practices and affect the dynamic relationships between the different agents involved therein. The proposed dialectical framework, which we term “a practice lens to study OSD,” integrates two research streams. One is the body of research on OSD, which focuses on the individual discloser. The second is the expanding body of critical research on datafication, which highlights the growing asymmetry of power and information between individuals and institutions that develop tools for self-disclosure and make use of personal information shared online (e.g. Couldry and Mejias, 2018; Kennedy, 2018; Turow, 2021; Zuboff, 2019).
The suggested framework is premised on the notion that perceptions of individual disclosers regarding self-disclosure and their OSD behaviors are affected by actions of peers and institutions, for example, the tools the latter develop, and their treatment of personal information. At the same time, practices of institutions are affected by individuals’ actions, including their choice of technologies for self-disclosure, the quality and quantity of the personal information they share, and more. Importantly, actions of both individuals and institutions are influenced by (local and global) collective structures including market structures, norms, regulations, existing infrastructures, and so on. By encompassing the dialectical relationships between people and institutions, and between agents (individuals and institutions) and structures, the integrated framework enriches our understanding of OSD.
Furthermore, the dialectical framework we propose here could enhance OSD research as it underscores structural elements like diversity and inequality. As demonstrated, in theorizing OSD, most studies rely on empirical evidence regarding self-disclosure attitudes and behaviors of Western-based, educated, upper-middle-class demographics. Yet, research on datafication has repeatedly shown that self-disclosure risks are unequally distributed among the population.
To account for the dialectical relationships between different agents that are involved in self-disclosure and that operate at both the individual and the collective level, we embrace the structuration perspective formulated by Giddens (1984). Giddens’ pioneering structuration theory recognizes that human actions are enabled and constrained by structures and that these structures are the result of previous ongoing actions. In Giddens’s model, structural properties, which he called “modalities,” encompass facilities (like technologies), rules (like norms), and interpretative schemes (perceptions) that human agents use in their habitual interactions. According to Giddens, these modalities mediate human action, while at the same time being reaffirmed through their use by human agents.
More recently, Wanda Orlikowski (2000) extended Giddens’s original structurational model into “a practice lens for studying technology,” which reveals “[h]ow people, as they interact with a technology in their ongoing practices, enact structures which shape their emergent and situated use of that technology” (p. 404). Based on Orlikowski’s schema, we constructed a practice lens for studying practices associated with OSD, employed by both institutions and individuals (See Figure 6—practice lens for studying OSD). In the remainder of this article, we briefly discuss it.
Practice lens for studying OSD.
A practice lens for studying OSD
In our framework, institutions are agents that facilitate (and often encourage) self-disclosure behaviors; they collect the data/information disclosed, disseminate it (sell or give it away), analyze it, and use it for a variety of purposes. These activities are constituted through the modalities institutions enact. One of these, termed here as “OSD facilities,” relates to resources deployed by institutions, including the tools and devices (like SNS, wearables, websites) they develop and utilize to enable individuals to disclose data and information about themselves. Another such modality is OSD norms, that is, laws, regulations, and so on that support and limit institutional practices, for example, leveraging the right to sell personal information to second and third parties. Yet another such modality is interpretative scheme, which relates to institutions’ perceptions regarding possible outcomes of practices associated with OSD. An example is apprehensions of losing consumers due to violation of privacy, and as a consequence, the construction of privacy settings that empower consumers.
The three modalities discussed earlier are also implemented by individuals in their ongoing OSD practices. For example, an individual enacts the facilities modality by deciding which wearables, platforms, apps, and so on to use for disclosing personal data and information. The norms modality is enacted when an individual accepts or rejects the terms and conditions of a certain privacy policy. Finally, the interpretative scheme modality is enacted when an individual discloses certain information (e.g. related to health) in a particular site thinking that this action will yield positive outcomes (like getting emotional support or required medical information).
In Giddens’s conceptualization, human agents (in this case, institutions and individuals) enact “structural properties of social systems,” which shape their practices. In the context of OSD, this may involve directing people’s behaviors on the basis of the data/information they knowingly and unknowingly self-disclose—a strategy that Yeung (2017) calls “hyper-nudging.” Such practices may also include surveilling and thereby exploiting disclosers by commodifying data and information about them (e.g. Couldry and Mejias, 2018; Zuboff, 2019). Another structural property is “personalization,” in the sense of providing people with personalized content. Additional instances of putting the “structural properties of social systems” to one’s advantage are displayed in Figure 6.
The practice lens for studying OSD offers a practical basis to address new challenges of the digital age. In the following section, we briefly provide two scenarios to demonstrate how the model can be applied.
In one scenario, a researcher studies the nature and effects of an emerging trend in healthcare called “data-driven personalization in healthcare” (DDPH). DDPH makes use of data generated by computer-mediated technologies to better understand the characteristics of an individual and deliver the “right” care to the “right” person at the “right” time. The operation of DDPH is grounded in two perspectives: of the individuals and of the healthcare institutions. For the former, personalized healthcare means being recognized as a unique individual with a unique health history, lifestyle, and circumstances, as well as receiving relevant content and services and achieving better health outcomes. For an institution, it means collecting big individual data, analyzing these, and using the insights obtained thereby to tailor products and services to the specific needs of the individual. DDPH thus provides a clear manifestation of the dialectical relationship between the two agents (individuals and heath institutions), which constitutes a structural property, that is, personalization of healthcare.
To gain a better understanding of the nature and effects of DDPH, a researcher may wish to explore perceptions of individuals and healthcare practitioners (operating in health institutions) regarding the risks and opportunities DDPH practices entail. Our model could be extremely useful for addressing issues of this kind. By observing the facilities modality, the researcher will be able to identify participants’ ideas about the tools that are currently employed for harvesting the data individuals disclose online. A focus on the norms modality could expose agents’ practices and show, for example, how practitioners in healthcare use (and plan to use) the personal data they harvest. An examination of participants’ interpretative schemes could reveal their evaluation of personalization practices, now and in the future. Do individuals and healthcare practitioners believe, for example, that personalization gains legitimize the collection of personal data individuals disclose in many contexts, including that outside healthcare?
A policy maker interested in this topic would also benefit from the model offered. Observing the so-called data-driven personalization practices through our practice lens framework, policy makers would be able to evaluate participants’ beliefs regarding the risks and opportunities DDPH practices entail and, on the basis of this, take socio-technological actions, like develop privacy-enhancing technologies (PETs) and establish rules that could minimize misuse of personal information by healthcare institutions, for example, acquire data and information about specific individuals and, on this basis, advise their addressees on the norms of healthy physicality and lifestyle and compel individuals that do not comply with these rules to fall in step.
Given that more and more individuals disclose personal information about themselves using a range of technologies and that an ever-increasing number of institutions stand to gain great social, political, and economic power by developing self-disclosure media and by making varied use of the data and information thus revealed, future studies on OSD need to account for the dialectical nature of these practices and of the role they play in the construction of the different “structural properties” (Giddens, 1984).
Conclusions
The dialectical model we have developed in this review by aggregating 10 years of research on OSD points to three important consequences and practical implications of this phenomenon. First, the view whereby OSD influences and is influenced by an interplay between agents and structures implies that such practices are politicized. Such a perspective acknowledges the issues of power involved in OSD and therefore also the need for appropriate socio-political measures to contain this phenomenon. Steps to this end need to be informed by supra-national regulation models that go beyond local individual-centric laws like the GDPR and that take into account the structural components that derive from big-data collection from individuals and the analysis of such information by both public and commercial entities. Second, our findings show that research on OSD overwhelmingly focuses on the individual. However, the individual discloser alone cannot be solely responsible for his or her protection, but rather, the burden must also lie on product developers, service providers, and those in charge of regulations, which must ensure transparency of data-driven institutions’ practices and constrain, if needed, harmful conduct. Third, insofar as dangers associated with OSD are unevenly distributed among the population, socio-technological actions are clearly warranted, including the development of PETs and regulations that could empower disadvantaged or underprivileged groups, thus minimizing existing power asymmetries among peers and between individuals and institutions.
With the sharp resurgence of scholarly interest in OSD following the advent of new network technologies, an urgent need arose to reconsider the research agenda regarding this phenomenon. Our review of the work carried out over 10 years which addresses the various aspects of these practices has yielded insights, conclusions, and recommendations that will hopefully help produce a more cohesive body of literature and thus help chart out actionable steps for individuals, managers, and policy makers.
Footnotes
Acknowledgements
The authors would like to thank Yarden Sher and Keren Westreich for their important contribution.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
