Online Political Participation: Evolution of a Concept PDF

INFORMATION, COMMUNICATION & SOCIETY https://doi.org/10.1080/1369118X.2021.2013919 REVIEW ARTICLE Online political participation: the evolution of a concept Christina Ruess a, Christian Pieter Hoffmanna, Shelley Boulianne b...

INFORMATION, COMMUNICATION & SOCIETY https://doi.org/10.1080/1369118X.2021.2013919 REVIEW ARTICLE Online political participation: the evolution of a concept Christina Ruess a, Christian Pieter Hoffmanna, Shelley Boulianne b and Katharina Hegerc a Institute of Communication and Media Studies, University of Leipzig, Leipzig, Germany; bDepartment of Sociology, MacEwan University, Edmonton, Canada; cWeizenbaum Institute for the Networked Society, Freie Universität Berlin, Germany ABSTRACT ARTICLE HISTORY The advent of online technologies has been triggering a wave of Received 3 August 2020 empirical examinations of online political participation (OPP) over Accepted 14 November 2021 the past twenty years. It also stimulated scholarly debate on how KEYWORDS to conceptualize political participation in a digital age. Scholars Online political participation; differ on whether to consider passive and expressive online digital media; systematic behaviors part of or a mere precursor to political participation. literature review; social This study argues that due to its rapid evolution as well as its media dependence on platform affordances, quantitative empirical studies on OPP may be prone to deviations between established, much-cited definitions and measurements applied in the field. Based on a systematic literature review of 289 international peer- reviewed survey-based and experimental studies, we analyze both definitions and measurements of OPP. We find a series of disconnections: Measures preponderantly address online activities, yet merely a small share of definitions focuses on the online sphere. While only few definitions account for passive activities (e.g., reading news about politics), many operationalizations include measures capturing such passive behaviors. Expressive activities are most popular in measures of OPP, but definitions rarely reflect this focus. Finally, while measures of OPP are prone to be platform-specific, definitions tend to neglect this characteristic. We conclude by reflecting the conceptual implications of common measurement practices for the study of OPP. Introduction Political participation is an ever-changing concept (Theocharis & van Deth, 2018b), garnering much attention both in research and practice. Since the middle of the twentieth century, definitions of political participation have moved beyond electoral participation (i.e., casting a vote and campaign activities) and have instead been extended to uncon- ventional political acts (e.g., protesting), and even non-political activities (e.g., volunteer- ing for a social cause) (Theocharis & van Deth, 2018a). This evolving understanding highlights new forms of political participation that are not tied to election cycles. CONTACT Christina Ruess [email protected] Institute of Communication and Media Studies, University of Leipzig, Nikolaistrasse 27-29, 04109 Leipzig, Germany Supplemental data for this article can be accessed at https://doi.org/10.1080/1369118X.2021.2013919. © 2021 Informa UK Limited, trading as Taylor & Francis Group 2 C. RUESS ET AL. The Internet has further extended the scope of political participation by facilitating new forms of mediated political action (Harris & Gillion, 2010; Kim & Hoewe, 2020). Social media, especially, afford new forms of political engagement – many of which tend to require little time, effort, and financial resources compared to established acts of formal participation. Against the backdrop of an observed decline in formal political participation in some Western democracies (e.g., Hooghe & Kern, 2017; Skocpol et al., 2000), such new online forms of participation raised hopes for a stronger engagement of politically under-represented demographics like women, ethnic minorities, youth, less educated, and low-income individuals (Correa & Jeong, 2010; Hirzalla & van Zoo- nen, 2010; Krueger, 2002; Sasaki, 2017). Critics, however, dismiss new, digital forms of participation as merely expressive or passive (e.g., Morozov, 2009) – and therefore as unworthy to be considered valid acts of political participation. Accordingly, the continual expansion of the notion of political participation in the digital age is accompanied by a fierce debate about its proper definition and delineation. Positions range from appeals to apply narrow conceptualizations (Verba & Nie, 1972) in order to avoid a ‘theory of everything’ (van Deth, 2001, p. 1) to calls for much broader definitions so that contemporary forms of engagement are not ignored and the richness of the phenomenon is not diminished (Ekman & Amnå, 2012; Norris, 2002; Oser, 2017; Pickard, 2019). While a range of publications have engaged in the conceptual debate of how to define and delineate political participation on the Internet (Theocharis & van Deth, 2018a), the present study contributes to this ongoing discussion by shining a light on how empirical research on online political participation (OPP) has operationa- lized and measured OPP. To date, there is a rich and dynamic body of empirical research on OPP, comprising hundreds of quantitative studies that apply dozens of measures, each implying a specific understanding of the concept. We argue that, first, examining empirical analyses of pol- itical participation sheds light on how the field is evolving and which conceptualizations of political participation are embraced in practice. Second, while in theory, empirical examinations should apply consistent definitions and measures of political participation, the quickly evolving social media landscape may in practice engender pragmatic, ad-hoc measurement decisions which could result in deviations between widely accepted definitions developed before the Internet, on the one hand, and commonly applied measures, on the other hand. Thereby, the measures applied in the field may in fact indi- cate an evolving understanding of OPP that can inform future conceptual deliberations in the field. In short, this study aims to contribute to the important discourse on the conceptual- ization of political participation in a digital age by examining and comparing both the definitions and measurements in quantitative empirical examinations of OPP. We con- duct a systematic literature review of 289 English-language peer-reviewed quantitative survey-based and experimental studies, and collect definitions and measurements applied in these studies. We employ Gibson and Cantijoch’s (2013) differentiated typol- ogy of active and passive modes of online political actions to categorize and contrast definitions with measurements of OPP. This will allow us to highlight not only the evol- ution and shifting focus of OPP measurements, but also emerging discrepancies between definitions and measurements applied in the field, and their conceptual implications. To summarize, this study will address the following research questions:  INFORMATION, COMMUNICATION & SOCIETY 3 RQ1: How is online political participation defined and measured in English-language quan- titative studies? RQ2: To what extent do definitions and measures of online political participation align? The political participation concept Definitions of political participation vary appreciably, even if focusing on contemporary approaches developed in the second half of the twentieth century. Early studies of politi- cal participation narrowly addressed election-related behaviors (Berelson et al., 1954). For Verba and Nie (1972), political participation includes all activities carried out by citi- zens ‘that are more or less directly aimed at influencing the selection of governmental personnel and/or the actions they take’ (p. 2). Milbrath and Goel (1977) define political participation as ‘those actions of private citizens by which they seek to influence or to support government and politics’ (p. 2). These definitions differ insofar as Milbrath and Goel’s (1977) understanding includes actions that are merely supportive or ceremo- nial in nature (like participating in a demonstration) that would be excluded from Verba and Nie’s (1972) understanding. Importantly, both definitions exclude passive behaviors or characteristics, such as the consumption of political news (Newton & van Deth, 2010). The traditional emphasis in political participation studies on active modes of behavior focused on affecting govern- ment decisions, is a key point of contention in the more recent conceptual literature. As noted above, a rich stream of research has since applied the political participation con- cept to unconventional political acts, such as protesting, and even various form of civic engagement that are not obviously political in nature (Theocharis & van Deth, 2018a). This increasing broadening of the scope of the political participation concept has been further accelerated in the digital age (Fox, 2014). As Gibson and Cantijoch (2013) point out, the Internet is ‘widening the repertoire of participatory actions’ (p. 701). Some scholars have put forth the argument that online and offline forms of partici- pation should be considered largely interchangeable. This view is particularly popular among those applying concepts of political participation developed during the pre-Inter- net era, and those skeptical towards the novelty and political quality of digital partici- pation (Schlozman et al., 2010). Since, to date, most countries have not implemented e-voting, however, online and offline political participation clearly differ. In its narrowest sense of choosing political representatives, political participation simply does not exist in the online realm, purely on practical grounds. Conversely, a rising number of studies apply conceptualizations of political participation focusing on activities that are imposs- ible to perform offline and therefore exist uniquely in the digital sphere, such as hashtag activism (Berg, 2017; Earl & Kimport, 2011). Consequently, there is an argument to be made that online and offline political participation should be differentiated conceptually (Gibson & Cantijoch, 2013). Online political participation Few scholars have attempted a consistent conceptualization of the OPP phenomenon (e.g., Theocharis & van Deth, 2018b; van Deth, 2014). The field is strongly driven by empirical 4 C. RUESS ET AL. examinations of the phenomenon – with each successive new online technology or plat- form sparking a slew of qualitative and quantitative analyses of their political uses. As a result, research on OPP is frequently shaped by pragmatic considerations, such as data or platform access (cf. Hargittai, 2021). Unavoidably, it is informed by platform function- alities or affordances. Hosch-Dayican (2014) declares: ‘Political participation is a dynamic concept, and its classical instrumental definitions are just too restrictive in the era of digital communication technologies’ (p. 345). In this light, widely accepted definitions of political participation and measurements commonly applied in empirical examinations of political participation on the Internet may actually deviate. In fact, given the dynamics of the field, such deviations could be particularly likely in studies on OPP. To date, the debate on how to conceptualize political participation in the digital realm is strongly shaped by empirical findings on political online behaviors. One such finding revolves around the fact that the populations engaging in political behavior online and offline differ. Consistently, studies find that younger citizens rely more heavily on online than offline participation (Di Gennaro & Dutton, 2006; Vitak et al., 2010). Second, differ- ent media platforms lend themselves to distinctive uses, and therefore attract differing user bases that may exhibit specific sociodemographic patterns (Brundidge et al., 2013; Vissers et al., 2012). Third, platform functionalities facilitate distinctive behaviors. Online, political behaviors emerge that do not have a direct offline equivalent, such as adorning profile pictures or tagging friends in political posts. Gibson and Cantijoch (2013) point out: ‘[…] among forms of engagement that are typically regarded as more passive – news consumption and expressive actions – there is evidence that the medium matters’ (p. 714). Again and again, discussions of the political participation concept in a digitally mediated context revolve around the question of which level of activity should be required to consider a political act as ‘participation’ (Gibson & Cantijoch, 2013; Theo- charis, 2015; Theocharis & van Deth, 2018b). Some forms of genuine online participation tend to require a minimum of activity – which is why they are often sweepingly dismissed as ineffective with regard to achieving political goals (Morozov, 2009). The term slackti- vism (or clicktivism) is used to refer to those expressive ‘feel-good forms of political par- ticipation’ (Vitak et al., 2010, p. 6). They are said to originate less from genuine political convictions with their purpose being merely to signal one’s support for a campaign, can- didate, or cause (Kristofferson et al., 2013). Halupka (2018) finds that clicktivism lacks legitimacy in part because of a perceived lack of ability to engender change. Christensen (2011) argues that worries of ‘slacktivist participation’ substituting more impactful forms of engagement are exaggerated. Indeed, empirical studies suggest that more passive forms of participation recruit citizens to get involved in politics in the first place (Bou- lianne, 2019; Cantijoch et al., 2016). Some authors attempt to relate online and offline activities by differentiating varying degrees of activeness or passiveness (Bakker & de Vreese, 2011). Based on the notion that participation is ‘concerned with doing politics, rather than with being attentive to poli- tics’ (Verba et al., 1995, p. 39), Hoffman (2012) declares ‘passive attention to politics [to be] more accurately defined as a communication activity’ (p. 218, emphasis in original) which is ‘essentially the dissemination of information between parties’ (p. 220). Similarly, Kim and Hoewe (2020) distinguish five ‘factors’ of political participation that include political talk, social media engagement, and information seeking.  INFORMATION, COMMUNICATION & SOCIETY 5 Gibson and Cantijoch (2013) distinguish acts of online and offline participation into nine categories or modes of action: (Active) Participation, thereby, is composed of (1) voting, (2) party/campaign activities, (3) protest activities, (4) contacting, (5) communal actions, and (6) consumerism, while so-called passive engagement consists of (7) news attention, (8) discussion, and (9) the expressive mode. In this categorization, Gibson and Cantijoch follow a distinction put forth by authors focusing on offline participation, arguing that various online behaviors lack the required level of organization or coordi- nation to be considered ‘participation’ (Schlozman et al., 2010). However, following Krueger (2002), Gibson and Cantijoch argue that online, such distinctions are difficult to uphold as ‘traditionally passive types of political engagement such as attention to news are becoming more active and genuinely participatory’ (p. 712). To summarize, over the previous decades, the concept of political participation has been adapted and extended to reflect societal circumstances. The advent of the Internet has added to this dynamic, as online media facilitate new forms of political action. While scholarly debate on the conceptualization of OPP has been lively – particularly on the distinction of passive and active political behaviors – little is known about its impact on empirical approaches to the phenomenon. Due to the dynamic development of the medium, constant adaptations of measurements of OPP can be expected. Such – fre- quently pragmatic –adaptations, however, may lead to deviations between conceptual foundations and measurements. Moreover, this interdisciplinary research field is shaped by political and communication science as well as sociology, among others. Over time, measurements may substantially deviate from established, much cited, ‘now anachronis- tic’ (Pickard, 2019, S. 73) definitions originating from political science, potentially result- ing in inconsistencies. Such inconsistencies can be problematic for several reasons: They could impede the replicability of findings within the research stream and make compara- tive research difficult. They may result in conflicting results about changes in levels of participation – and about who participates in which forms, why, and under which con- ditions. They may also result in ‘implicit theorizing’ (Leontief, 1937), where empirical measures come with conceptual implications that are not made explicit. This study therefore aims to contribute to the discourse on the conceptualization of OPP by examining the definitions and measurements applied in the field. We submit that the dynamics of this research stream, heavily shaped by empirical phenomena such as new platforms and their functionalities, and the ensuing pragmatic research decisions, facilitate a dynamic development of OPP measures. Further, these dynamics may not always be mirrored by corresponding conceptual developments, resulting in dis- crepancies between definitions and measures. Based on the framework proposed by Gib- son and Cantijoch (2013), our analysis will focus on the incorporation of active and passive political behaviors in definitions and measurements of OPP in the extant body of research, as this constitutes a key point of contention in the conceptual debate. Method Selection of studies In order to identify relevant research on OPP, we conducted a systematic literature review. The search procedure was guided by the search string ‘(online OR digital OR 6 C. RUESS ET AL. Internet OR web OR social media OR Facebook OR Twitter OR YouTube OR Snapchat OR Instagram) AND (political engagement OR civic engagement OR political partici- pation OR civic participation OR activism)’. All in all, the combination of these 15 key- words lead to 50 search strings (e.g., ‘digital political participation’ or ‘YouTube activism’). The search terms allowed us to capture distinct concepts of OPP that are oftentimes considered synonymous while avoiding redundancy. The search was con- ducted in the academic databases Scopus, Web of Science, and EBSCOhost.1 This approach was intended to ensure an exhaustive overview of the research field. The search was conducted from June to August 2018. Seven criteria were used to search for relevant publications in a first step of the search procedure (for an overview, see Appendix C): We restricted the search to (1) peer-reviewed (2) journal articles and conference proceedings (3) in English that have been (4) published until the middle of 2018 and (5) contained a combination of the words specified in the search string. Furthermore, we selected articles containing (6) quantitative data analysis of (7) surveys and experimental studies. In total, this initial search yielded 22,096 search results. In step 2, we scanned titles, abstracts, and keywords of the results to ensure that the papers actually deal with the research topic and to assess whether they might contain quantitative data analyses. After mer- ging the samples of the three databases and removing duplicates (step 3), the remaining 346 publications were carefully examined in step 4 to ensure they actually contained quantitative data analyses and had undergone peer-review. In the final step, the articles’ operationalizations of OPP were scrutinized whereby publications with operationaliza- tions that were merely offline-related or restricted to consumptive forms of engagement were excluded. This procedure of elimination resulted in a substantially reduced number of publi- cations per database: 173 for Scopus, 173 for Web of Science, and 175 for EBSCOhost. When merging the papers of all databases, the final sample consists of 289 relevant pub- lications (see Appendix D). Data used in these publications were predominantly from the United States (45.0%), then from the United Kingdom (8.3%), and from China (6.2%) (see Appendix B). Coding We collected general information concerning the publication of the article (e.g., year, journal) and study design (e.g., years and countries of data collection) of the remaining articles. At the heart of our analysis was the distinction of active versus passive political behaviors. We followed the framework proposed by Gibson and Cantijoch (2013) which distinguished (active) participation and passive engage- ment. For the analysis of definitions, we also distinguished whether definitions refer to ‘the most passive types of engagement’ (Gibson & Cantijoch, 2013, p. 709) corresponding to Gibson and Cantijoch’s mode of news attention and affording mainly psychological involvement, e.g., reading political articles – or if they refer to political behaviors that afford more physical, observable activity on the part of the individual. The examination of OPP measurements utilized in the studies allowed for an even more fine-grained analysis than the definitions, as measurements are frequently  INFORMATION, COMMUNICATION & SOCIETY 7 constituted of a list of specific online behaviors. Therefore, we examined the complexity of measurements in the analysis by coding how OPP is operationalized, e.g., measured as multidimensional construct. This was a necessary first step as single-item measurements are rarely specific enough to be ascribed to a particular mode of OPP. Next, we gathered all items representing political behaviors. If applicable, items were split up into their elements (e.g., the item ‘take part in online discussion in order to express own political views online’ was coded as ‘join online discussion’ and ‘express/spread own political opinion’).2 This initial step of the analysis resulted in a list of 65 items (see Appen- dix A). These items were then coded based on the nine modes of action distinguished by Gibson and Cantijoch (2013, p. 707), which these differentiated into (active) participation and passive engagement (see Appendix A). In accordance with the procedure above, we further coded the activity of measurements to capture the most passive political behaviors (e.g., reading a political blog) as termed by Gibson and Cantijoch (2013). For both definitions and measures we also coded the sphere: whether they explicitly relate to political participation online or offline. For definitions and measures specifi- cally relating to the online sphere, we determined whether there are any references to digital platforms (specificity), e.g., social networking sites, microblogging sites, video portals, blogs, e-mail. Measures and definitions without platform reference were declared as platform-independent, meaning that they refer to activities, e.g., ‘on the Internet’ or ‘online’. The values of Krippendorff’s ranged from α =.966, 95% CI[.897, 1.00] for rather mani- fest items (i.e., platform references in OPP measures) to α =.734, 95% CI[.609,.844] for the coding of specific OPP items in accordance to the item list (see Appendix A). All in all, this indicated an adequate intra-coder reliability for the coding of the publications’ information and parameters regarding definitions and operationalizations. Results Analysis of the body of research In the following section, we present an overview of results of the descriptive analysis. The sample of identified publications consists of 286 journal articles and three conference proceedings. Since several publications refer to more than one empirical study, we base our analyses on a total of 319 studies. Figure 1 illustrates the continuous increase of publications containing measure- ments of OPP and thus the empirical relevance of the topic throughout the last decade. The first publication meeting all of our criteria is by Wellman et al. from 2001; its data collection dates back to 1998. The growing number of studies conducted since 2005 may be explained in terms of the emergence of Web 2.0, which was accompanied by launches of popular social media platforms like Facebook and Twitter (founded in 2004 and in 2006). Many studies in the sample were conducted in 2008 and 2012; since a disproportionate number of studies were conducted in the US (see Appendix B), this is most likely connected to the presidential elections, and the intensive use of social media for campaign activities, most notably by Barack Obama’s campaign. Since 2009, the number of publications analyzing quantitative data on OPP has been rising quite steadily.3 8 C. RUESS ET AL. Figure 1. Years of data collection of single studies (specified in n = 289 of a total of 319 studies) and years of publication (n = 289). Analysis of OPP definitions Research Question #1 relates to how the OPP concept is defined and measured in existing research, with a focus on the distinction on active and passive behaviors. Table 1 presents the results of the deductive analysis of definitions, which allows for an overview of the extent of activeness of definitions and reference to the online environment found in the literature and of their pervasiveness. In total, only 151 publications provide an OPP definition of some sort. We find that 32.4% of publications present a definition exclusively referring to the online sphere, whereas in merely 5.3% of cases, authors opted for traditional definitions of (offline) pol- itical participation, in 31.8% of cases, online and offline participation are considered jointly. The analysis reveals that many definitions of political participation lack information regarding the sphere in which political action is performed (30.5%). Thereby, authors implicitly assume the transferability of the concept of political participation from the offline to the online sphere. Limiting the analysis to definitions that refer to the online con- text, 60.2% mention at least one online platform as the context for political participation – most frequently social networking sites, then e-mail and blogs (specificity). Thereby, if definitions expressly address the Internet as the sphere of political participation, they mostly assume that platforms affect the expression of political behaviors online. Table 1. Parameters of OPP definitions. Parameter Specifications Sphere online (excl.) offline (excl.) online and offline n.s. 32.4% 5.3% 31.8% 30.5% a Specificity platform-specific platform-independent 60.2% 39.8% Activity active PP (excl.) passive PP (excl.) active and passive PP n.s. 74.8% 0% 22.5% 2.7% Note: n = 151; n.s., not specified; excl., exclusively; PP, political participation. a Only definitions referring to the online sphere (n = 98).  INFORMATION, COMMUNICATION & SOCIETY 9 In terms of the activity attributes ascribed to political participation, only 22.5% of definitions explicitly account for passive behaviors (such as reading or being attentive to political news) alongside active ones (such as marching in protests). Conversely, 74.8% of definitions exclusively refer to active political behaviors. Analysis of OPP measurements Many studies rely on lists of political activities, partly borrowed from offline concepts of political participation, and approximate the concept by constructing indices of activities (see Table 2). This may partly be due to the fact that online and offline political partici- pation tend to be handled as a set of interchangeable activities carried out through differ- ent channels (Gibson & Cantijoch, 2013). Thereby, when considering the complexity accounted for in OPP measurements, a high percentage constitute scales created from more or less extensive item lists of participatory acts online (and offline) (79.0%), allow- ing us to more closely examine the activities or behaviors at the center of empirical approaches to OPP. OPP indices in the sample consist on average of six items (SD = 3.5). Despite there being a broad consensus in theory about the multidimensionality of political participation, we found ten cases in which scholars resort to one-item measures when trying to capture OPP. Examples of those holistic measures are ‘use social network- ing site politically’ or ‘participating in politics on the Internet’. In terms of sphere, online and offline forms of engagement are not distinguished from each other in 9.6% of measurements, but instead surveyed collectively. 90.4% of measure- ments instead clearly focus on online media. In the wake of the differentiation of online platforms, researchers have customized operationalizations to fit political activities that are (exclusively) executable on these specific platforms. ‘Liked a political party or candi- date page on Facebook’ or ‘tweet about politics’ could serve as examples for platform- specific items. While 20.8% of the documented measurements in the sample lack refer- ences to any specific platform, 79.2% relate to at least one specific online platform. We plotted the number of references to specific platforms in the collected studies over the years of data collection in Figure 2. While the number of references to political activities via e-mail has decreased since 2009, references to social networking sites and to microblogging services like Twitter have gained popularity throughout the last decade. In recent years, studies examining Table 2. Parameters of OPP measurements. Parameter Specifications Complexity single item(s) scale (one-dimensional) OPP dimensions dimensions, but not excl. OPP 15.2% 79.0% 3.8% 2.0% Sphere online PP (excl.) offline (excl.) online and offline PP n.s. 90.4% - 9.6% 0% Specificity platform-specific platform-independent 79.2% 20.8% Activitya active PP (excl.) passive PP (excl.) active and passive PP n.s. 56.7% 1.2% 40.4% 1.8% Mode participation (excl.) passive engagement (excl.) participation and passive n.s. engagement 12.3% 29.8% 56.1% 1.8% Note: n = 342; n.s., not specified; excl., exclusively; PP, political participation. a Percentages do not add up to 100% due to rounding. 10 C. RUESS ET AL. Figure 2. References to platforms in OPP measurements over time. social media platforms, like social networking sites, blogs, YouTube, or microblogging, constitute a large majority, lending support to the notion that current understandings of OPP are intricately intertwined with the affordances of these participatory platforms. Again, the high number of references to social networking sites and e-mail in 2008 and 2012 can be explained by the many studies elicited by the US presidential elections and the campaigns’ social media strategies; the peaks in 2010 and 2014 coincide with US mid- term elections. As for activity, more than half of measures (56.7%) did not include passive partici- pation. However, active as well as passive political behaviors were used side by side in 40.4% of publications in the sample. In total, the literature analysis captured 342 measurements of OPP. The identified activities addressed by these measurements were coded as modes of participation based on Gibson and Cantijoch’s (2013) framework. Figure 3 now illustrates how the focus on particular modes of participation has changed over time. Modes of ‘passive engagement’ are presented in dotted lines, modes of (active) ‘participation’ in continuous lines. The numbers for all participation modes in the years of 2008, 2010, and 2012 are again marked by peaks that can be ascribed to US elections raising interest in OPP. Strik- ingly, the prevalent mode of participation empirically examined in the documented pub- lications is the expressive one, followed by news attention, whereas consumerism and voting play a subordinate role in OPP operationalizations. Therefore, even if the debate about the legitimacy of those low-threshold, convenient forms of participation persists, the frequent use of expressive activities in operationalizations of OPP demonstrates  INFORMATION, COMMUNICATION & SOCIETY 11 Figure 3. Modes of participation in measurements over time, categorized according to Gibson and Cantijoch (2013). their acknowledgment as salient, possibly even dominant elements of political engage- ment in the digital sphere by current research. Comparing definitions and measurements of OPP To address Research Question #2, we contrasted the measures with the conceptual focus evident in OPP definitions along parameters ascertainable for both (Table 3). To rule out the possibility that different bases may cause deviations between definitions and measures, the following comparisons are based on publications that provide a definition alongside an empirical examination, which equals 185 measures. Table 3. Comparison of OPP definitions and measures. Parameter Specifications Sphere online (excl.) offline (excl.) online and offline n.s. Definitions 32.4% 5.3% 31.8% 30.5% Measures a 86.5% - 13.5% 0% Specificityb platform-specific platform-independent Definitions 60.2% 39.8% Measuresa 78.4% 21.6% Activity active PP (excl.) passive PP (excl.) active and passive PP n.s. Definitions 74.8% 0% 22.5% 2.7% Measuresa 58.9% 0% 40.0% 1.1% Note: n = 151; n.s., not specified; excl., exclusively; PP, political participation. a Only publications offering a definition of OPP or a similar concept (n = 185). b Only publications referring to the online sphere (n = 98 for definitions). 12 C. RUESS ET AL. The largest divergence between conceptual foundation and measurement can be specified in terms of sphere: While only 32.4% of definitions reference the online sphere exclusively, a majority of measures does (86.5%). We find that 31.8% of definitions address both the online and offline context, but only 13.5% of measurements do. Regarding specificity, 39.8% of definitions are platform-independent, yet 78.4% of measures specifically reference a platform. A divergence also emerges when taking the level of activity of the analyzed political behaviors into consideration: While 74.8% of definitions focus entirely on active partici- pation, only 58.9% of measurements do so. Instead, 40.0% of measures account for pas- sive (along active) participation – which only 22.5% of definitions do. Discussion Based on an international interdisciplinary systematic literature review of peer-reviewed English-language empirical studies reporting on quantitative surveys and experimental studies, this analysis provides an overview of the conceptual understanding of OPP as evidenced both explicitly by the definitions provided and implicitly by the measurements applied. Thereby, this study contributes to the ongoing debate on how to conceptualize political participation in an online sphere by highlighting how research on OPP has in fact defined and operationalized the concept. We document a dynamic growth in the body of research over the past twenty years, gaining momentum with the establishment of social media platforms around 2005 as well as upticks related to the US presidential elections (2008, 2012). This finding aligns with our argument that a rapidly evolving field such as OPP studies, strongly shaped by the emergence of new platforms for political online behaviors – each coming with specific technological affordances and restrictions – may be especially prone to pragmatic, ad- hoc decisions on how to empirically capture, i.e., measure, its core phenomenon. We compare the definitions to the measures of OPP and find several disconnections. First, while the shift in the conceptualization of political participation to the online sphere is clearly reflected in the applied measures, it is not equally reflected in the provided definitions. While most measures almost exclusively address online activities (about 90%), only few definitions focus on the online sphere (about 30%). This may be due to the fact that the field still heavily relies on definitions of political participation developed in a pre-digital context. Definitions, thereby, tend to be ambiguous as to the sphere in which the examined political behaviors occur. Theocharis (2015) remarked, ‘[t]here is no other way to concretely find out how much this type of participation matters to citizens and how strongly we should expect it to affect democracy, unless we measure it’ (p. 11). However, if authors consider online activi- ties sufficiently salient to be included in a measure of political participation, this evolved understanding of the phenomenon should be made conceptually explicit. Second, while overall, only 22.5% of definitions account for passive activities (e.g., read- ing the news), 40.0% of operationalizations include items capturing such passive behaviors. In this analysis, we apply the typology developed by Gibson and Cantijoch (2013) dis- tinguishing nine modes of OPP and provide an overview of the prevalence of various activities covered by empirical OPP research. It should be pointed out that using the term ‘passive engagement’ to denote online discussion and expression may be somewhat misleading, as these activities require heightened levels of attention and engagement.  INFORMATION, COMMUNICATION & SOCIETY 13 Gibson and Cantijoch (2013) themselves point out that in an online setting, such activities should be considered ‘participatory’ and can quickly morph into coordinated active par- ticipation. Accordingly, new conceptualizations tend to position information-seeking activities as a mode of (active) political participation (Kim & Hoewe, 2020). Third, the results show that with the emergence of social media platforms, OPP research has focused to a large degree on expressive modes of online behavior. Expressive activities are popular measures in indexes of OPP, but the existing definitions of political participation rarely address the expressive dimension of participation. Focusing on the measurements of OPP applied in the field, it could be argued that ‘passive engagement’ (as defined by Gibson & Cantijoch, 2013), i.e., expressive partici- pation, discussion, and attention to news, plays a prominent role in attempts to capture OPP because these are the behaviors afforded by the online platforms most heavily employed for political purposes. This demonstrates the mediated nature of OPP. Empiri- cally, OPP is inextricably intertwined with the functionalities offered by online platforms (about 80% of measures are platform-specific). Interestingly, though, only 60% of definitions discuss specific online platforms. We find that around 2008, social networking sites became the focus of OPP research, which they remain to date. We also see an increase in the focus on expressive activities. These two trends may be connected as social networking sites facilitate expressive forms of OPP. In this context, it should be noted that recent studies highlight the sociodemographic heterogeneity of different social media platforms’ user bases (e.g., Blank & Lutz, 2017) – another point in favor of a con- textualized conceptualization of OPP. To summarize, we find that the measures applied in the field tend to implicitly acknowledge that OPP is not merely an online equivalent of traditional, offline political participation. Instead, it is shaped by and contingent upon the platforms on which pol- itical participation is conducted – which, in turn, raises the salience of expressive forms of participation as well as political discussion and attention to news. Current definitions, however, if provided at all, tend to uphold the notion of the offline/online equivalence of political participation. Many are adaptions of definitions dating back as far as the 1970s (e.g., Conway, 1985; Verba et al., 1995; Verba & Nie, 1972). Our systematic literature review thereby highlights a need to clarify the conceptual understanding of OPP, and, more specifically, to bring it in line with the measures actu- ally applied in the field. We find evidence of a form of implicit theorizing, in that measures have quickly evolved to embrace the mediated nature of OPP, its platform- specificity, and its reliance on expressive and discursive modes of political engagement. This implicit conceptual evolution, however, is largely relegated to the shadows of the methods section, in many instances even footnotes or appendices. Bringing it to light by providing an adequate definition of the concept is necessary not only to avoid incon- sistencies within studies, but to ensure that seemingly contradictory findings caused by conceptual confusion may be avoided, rendering analyses more comparable. Limitations and suggestions for future research This systematic review is limited to measurements of OPP in quantitative survey- based and experimental studies. However, important conceptual developments have emerged from the qualitative literature. So, qualitative studies and conceptual 14 C. RUESS ET AL. pieces may offer additional insights into the complexities related to defining and measuring OPP. We find that the quantitative field of research would benefit from a more careful review of reporting practices. We can conclude from our descriptive results that Eng- lish-language quantitative empirical studies on OPP largely lack definitions of the con- cept with regard to the sphere in which political participation takes place. It is striking that 138 of 289 publications lack definitions of the concept entirely. In addition, when definitions are applied, they vary across the elements they cover. The identified lack of a theoretical foundation might be the cause for the inconsistent operationalizations applied to the phenomenon and ‘arbitrary measures’ (Theocharis, 2015, p. 4), which has led to conflicting conclusions about whether online participation is more or less equitable than offline forms of participation. This is also associated with varied measurements: Measures tend to show much higher levels of specificity and thus on average more elaboration than the preceding definitions or lack thereof – indicating that the field is structured by an ad-hoc approach based on prior studies, state of the art, and implicit theorizing instead of explicit theoretical or conceptual reasoning. We recommend future research to develop a conceptual understanding or definition of OPP that precedes the operationalization (Hooghe, 2014, S. 339). Another issue in current research on OPP is a lack of transparency as to the applied methodology. The exact operationalization of OPP is either not fully or not at all dis- closed in 34 cases. Beyond that, information about the internal consistency of indices is missing in 39.3% of OPP measurements, and several publications abstain from indicat- ing the year(s) of data collection (n = 27), the country of data collection (n = 2), the size of the sample (n = 5), or the survey mode (i.e., data collection online, via telephone, face-to- face, etc.) (n = 38). As for future research, we acknowledge that existing research is dominated by Wes- tern, well-established democracies. As such, there are gaps in our understanding of OPP in non-democratic states. This is an important gap, because these states may not provide opportunities for offline participation and therefore political activism through channels like social media becomes relevant, as shown in the context of the Arab spring (Wolfsfeld et al., 2013). In addition to systemic characteristics, the Western bias of OPP research also comes with a culture-specific, most likely hegemonial understanding of pol- itical participation, its spaces and restrictions (Almond & Verba, 1963), that has an impact on both theoretical debates and empirical measures, thus narrowing the general- izability of OPP research. Future research might consider the affordances of different platforms in providing opportunities for participation, particularly around self-expression. Given the focus on expressive and passive forms of engagement in contemporary empirical OPP research, scholars might explore how these are correlated with participants’ sense of internal and external efficacy, as opposed to more conventional political acts. At the same time, future studies might pay attention to more complex types of digital participation (like engaging in hacktivism) compared to expressive activities which tend to require little resources and still appear to be at the center of the OPP discussion. In this context, future research might also offer more sophisticated measures of participation that are scaled by level of effort, rather than indexed, as though all forms of partici- pation are of equivalent effort and impact (e.g., use Mokken scaling), and that