Agency is recognized as a key behavioral motive in a wide range of environmental communication research (Bandura, 2000). Two factors are commonly used in such investigations - the first is variously referred to as capability or perceived behavioral control, while the second is variously called response efficacy or outcome expectancy (Koletsou and Mancy, 2011). We use the terms capability and response efficacy. Both forms operate at different levels of analysis, from the individual to the collective: definitions can be found in Table 1.

Research to date with the PRIA, RISP, and PRISM appears to have largely constrained investigations of efficacy to capability rather than response efficacy, and to the personal level rather than collective or proxy levels (Hmielowski et al., 2019; Koletsou and Mancy, 2011). Environmental IA research continues this focus on individual-level capability efficacy and has typically investigated it in relation to information attainment. The small number of results in this area are mixed. Researchers working in the context of climate change found both negative relationships between capacity to find information and IA (Yang and Kahlor, 2012) and null effects in the area of Great Lakes health (Kahlor, 2006). However, Dunwoody and Griffin (2014), researching water resource risks, exhibited mixed findings, with one project showing a negative relationship between capacity to find information and IA, while another found the opposite. None of these studies appear to have examined the efficacy of avoiding information, or response efficacy. Of note, a recent IA meta-analysis did find that efficacy beliefs exerted a stronger relationship on IA in environmental than health contexts, but these beliefs were examined as a universal concept, ‘confidence beliefs’, without specifying whether they related to seeking or avoidance, or response versus capability efficacy (Liu and Chen, 2024). The only research we know of to date regarding response efficacy and environmental IA showed that the perceived financial burden of undertaking risk mitigations was positively related to IA (Losee et al., 2020).

The PRIA focuses on social norms as a key socio-cultural factor to the exclusion of others such as values (Steg and de Groot, 2012) or social support (Ou and Ho, 2022). Social norms are conduct codes that guide our behavior (Link et al., 2023). Norms drive IS (Ou and Ho, 2022; Liu et al., 2022), and recent research in health IA shows that IA itself can be seen as a non-normative behavior, compared to IS (Heck and Meyer, 2019; see also Fung et al., 2024). Work on IA and norms is in the beginning stages. Seeking to differentiate avoidance from seeking norms, researchers have now shown avoidance norms significantly increase avoidance intentions (Link, 2021; Qu et al., 2021). Mixed results within the small number of studies on the topic in environmental contexts might be due to differences between seeking versus avoidance norms, as well as between descriptive and injunctive norms (see, for example, Link, 2024).

Continuing this trend in environmental communication, research found both positive and negative relationships between seeking norms and avoidance in relation to climate change (Kahlor, 2006; Yang and Kahlor, 2012), a positive relationship between seeking norms and IA in relation to PFAS contamination (Liu and Yang, 2023), and positive relationships between avoidance norms and intentions in the context of manmade earthquakes (Kahlor et al., 2018). However, other researchers found null relationships between social norms and avoidance (Hwang and Jeong, 2020). Notably, the manmade earthquake study was the first to look at avoidance, as opposed to seeking, norms in environmental contexts. Overall, these mixed findings indicate more research is needed.

IA research in communication typically adopts a dimensional approach to emotions that is concerned with valence or arousal as key predictors of IA where valence reflects emotional quality and arousal reflects emotional intensity (Deline and Kahlor, 2019; Leblanc et al., 2015). Valence research has shown that negative affect is negatively related to IA intentions (Kahlor et al., 2018), while arousal research has also shown effects: for example, overwhelming worry is associated with individuals avoiding cancer screenings (Peng et al., 2019). However, results here are also inconsistent, with recent research in the context of COVID-19 finding null effects between fear and IA (Xu et al., 2023; Zhou et al., 2021).

Contextualizing these factors within the environmental arena also shows mixed results. For example, research found that worry (identified as a negative emotion) was unrelated to attitudes about avoiding environmental health risks (Fung et al., 2018), but a recent meta-analysis showed that worry was more strongly related to IA in environmental, as opposed to health, contexts (Liu and Chen, 2024). On the other hand, researchers found that those with positive attitudes about climate change were more likely to avoid information on the topic, suggesting participants were seeking to maintain positive mood (Yang and Kahlor, 2012). And exploratory, inductive research has shown a desire to maintain hope as a form of mood management in relation to IA (Barbour et al., 2012); therefore, emotional motives also appear key to IA processes.

This theme described what occurred when participants used avoidance to manage their feelings. Participants described both emotional valence and arousal. For example, valence ranged from positive, described as ‘…good thoughts…’(32) to negative, such as one participant stating “Sometimes I just do not want the doom and gloom. It can be depressing to read about that all the time” (117). Arousal ranged from low levels [described as ‘…being upset…’ (185)] to high levels [such as being ‘…overwhelmed…’(12)]. As one participant with high arousal stated “I cannot pinpoint it but it’s definitely happened multiple times. It’s almost like an impending doom feeling that sometimes I just cannot handle” (152). In addition to addressing their own valence or arousal, participants also referenced how mediated or interpersonal messages felt. For example, referring to the valence of mediated messaging, one participant noted, “You often see negative things in the news like forest burning and the garbage in the oceans. Its emotionally negative to always hear about all the bad news” (20) while another participant, referring to arousal regarding interpersonal messages noted, “Sometimes it gets overwhelming and I need to decompress before continuing the conversation” (69). These motives seem similar to mood management, a coping behavior to regulate one’s mood by attending to both emotional valence and arousal (Zillmann, 1988).

Our participants expressed agency motives by assessing their own and others’ perceived capability to address environmental issues, as well as whether they thought their efforts or those of others would result in consequential change. These motives seem to align with capability and response efficacy at different levels; definitions and exemplars for these levels are found in Table 1.

Turning first to capability, participants described a range of perceptions. For example, here a participant refers to their capability motivating IA: “If I’m doing all I can, there’s not much use in getting stressed about it” (169) while another participant described how incapability perceptions motivated IA: “Sometimes the issue is to [sic] depressing because I know I cannot do anything to help or prevent it. I try to avoid or pretend it’s not happening…” (126).

In addition to differing capabilities, participants also ranged in response efficacy, from not thinking that their own or others’ actions would be consequential to thinking that they would be. For example, here one participant used IA to ensure their actions felt consequential:

On the other hand, participants often described the perceived inconsequentiality of their actions as a reason for avoiding information:

Participants often appeared to perceive environmental issues as hazards motivating IA – that is, threats to themselves and what they valued (Paek and Hove, 2017). Our sample focused on how severe the threat was, as well as susceptibility, the timing of the threat, and who the threat was directed at. Threats in the literature are often defined in terms of perceived severity and susceptibility: perceived severity refers to beliefs about how serious the threat is, while perceived susceptibility refers to perceptions about one’s odds of experiencing a threat (Witte, 1992). Accordingly, participants ranged in how severe they perceived environmental threats to be, describing them as less severe ‘…problems…’ (49) to more severe ‘environmental disaster(s)’ (39; 79) and ‘destruction’ (19) scenarios. Discussion of severe harm was common; as an example, one participant noted an avoidance motive as “catastrophe burnout” (11), while another stated, “I cannot bear to think about death of so many individual animals due to fires, ocean temperature rise, etc. Let alone species extinction” (60). Our participants were also motivated by perceived susceptibility, ranging from being sure to unsure about their vulnerability to environmental hazards. One participant described their surety about not being vulnerable this way:

On the other hand, participants also described being unsure about their vulnerability, as expressed by this participant: “Sometimes it’s scary or overwhelming to hear about the demise of the planet, and I wonder if it’s being overblown, and why we do not hear from the scientists that denounce climate change” (145).

In addition to perceived severity and susceptibility, participants ranged from current to future hazard time orientation; here a participant discussed future hazards: “There are times when the weight of it all feels like too much. I do not think we are doing enough to stop global warming, and I worry what the future will look like for my kids” (127).

Finally, participants ranged in who they considered to be under threat. For example, here a participant referenced themselves as under threat: “Sometimes it’s all bad news and I’ll tune it out for my mental health…” (33).

On the other hand, many participants referred to other referents – both human and non-human - as motives: “Sometimes I’m just not in a place where I can hear about harm coming to people or animals” (196). This motive appears similar to environmental concern values which represent three types of valued referents – concern for the self (egoistic); others (altruistic); and concern for animals, plants and ecosystems (biospheric) (Schultz, 2001).

Relationship frames represent cognitive schema about the qualities of relationships (Solomon and McLaren, 2008), and our participants appeared to use dominance-submission and affiliation-disaffiliation frames, indicative of relational framing theory (Solomon and McLaren, 2008). Dominance-submission “…refers to the degree to which someone controls, influences or has status over the other,” while affiliation-disaffiliation “…captures the appreciation, esteem, or solidarity one person has for the other” (Solomon and McLaren, 2008, p. 3).

Participants described others who were communicating about environmental issues as unlikable and/or as overbearing, suggesting disaffiliation and dominance frames. Here, a participant described their dislike: “I have heard enough from and seen enough of Greta Thunberg to last a lifetime. Not impressed” (2). Another participant detailed overbearingness: “It’s not any specific environmental issue--it is people who just will not let it go. It is people who carry on and harass people if they do not shore {sic} their feelings” (177). Notably, our participants did not use affiliation or submission frames of others as IA motives.

Participants often referenced information exposure as an IA motive. Many participants described the volume of information and messages about environmental issues as excessive: as one participant stated, “Sometimes it’s just too much information, too sad, and I’ll change the channel or move a conversation in another direction to avoid the topic” (132).” Many participants also referenced the repetitive nature of environmental information to which they were exposed: “Sometimes you hear about the same issues over and over and over. The content remains the same just a different talking head” (67). They detailed being tired and bored by such information: “I’m at the point where I’m tired of hearing about almost all issues. Especially the same old talk,” (194) and “It’s boring” (5). These four factors together comprise chronic message fatigue: “…an aversive motivational state of being exhausted and bored by overexposure to similar, redundant messages over an extended period of time” (So et al., 2017, p. 10). However, some participants only discussed being overexposed to information without qualifying it further, such as this participant: “Sometimes you get so much over load on an issue you just want to turn it off” (107). We must therefore consider that without those other chronic message fatigue factors, expressions of only overexposure mean those participants may be experiencing other information exposure phenomenon, such as information overload, which refers to being overwhelmed by information volume (So et al., 2017).

Participants referred to assessments of whether information was credible as an IA motive, and focused on both the information content and who the information was coming from. Content was assessed by participants using criteria such as accuracy, sufficiency, disinformation, and argumentation quality. For example, as one participant using accuracy and disinformation criteria stated:

Regarding information sources, participants ranged in referring to those external to themselves via mediated and interpersonal channels, as well as themselves as an information source. When assessing others as credible sources, participants used a range of assessment criteria, from ideology to beliefs to accuracy to disinformation to education levels. For example, as one participant using education criteria noted:

On the other hand, participants who spoke of themselves as sources often referenced their belief in their knowledge or information salience as assessment criteria. For example, as one participant noted: “I know about the environmental issues facing us today and I do my part to try to help. I do not need to hear about something I already know about” (37). These different foci on content and source credibility appear like the concept of information credibility, defined as “…the believability of a source or information…evaluated at the medium, source and message levels” (Metzger et al., 2020). Occasions when participants assessed themselves as credible sources appear similar to the concept of epistemic authority, in which the self is considered a reputable source in source credibility decisions (Kruglanski, 2012).

Participants sometimes referred to a specific topic as a reason in and of itself for avoiding environmental information, and topics ranged from climate change to mining. As an example, one participant noted: “We all have tried to hide our heads in the sand at one time or another. Especially [sic] concerning the ozone layer and coal mining” (81). Such a statement is reminiscent of recent findings from Link (2024) that specific topics themselves induced IA, e.g., that information on disease-states was typically avoided, but other health related information was not.

Our participants were motivated by information credibility assessments to avoid information. Research in IS shows that information characteristics are an important factor for understanding that information behavior (Ou and Ho, 2022), and this motive seems to indicate the same importance in relation to IA. In risk communication, when credibility research is studied with models like the RISP, it is often subsumed under a factor called relevant channel beliefs that appears to focus on medium credibility (Griffin et al., 1999; Hwang and Jeong, 2020; Metzger et al., 2020). These beliefs represent holistic assessments of information providers, ranging from doctors to TV programs, and are among the least studied factors in both RISP and PRISM research (Yang et al., 2014, 2022). Yet, the emphasis people placed on credibility in this study, as well as in other exploratory health studies (Barbour et al., 2012; Link, 2024) speaks to the importance of this concept to IA.⁸ Recent research supports this significance, at least in relation to IS – health meta-analyses have found that information factors exhibited larger effect sizes on IS than cognitive or affective factors (Ou and Ho, 2022; Wang et al., 2021). Given these findings in relation to IS, mentions of it in exploratory health IA studies, and our participants’ use of the factor, we suggest that information credibility therefore be added to the PRIA under its own category of Information Characteristics. IA researchers could use this factor to push the field forward, for example by investigating if IA credibility standards differ from IS standards, and the importance of those standards to different audiences.

Considering information characteristics also opens opportunities to consider IA relative to the news media. Today’s media environment, with its proliferation of social media platforms and personalized algorithms has changed audiences’ news practices (Swart, 2021; Toff and Nielsen, 2018). One way to reflect on these changes is by considering active versus inactive information engagement and IA.

According to media scholars, in the 20^th century direct discovery was afforded by practices like watching TV broadcasts or reading newspapers, where people actively engaged with information directly from publishers. In contrast, the 21st century is characterized by distributed discovery, where digital intermediaries between publishers and audiences occur in the form of search engines and social media (Toff and Nielsen, 2018). In these environments, information is frequently curated by information proxies, such as algorithms (Stoldt et al., 2023; Swart, 2021). Using these proxies seemingly represents a degree of divestment from active information engagement (Swart, 2021). An example is the ‘news finds me’ (NFM) phenomenon, which refers to the belief that direct news engagement is unnecessary because exposure to news through ones’ peers and platforms is sufficient (Toff and Nielsen, 2018). It is composed of three factors: a belief that one is adequately informed about public issues; a reliance on peers and platforms for news updates; and a corresponding belief that one has little need for active news engagement (Skurka et al., 2023b; Song et al., 2020). NFM is on the rise – recent research shows roughly over 1/3 of Americans hold the perception, and it has been studied as a form of IA (Gil de Zúñiga et al., 2020; Toff and Nielsen, 2018). How might considering distinctions between active and inactive information engagement inform our understanding of IA moving forward? Below we reflect on this question in relation to another information behavior, inertia (inaction in service to the status quo; Polites and Karahanna, 2012), and by conceiving of NFM as a process.

Turning first to information behavior, we recognize that active information engagement is a fundamental assumption of IA conceived of in the PRIA, given its theoretical roots in 20^th century theories and models, like the RISP (Deline and Kahlor, 2019). Given NFM is a form of distributed discovery, associated with divestment from active information engagement via information proxies, we therefore question whether we should consider NFM as a form of active avoidance, or a related concept: inertia. Two NFM characteristics seemingly support an inertia classification: passive exposure and information proxy use.

The first factor, passive exposure, described as ‘ambient news’, occurs when one is exposed to information while doing other things on social media (Toff and Nielsen, 2018). Such exposure is also defined as information scanning – “…information acquisition that occurs within routine patterns of exposure to mediated and interpersonal sources…” (Niederdeppe et al., 2007, p. 154). This non-intention to engage with information one is exposed to seems similar to inertia’s ‘inaction’. The second factor aligns with Franklin’s (1999) observation that technology is a practice, and the status quo practice designed into social media platforms and personalized algorithms is a seeming divestment of active information engagement to information proxies who provide a steady stream of curated and personalized information for one to be exposed to Swart (2021). Actively engaging with information by avoiding information that one is motivated to avoid (IA) is a different practice than passively encountering information within one’s news feed that can be ignored (inertia).

But it is also possible that suggesting NFM as inertia obscures different forms of avoidance within NFM stages. In other words, might NFM contain both IA and inertia behaviors? And if so, how do the behaviors inform each other through the process? For example, part of NFM implicitly includes the active choosing or installation of apps and/or peer networks, representing a form of initial active information engagement with an information proxy (Barnidge and Xenos, 2024; Schäfer, 2023). If so, are source credibility considerations part of this process? Do people avoid certain platforms or peer networks? If they do, how and why do they do so? Further questions could examine whether proxy avoidance increases or decreases subsequent perceptions of quality or accuracy in the information one is exposed to. Considering IA relative to these factors can further inform explanatory and predictive power of the concept, as well as sharpen our understanding of distributed discovery processes like NFM moving forward.

Finally, information credibility could also be used to explore uneven findings in relation to other cognitive IA factors, such as efficacy. For example, one question potentially fruitful for future research would be whether, and if so, how, low information credibility assessments were related to collective response efficacy as an IA motivator. This is because collective response efficacy relies on perceptions of collective agreement levels, in this case, about whether the collective behavior will result in meaningful action (Koletsou and Mancy, 2011). However, perceptions of social agreement can be inaccurate, as discussed in literature discussing theories ranging from co-orientation to pluralistic ignorance (Geiger and Swim, 2016; Leong et al., 2007; Mildenberger and Tingley, 2019). It therefore seems reasonable to question whether credibility assessments might affect not only the trustworthiness of information about environmental issues, but also perceptions of collective agreement. For example, if I do not consider the environmental information I’m exposed to as credible, perhaps that credibility assessment spills over to perceptions that levels of collective actions in response to the environmental issue are also untrustworthy, decreasing my perception of collective response efficacy and further increasing IA. More research into these factor relationships through abductive research is warranted.

The second unexplored motive we identified is relationship frames, which appears consistent with interpersonal relational framing theory (Solomon and McLaren, 2008). There are few studies that we know of pertaining to interpersonal communication within environmental communication, although a burgeoning field of work has experimentally examined a form of interpersonal communication, aggressive communication (see, for example, Yuan and Lu, 2020). In contrast to the environmental communication field, risk communication has addressed interpersonal factors more prominently with models such as the social amplification of risk, where social interaction increases or decreases perceived risk significance among laypeople (Kasperson et al., 1988; Rickard, 2011).

Given these findings and their potential relationship to how environmental risks are socially constructed through interpersonal communication (Cantrill, 2010) we suggest the inclusion of a new PRIA category – interpersonal factors, which would include relationship frames as an item for further investigation. For example, our participants who used these frames frequently referred to activists, consistent with other findings that environmental activists are often assessed as overbearing and unlikeable (see, for example, Klas et al., 2019). Future research could examine, for example, whether changes in levels of dominance or disaffiliation cues in activist messaging would affect IA.

Including relationship frames in the PRIA could also inform research in other environmental communication areas, such as environmental dialogue. Dialogue refers to developing shared understandings through communication processes that foster openness and respect, making civility important (Black and Wiederhold, 2014; Gastil and Black, 2018). Civility is often expressed through interpersonal communication norms – such as not interrupting others, or not using offensive language (Bonotti et al., 2024). Yet there are two major understandings of civility and we wonder whether, and if so how, misunderstandings about parties’ conceptions of civility could affect environmental dialogue.

For example, one understanding of civility appears to be at the individual level, where politeness norms are enacted in service of tolerance - a basic acknowledgement of others as social actors that we exist with (Bardon et al., 2023; Bonotti et al., 2024). Bardon et al. (2023) call this civility as politeness. Another understanding is of civility at the community or social level, where civility is understood as a collective, civic goal of recognizing others’ moral worth (Bardon et al., 2023; Bonotti et al., 2024). Bardon et al. (2023) call this civility as public-mindedness. Those who hold this second understanding of civility sometimes enact strategic incivility – called critical impoliteness – to secure public-minded civility (Bonotti et al., 2024). As Bonotti et al. (2024) note, “Grassroots movements…may sometimes need to employ impolite means in order to advance their public minded goals” (p. 8). However, this understanding of critical impoliteness, vis-a-vis dominance or disaffiliation behaviors, might not be shared with social actors like our participants. Instead, actors might instead ascribe an individual lens to these actions, and see critical impoliteness as violating basic acknowledgements of people as social actors, motivating their IA.

If this is the case, it might suggest a context of false consensus – where advocates and activists assume audiences, like our participants, share their conception of impoliteness as needed for the greater good of public-minded civility, but in fact this shared understanding is false (Leong et al., 2007). If this is the case, potential remedies could include efforts to correct civility misperceptions through activities like facilitated dialogues and workshops (Leong et al., 2007). Considering how relationship frames and civility operate relative to environmental dialogue and IA could therefore be a potentially fruitful avenue for research moving forward.

The last motive we address here, emotional arousal, is not new to the PRIA but is underexplored, represented in the PRIA with the affective response factor (Deline and Kahlor, 2019). We do not know much about arousal in relation to IA, providing a generative research gap. For example, traditionally RISP and PRISM studies have focused on valence to the exclusion of arousal, but a significant body of research suggests such uni-dimensional approaches limit our understanding of the full(er) extent of how emotion influences IS and avoidance behaviors (Hmielowski et al., 2019; Myrick and Nabi, 2017). As an example of such limitations, a recent IA metanalysis showed that negatively valenced emotions were differently related to IA, with anxiety positively related, and worry negatively related (Liu and Chen, 2024). These inconsistencies suggest additional emotional factors, such as arousal, be used to explain and predict IA relationships.

Future research could take a more comprehensive approach by using models that integrate the two factors: for example, the circumplex emotional model considers emotions to be governed by both valence and arousal (Leblanc et al., 2015; Russell, 1980). In this model, fear evinces high arousal and negative valence, while excitedness is comprised of high arousal and positive valence, countering the common misunderstanding that arousal necessarily equates anxiety (Burgoon et al., 1989) and paving the way for better understanding how a variety of participant experiences of arousal contributes to IA processes. For example, researchers could use the circumplex model to investigate the effects of both excitedness (positive valence, high arousal) and contentment (positive valence, low arousal) on IA (Russell, 1980).

Our participants described multiple levels of response and capability efficacy motives, ranging from the self to proxies (see Table 1). This suggests participants perceived environmental issues to be multi-scale phenomena, reflecting how such issues, like climate change, are conceived of in the literature (Koletsou and Mancy, 2011). This indicates that multiple levels of efficacy, from the self to governmental proxies be considered when dealing with multi-level environmental issues and IA (see Table 1). What little work we have on these factors’ levels indicates more investigations are needed into issue scale in both environmental communication and IA research (Doherty and Webler, 2016; Koletsou and Mancy, 2011). For example, environmental communication research shows that collective efficacy is a stronger predictor than self-efficacy regarding pro-environmental behavior (Chen, 2015; Choi and Hart, 2021) and proxy efficacy - a belief that a third party’s involvement is necessary to achieve one’s goals - is perceived to be more effective at dealing with environmental issues writ large (Bostrom et al., 2019; Li, 2018). Establishing participants’ perceptions of environmental issue scale therefore appears warranted to understand which levels of efficacy constructs might be at work. Pilot studies (commonly called elicitation studies in health communication: see Ajzen, 1991 and Niederdeppe et al., 2007) can be used for these purposes; to our knowledge, no such work to date in relation to environmental IA has occurred.

In addition to scale, one of our participants’ motives – information exposure – represents their attention to issue duration and time, another boundary condition (Zaheer et al., 1999). Many environmental issues are labeled slow crises – that is, they extend over years (Lu, 2022). For example, toxic waste contamination is entrenched in communities of color in the US, a form of environmental racism (Bullard, 1993). Two of our participants’ top three issues – climate change and waste – are of long duration and therefore factors such as chronic message exposure (Lu, 2022) and associated fatigue (So et al., 2017) would be fruitful concepts to consider in future research efforts, especially given our participants’ references to chronic fatigue factors. Moreover, in addition to our findings on the topic, a recent inductive study on health IA identified two different types of IA based on temporal factors – a short term version to manage emergent emotions, and another related to what is perceived to be long-term threats (Link, 2024); in addition, recent deductive research shows fatigue’s positive relationship to IA (Ford et al., 2022; Gurr and Metag, 2022).

Considering duration as a boundary condition in relation to IA will have practical benefits – for example, chronic message fatigue has been found to be harmful to laypeople’s responsiveness to evolving environmental challenges (Siebenhaar et al., 2020) such as climate change (Lu, 2022), with the potential to hinder timely and effective mitigation and response. Future research could examine how IA is implicated in these relationships.

Participants attended to non-human stakeholders, which we suggest represents biospheric values – concern for animals, plants, and ecosystems (Schultz, 2001). Non-human referents are implicitly characteristic of natural resource depletion issues, one of our participants’ other top three environmental issues. Additionally, non-human referents are not normally part of health communication discourse (Lapinski et al., 2023), which supports our contention that boundary conditions be examined for different communication spheres relative to the PRIA. To our knowledge, no work to date has examined these values in relation to IA.

There are two major opportunities for future research associated with the boundary conditions suggested by these motives. The first approach is to consider which motives might be elicited by different boundary conditions, which can increase predictive and explanatory power. As de Young (2000) notes, there is no one motive that operates under all conditions; for example, proxy efficacy is not likely to be considered salient as an IA motive if the environmental issue is not perceived to be a collective level issue, requiring proxies’ involvement. Likewise, chronic message fatigue is not likely to occur if the environmental issue is not perceived to be of long duration. This speaks to McPhee and Poole’s (2016) assertion that theory development needs to consider context. Future research could map whether specific boundary conditions occasioned by different environmental issues might induce different IA motives.

Second, considering boundary conditions and the IA motives that they elicit can enrich our understanding of different IA processes (McPhee and Poole, 2016). For example, to our knowledge, no work has occurred on socio-cultural values in relation to environmental IA, however, our participants attended to biospheric referents as an IA motive, representing biospheric values (Steg and de Groot, 2012). Research links these values to environmental identification, which is a perception that nature is a part of who that person is (Clayton et al., 2021). It seems reasonable to suggest biospheric referents could therefore either evoke or signal environmental identification. If this were the case, threats to these species/ecosystems could constitute an identity threat that might be defensively avoided. Forwarding research in this way could contribute to bodies of environmental communication research. For instance, investigations into unexpected campaign effects on harm to animals to date has largely prioritized emotional factors (Lu, 2022; Swim and Bloodhart, 2015), but considering such effects through an identity based defensive avoidance process (see, for example, Braun and Niederdeppe, 2012) could broaden our understanding of the issue.