To respond or not to respond: The effects of interpersonal goals on responsive Facebook behaviours

Introduction

Interpersonal interactions have been redefined by Social Networking Sites (SNSs; Kross et al., 2021). Globally, there are over four billion users, averaging approximately two and a half hours daily on SNSs (Kemp, 2023). These platforms continue to rise as a way to maintain and foster social connections (Carpenter et al., 2018; Kemp, 2024).

However, the harms versus benefits of SNSs are often debated. Some link SNS use to negative well-being-related outcomes, such as psychological distress, reduced life satisfaction and lower affective well-being, via upward social comparison, envy (Krasnova et al., 2015), and cyberbullying (Giumetti & Kowalski, 2022). Conversely, others highlight benefits, including enhanced social capital (Liu et al., 2016) and fulfilment of social needs. But outcomes can be contingent on individual differences, motivations and behaviours (Kross et al., 2021; Valkenburg et al., 2022). Research analysing beneficial SNS use has considered active versus passive use (Trifiro & Prena, 2021; Verduyn et al., 2021) and relationship maintenance behaviours (Burke & Kraut, 2016; Roper & Tobin, 2022).

SNSs can effectively and efficiently facilitate the cultivation and maintenance of different relationships (Bryant & Marmo, 2009; Ellison et al., 2014). Facebook offers various response modalities to actively engage with others, including one-click reactions, comments, and private messages. These options allow people to respond publicly or privately and with varying levels of thought and effort. These responsive behaviours play an important role in relationship outcomes and well-being (Burke & Kraut, 2014, 2016). For example, warm, targeted, active SNS use, like writing a supportive comment on a post, has been emphasised in Verduyn and colleagues’ (2021) extended active-passive SNS use model as benefiting well-being. However, how users respond to posts can be influenced by aspects including post valence (Ziegele & Reinecke, 2017), desires to maintain relationships or please others (S.-Y. Lee et al., 2016), and the publicness and effort associated with different response modalities (Burke & Kraut, 2014). The current study considered these factors and aimed to expand our understanding of what motivates responsive behaviours on Facebook by examining interpersonal goals.

Compassionate and self-image interpersonal goals can energise social interactions and communication patterns (Canevello & Crocker, 2010, 2011, 2015, 2017). Compassionate goals focus on the well-being of others, while self-image goals focus on being perceived as desirable (Crocker & Canevello, 2008, 2018). Recent research has documented the influence of these goals on SNSs (Roper & Tobin, 2022; Tobin et al., 2020; Toh & Lee, 2022). Specifically, compassionate goals have predicted more responsive behaviours on Facebook (Roper & Tobin, 2022; Tobin et al., 2020). However, this research has not distinguished differences in response modalities, and the correlational findings cannot determine whether compassionate goals cause greater responsiveness on SNSs.

The current study experimentally tested whether compassionate and self-image goals predict specific responsive behaviours to Facebook posts. This approach addresses the limitations of previous research and extends understanding of what could motivate SNS behaviours. Our study was informed by foundational work on interpersonal goals in offline contexts (Canevello & Crocker, 2010; Crocker & Canevello, 2008) and considers how novel aspects of the online context could affect their operation.

Interpersonal Goals

According to ecosystem and egosystem theory, interpersonal behaviour is guided by two underlying motivational systems (Crocker & Canevello, 2018). Ecosystem motivation promotes prosocial behaviours aimed at benefiting others and avoiding harm, typically seen in compassionate goals. Conversely, egosystem motivation promotes behaviours aimed at maintaining, enhancing and protecting a desired self-presentation, typically seen in self-image goals (Crocker & Canevello, 2008, 2018; Crocker et al., 2009). These interpersonal goals can change as a function of internal and external factors (Canevello & Crocker, 2015). Studies with college roommates using daily measures of compassionate and self-image goals over several weeks found that these goals can fluctuate based on interactions and internal states, and the researchers position them as dynamic states (Canevello & Crocker, 2010; Crocker & Canevello, 2008).

Compassionate Goals

Compassionate goals uniquely and positively influence behaviours. These goals are associated with increased responsiveness and improved relationship quality (Canevello & Crocker, 2010) and predict increased regard for others and improved well-being (Canevello & Crocker, 2017; Crocker et al., 2010). Notably, compassionate goals are distinct from other constructs like empathy, compassionate love, and communal orientation. Canevello and Crocker (2020) found that although compassionate goals intersect with common measures of prosocial orientation, they uniquely predict gratitude, giving, and reasons for giving. These findings remained consistent despite gender, social desirability, and relationship length.

Research on interpersonal goals has mainly considered close relationships in offline contexts (Canevello & Crocker, 2010, 2011, 2017; Crocker & Canevello, 2008; Crocker et al., 2010; Duarte & Pinto-Gouveia, 2015). In these contexts, responsiveness refers to providing a particular form of support that is sensitive to others’ desires to feel understood and valued (Canevello & Crocker, 2010; Reis & Clark, 2013; Reis et al., 2004). Associated with increased social support and responsiveness, compassionate goals are influential when interacting with others (Canevello & Crocker, 2010; Crocker & Canevello, 2008). This association persists irrespective of other impactful relationship factors like attachment security and trust (Jiang et al., 2023).

Self-Image Goals

Self-image goals can encourage impression management to influence others’ perceptions of one’s relational value (Crocker & Canevello, 2008, 2018). However, self-image goals can backfire (Canevello & Crocker, 2011, 2015). In close offline relationships, these goals are linked to decreased responsiveness and relationship quality (Canevello & Crocker, 2010). Pursuing self-image goals can contribute to increased anxiety and lower responsiveness and support when interacting with others (Crocker, 2011).

Self-image goals can also moderate the impact of compassionate goals on social support and responsiveness. Crocker and Canevello (2008) found that high self-image goals undermined the positive impact of high compassionate goals in college roommates. When participants were low in self-image goals and high in compassionate goals, they had roommates who gave and received more support. However, when participants were high in self-image goals, there was no significant association between participants’ compassionate goals and the level of support their roommates received and reciprocated. In another study, the positive association between compassionate goals and responsiveness was greater when self-image goals were low compared to high (Canevello & Crocker, 2010). The effects of compassionate goals seem to be strongest when self-image goals are low, illustrating the importance of considering how these goals interact when examining responsive behaviours.

Interpersonal Goals in Social Media Contexts

Most research on interpersonal goals has focused on in-person interactions, but these goals also matter when using SNSs and are likely influenced by contextual and environmental factors. In-person interactions involve rich, real-time communication and reactions that can foster deep connections; however, they require proximity, can be time-consuming, and usually involve fewer people. SNS interactions are flexible and allow maintenance of a wide range of relationships; however, they lack crucial non-verbal cues and can be superficial (Lieberman & Schroeder, 2020).

Research on interpersonal goals in SNS contexts has examined a collection of active, passive (Tobin et al., 2020), and relationship maintenance behaviours, including responding to others when they share good or bad news or seek advice (Roper & Tobin, 2022). Like offline findings, compassionate goals predict more responsive behaviours on Facebook and, subsequently, more resources and benefits of social connections (Tobin et al., 2020) and well-being (Roper & Tobin, 2022). Responding to Facebook posts can serve as a signal of attending to others, thereby maintaining social connections, fostering reciprocity, and building social capital (Ellison et al., 2014). Compassionate goals also predict greater life satisfaction through increased social connectedness from actively attending to Facebook friends (Roper & Tobin, 2022). In contrast to offline findings associating self-image goals with less responsiveness (Canevello & Crocker, 2010; Crocker, 2011), self-image goals are not significantly associated with responsiveness on Facebook (Roper & Tobin, 2022; Tobin et al., 2020). Thus, interpersonal goals, particularly self-image goals, may operate differently in SNS contexts.

Initial investigations of interpersonal goals and responsiveness on SNSs are limited. First, they examined overall responsiveness without considering contextual variables (e.g., response modality). Second, they did not examine whether self-image goals moderated the impact of compassionate goals on responsiveness. Third, their correlational designs did not permit causal interpretations. Our study sought to address these limitations.

Relationship Strength and Maintenance on Facebook

Relationship strength can influence responsive behaviours. Strong ties signify close, intimate relationships distinguished by repetitive interactions, high support, and trust. Conversely, weak ties are distinguished by lower intimacy and interaction frequency, although they provide access to a broader network and resources (Granovetter, 1973).

Facebook allows maintenance of different relationship types through varied behaviours accompanied by distinct social rules (Bryant & Marmo, 2012). Bryant and Marmo (2012) found that relationship maintenance rules on Facebook apply to casual and acquaintance relationships more than close relationships. Close relationships involve intimacy and substantial interaction and comprise a small portion of Facebook friends. Instead, Facebook friends tend to be casual and acquaintance relationships. Casual relationships involve occasional interaction and some social support but lack intimacy, whereas acquaintance relationships are people who loosely know each other before connecting online (Bryant & Marmo, 2009, 2012).

Acquaintance relationships can be common on Facebook as it is a convenient environment to maintain these connections with minimal effort (Donath, 2007; Ellison et al., 2010). Nevertheless, Bryant and Marmo (2009) found that Facebook is a vital communication platform most used to maintain casual relationships. These connections are substantial enough for interaction on Facebook (e.g., commenting on posts) without seeming intrusive, whereas maintaining acquaintance relationships typically involves more surveillance-type behaviours (e.g., viewing posts without responding). Due to the unique prevalence, likelihood of responding, and high endorsement of relationship maintenance rules, the current study focused on how interpersonal goals shape responding to casual friends on Facebook.

Response Modalities

Response mode relates to responsiveness and relationship maintenance on SNSs (Burke & Kraut, 2016). SNS response options vary in publicness, effort, and implied relationship signals. Although research has considered interpersonal goals when investigating responsiveness on SNSs (Roper & Tobin, 2022; Tobin et al., 2020), these studies did not distinguish between different response modalities, an important factor for understanding how relational investment is signalled.

Offline, self-image goals predict less responsiveness (Canevello & Crocker, 2010), but SNSs offer a unique context where these goals can encourage responding to manage impressions with little effort. One-click reactions, specifically ‘liking’, can be strategically used for self-presentation (Lowe-Calverley & Grieve, 2018). S.-Y. Lee et al. (2016) also found that pleasing others, a characteristic of self-image goals (Crocker & Canevello, 2012), motivated responding by clicking like on Facebook. Similarly, factors linked to self-image goals (e.g., greater public self-consciousness (Crocker & Canevello, 2008) have also predicted ‘liking’ Facebook content (Hong et al., 2017). These findings suggest that self-presentation concerns, which are typically prominent when holding self-image goals (Crocker & Canevello, 2012), could prompt responsive behaviour on Facebook with the intention of managing impressions. This behaviour is more likely to involve consideration of the self rather than others’ needs and encourage less effortful responding. Accordingly, we predicted that:

When Facebook users hold high relative to low self-image goals, they are more likely to respond with a one-click reaction (H1).

When considering convenient, low-effort responses, like one-click reactions, different motivations can lead to the same behaviour. Lowe-Calverley and Grieve (2018) found that while people do consider reputation before ‘liking’ content on SNSs, they also see ‘liking’ as a way of providing social support, a typical component of compassionate goals (Canevello & Crocker, 2010; Crocker & Canevello, 2008). Furthermore, S.-Y. Lee et al. (2016) found that maintaining interpersonal relationships, a characteristic of compassionate goals (Crocker & Canevello, 2012), also motivated responding by clicking like on Facebook. While one-click reactions such as ‘likes’ may appear simplistic, they can be driven by complex motivations, such as maintaining relationships and offering social support, aligning with compassionate goals. Accordingly, we predicted that:

When Facebook users hold high relative to low compassionate goals, they are more likely to respond with a one-click reaction (H2a).

Conversely, research and theory have argued that interpersonal goals are associated with variations in responsiveness (Canevello & Crocker, 2010; 2017; Tobin et al., 2020). These variations reflect more effortful responding behaviours when holding compassionate goals, providing clearer signals of relational investment (Canevello & Crocker, 2017). D. S. Lee et al. (2021) found that compassionate goals encouraged perspective-taking, and in turn, the provision of support that better met people’s needs.

On Facebook, direct, targeted communication (e.g., comments, private messages) can more effectively signal relational investment and encourage reciprocity than public, non-targeted communications (e.g., status updates; Bazarova et al., 2012; Burke et al., 2011). Receiving targeted, written responses on SNSs, can improve well-being (Burke & Kraut, 2016), as the expenditure of time signals relationship commitment (Donath, 2007). If SNS users are motivated by compassionate goals to be responsive and supportive, and responding to casual friends on Facebook is an endorsed relationship maintenance strategy (Bryant & Marmo, 2012), responding with a comment or private message could signal increased relational investment. However, given that self-image goals can undermine the positive effects of compassionate goals (Crocker & Canevello, 2008), higher compassionate goals may lead to more effortful responding only if self-image goals are low. For example, someone motivated to support a casual friend posting about feeling down (compassionate goal) could comment with words of encouragement; however, if simultaneously they are concerned about being perceived negatively by others for this response (self-image goal), they may not respond in this way. Accordingly, we predicted that:

When Facebook users hold high relative to low compassionate goals, they are more likely to respond to a Facebook post with a comment. This effect of compassionate goals is stronger when users hold low relative to high self-image goals (H2b). When Facebook users hold high relative to low compassionate goals, they are more likely to respond to a Facebook post with a private message. This effect of compassionate goals is stronger, particularly when users hold low relative to high self-image goals (H2c).

Post Valence

The valence of SNS posts can affect whether people respond publicly. Positive compared to negative posts elicit more public responsive behaviours, such as one-click reactions and comments (Y. Lee et al., 2021; Vogel et al., 2018; Zhang et al., 2022). On Facebook, people are more likely to convey positive than negative emotions (Qiu et al., 2012). Reflective of this preference, people may be more likely to endorse a positive post publicly. Research investigating social support on Facebook, measured as a combination of clicking like and commenting, supported the notion that positive versus negative posts receive greater public endorsement (Vogel et al., 2018). Accordingly, we predicted that:

Facebook users are more likely to respond with a one-click reaction to positive than negative posts (H3a). Facebook users are more likely to respond with a comment to positive than negative posts (H3b).

While public responses are more likely for positive posts, private messages are more likely for negative posts (Zhang et al., 2022; Ziegele & Reinecke, 2017). Ziegele and Reinecke’s (2017) observation that Facebook users were less inclined to respond with a comment to a negative status update was partially attributed to the perceived inappropriateness of posting such content. However, perceptions of support urgency partially contributed to a greater willingness to respond privately. These findings highlight the intricate dynamics of responsive SNS behaviours, emphasising valence, modality, and norms in shaping interactions. Accordingly, we predicted that:

Facebook users are more likely to respond with a private message to negative than positive posts (H3c).

The Current Study

Understanding how interpersonal goals guide responsive SNS behaviours can expand knowledge about modifiable psychological processes that could enhance social connection. Offline research links compassionate goals to increased responsiveness and positive social and emotional outcomes, whereas opposing effects were found for self-image goals (Canevello & Crocker, 2010; Duarte & Pinto-Gouveia, 2015). Relatedly, compassionate goals have been associated with responsiveness on SNSs and, subsequently, greater social capital benefits (Tobin et al., 2020), social connectedness, and life satisfaction (Roper & Tobin, 2022). However, the causal effects of interpersonal goals on specific responsive SNS behaviours remain unknown. The current study aimed to extend past correlational findings and address research design limitations by contributing experimental evidence to understand how these goals influence responsiveness toward casual friends on Facebook. Gaining this understanding could expand SNS research by revealing what could motivate responsiveness on SNSs and broaden insights into how users signal relationship investment.

SNSs uniquely offer increased opportunities to connect with casual friends (Bryant & Marmo, 2009) and communicate broadly via posts that vary in valence and topic (Wang et al., 2013; Ziegele & Reinecke, 2017). On Facebook, it is common to share posts related to work and social experiences (Wang et al., 2013). To create stimuli that represent these common topics, the posts used in the current study reflect work and social-based domains. Unlike post valence, post domain variations represented content typically posted on Facebook and were not anticipated to influence responses. Hence, responses were to posts about two positive and two negative experiences, one related to work and one to friends.

The materials, data, and syntax for the studies reported in this manuscript can be found on the Open Science Framework (OSF) at https://osf.io/kuxb3/.

Pilot Study

A pilot study was conducted (after the main study) to confirm that the post stimuli used accurately conveyed the intended valence and were considered appropriate to post on Facebook.

Initially, 101 participants were recruited via Prolific; however, six were excluded for failing attention checks¹. The final sample (N = 95) included participants aged 18–71 (M = 35.63, SD = 11.78), with almost equal proportions from Australia (n = 47; 49.5%) and the United States (n = 48; 50.5%), and a nearly even representation of males (n = 47; 49.5%) and females (n = 48; 50.5%).

Participants evaluated the four posts used in the main study as part of a larger set, three reflecting each of the four combinations of post valence (positive, negative) and domain (work, social). Participants rated the perceived valence and appropriateness of each post using 7-point scales (1= very negative and 7= very positive; 1= very inappropriate and 7= very appropriate). Among these posts were the four posts used in the current study: feeling excited about winning an award at work (positive, work); having a great time at a party with friends (positive, social); feeling exhausted and unproductive at work (negative, work); missing a friend who passed away last year (negative, social). Ideally, we could hold appropriateness constant. However, Bazarova (2012) found that negative Facebook posts were perceived as less appropriate than positive posts. Consequently, it was anticipated that positive posts would be perceived as more appropriate to post than negative posts.

A series of 2 (valence: positive, negative) x 2 (domain: work, social) x 2 (country: Australia, United States) x 2 (gender: male, female) mixed model ANOVAs were run to test whether the hypothetical post stimuli accurately depicted the intended valence and were perceived as appropriate. When testing the perceived appropriateness and valence of the posts, we wanted to test whether the participant’s gender or country affected results, hence including these variables in the analyses. The within-subjects variables were valence and domain, with country and gender as between-subjects variables. Results showed only a significant main effect of valence on perceived valence, F(1, 91) = 1063.21, p < .001, whereby participants rated the negative posts as negative (M = 2.45, SE = 0.09) and positive posts as positive (M = 6.26, SE = 0.07).

Results for appropriateness indicated a significant main effect of valence F(1,91) = 184.69, p < .001, and domain F(1,91) = 121.63, p < .001, such that participants perceived positive posts (M = 5.95, SE = 0.10) as more appropriate than negative posts (M = 4.33, SE = 0.12); and social posts (M = 5.76, SE = 0.10) to be more appropriate than work posts (M = 4.52, SE = 0.12). Additionally, the valence x domain, F(1, 91) = 77.90, p < .001, and valence x domain x gender, F(1,91) = 6.74, p = .011, interactions were significant. Follow-up pairwise comparison of the valence x domain interaction cell means, detailed in Table 1, revealed that when posts had either a positive, p < .001, or negative valence, p = .019, participants rated social compared to work posts as significantly more appropriate. However, the difference was larger for negative posts. To follow up the significant valence x domain x gender interaction, we examined the simple effects of gender with pairwise comparisons. These analyses indicated no significant simple effects of gender. Collectively, each post was perceived to accurately reflect the intended appropriateness and valence (positive or negative). Notably, despite negative events being perceived as less appropriate to post about, the negative social post, missing a friend who passed away, was seen as more appropriate than the negative work post (see Table 1).

Table 1. Means and Standard Deviations for Perceived Valence and Appropriateness of Posts.

Methods

Participants

In December 2021, 402 Facebook users from the United States were recruited via the online research participation platform Prolific. This recruitment method was deemed appropriate as participants on Prolific are more diverse, naïve, less dishonest, and produce higher quality data than participants from other online crowdsourcing platforms (Peer et al., 2017). Participants had to have a Facebook account, be at least 18 years old, and be located in the United States to be eligible to participate. Following Prolific pay rates and the proposed 10-minute study completion time, participants received compensation of USD 1.20. Participants gave their informed consent, and approval was obtained from the Human Research Ethics Committees at the authors’ institutions.

The final sample included 389 Facebook users, as 13 participants were excluded for failing the attention checks. Participants’ ages ranged from 18–92 (M = 36.51, SD = 14.78). Nearly equal proportions identified as male (n =191; 49.1%) and female (n = 195; 50.1%), with three (0.8%) noting another identity. Most participants identified as Caucasian/White (n = 322; 79.3%), then subsequently, Asian (n = 49; 12.1%), African American/Black (n = 20; 4.9%), American Indian/Alaska Natives (n = 7; 1.7%), other (n = 7; 1.7%) and Native Hawaiian/other Pacific Islander (n = 1; 0.2%). A further 25 participants (6.4%) indicated their ethnicity as Hispanic/Latino and 380 (97.7%) were native English speakers. Within the sample, 155 (39.8%) indicated their active Facebook use was less than 10 minutes per day, 147 (37.8%) reported 10–30 minutes per day, 48 (12.3%) reported 31–60 minutes per day, 30 (7.7%) reported 1–2 hours per day, 3 (0.8%) reported 2–3 hours per day and 6 (1.5%) reported more than 3 hours per day.

Experimental Design

The current study was a 3 (mode: one-click, comment, private message) x 2 (self-image goals: high or low) x 2 (compassionate goals: high or low) x 2 (post valence: positive or negative) x 2 (post domain: work or social) mixed-model design. Interpersonal goals (compassionate and self-image goals) were manipulated between-subjects and mode (one click, comment and private message) and scenario variables (valence and domain) within-subjects. Participants were randomly assigned a set of interpersonal goals, which they consistently held while responding to four hypothetical Facebook posts presented in a random order. Unlike post valence, which was anticipated to influence responding behaviours, post domain variations were included to mimic typical Facebook posts. Comparisons of these post domains were exploratory and included to potentially provide insights for future research. The dependent variables were the likelihood of responding with a one-click reaction, comment, and private message.

Procedure

Participants took part in the online experiment using Qualtrics survey software. After participants were verified to be in the United States, they needed to pass a Captcha test before they could begin the experiment.

Participants were informed that they would be presented with four possible Facebook posts, and they should imagine they are seeing them when looking through their Facebook News Feed. To hold constant the strength of the relationship, participants were asked to imagine the posts were from friends whom they occasionally interact with (i.e., casual friends). Participants were randomly assigned to one of the four goal conditions. Depending upon the condition, participants were given instructions to adopt different motivations when responding to the posts. They were asked to imagine: that when you respond, you are motivated to support other people’s well-being. You also are motivated to convey a desirable image of yourself to other people (high compassionate, high self-image goals condition); that when you respond, you are motivated to support other people’s well-being. However, you are NOT motivated to convey a desirable image of yourself to other people (high compassionate, low self-image goals condition); when you respond, you are NOT motivated to support other people’s well-being. However, you are motivated to convey a desirable image of yourself to other people (low compassionate, high self-image goals condition); or when you respond, you are NOT motivated to support other people’s well-being. You also are NOT motivated to convey a desirable image of yourself to other people (low compassionate, low self-image goals condition). Participants were presented with these instructions and reminded of their motivations before being shown each post.

Participants answered the same three questions about the likelihood of responding via a one-click reaction, a comment, and a private message when presented with each hypothetical Facebook post. They further indicated what they would select from seven one-click reactions and what they would write in a comment and private message². Participants then completed the manipulation checks, attention checks, and demographic questions. Lastly, participants were debriefed and thanked for their participation.

Measures

Response Likelihood

For each hypothetical Facebook post, participants rated their agreement on a 5-point Likert scale (1 = not at all likely and 5 = very likely) with: How likely would you be to click on one of the reaction options?, How likely would you be to write a comment? and How likely would you be to send a private message?. Although the researchers developed these items, other studies have used hypothetical scenarios (Gilbert et al., 2019; Malloch & Feng, 2022) and similar scales to measure SNS responding behaviours (Malloch & Feng, 2022; Ziegele & Reinecke, 2017).

Interpersonal Goal Manipulation Checks

Using a 5-point Likert scale (1 = not at all and 5 = to a great extent), participants responded to two manipulation check items related to goals: To what extent were you trying to support other people’s well-being through your responses? and To what extent were you trying to convey a desirable image of yourself to other people through your responses?.

Attention Checks

Following Newman and colleagues’ (2021) suggestions, four attention checks³ were included to identify inattentive participants, enhance statistical conclusions and internal validity. Participants needed to pass all attention checks to be included in the final sample.

Results

Table 2 presents goal condition sample sizes, descriptive statistics for the manipulation checks, and response likelihood ratings for each scenario.

Table 2. Means and Standard Deviations for Manipulation Checks and Response Likelihood Within Interpersonal Goal Conditions.

Manipulation Checks

Compassionate Goal

A 2 (self-image goal) x 2 (compassionate goal) between-groups factorial ANOVA on the compassionate goal manipulation check revealed only a significant main effect of compassionate goals, F(1, 385) = 153.50, p < .001, η_p² = .285. Specifically, participants in the high compassionate goal condition (M = 4.11, SD = 0.96) reported trying to support the well-being of others in their responses more than participants in the low compassionate goal condition (M = 2.58, SD = 1.44). These results indicate that the manipulation of compassionate goals was effective.

Self-Image Goal

A 2 x 2 between-groups factorial ANOVA on the self-image goal manipulation check revealed a significant main effect of self-image goals, F(1, 385) = 202.38, p < .001, η_p² = .345, such that participants in the high self-image goal (M = 3.42, SD = 1.41) condition reported trying to convey a desirable image of themselves to others more than participants in the low self-image goal condition (M = 1.67, SD = 1.04). There was also a significant interaction effect of compassionate and self-image goals, F(1, 385) = 17.44, p < .001, η_p² = .043. Follow up analyses of the simple effects compared cell means detailed in Table 2. These comparisons indicated that self-image goals significantly increased the extent participants tried to convey a desirable image of themselves in their responses in the high compassionate goal, F(1, 385) = 50.64, p < .001, η_p² = .12, and also the low compassionate goal conditions, F(1, 385) = 168.85, p < .001, η_p² = .31. These results indicate that participants in the high self-image goal condition tried to convey a more desirable image of themselves in their responses more than those in the low self-image goal condition, despite compassionate goal condition. Notably, a larger effect of self-image goals was detected in the low compassionate goal condition. Overall, these results show that the manipulation of self-image goals was effective.

Main Analyses

A 3 (mode) x 2 (self-image goal) x 2 (compassionate goal) x 2 (valence) x 2 (domain) mixed model ANOVA tested how interpersonal goals affected the likelihood of responding with a one-click reaction, comment, and private message to hypothetical Facebook posts. ANOVA test results are detailed in Table 3. Due to violating the assumption of sphericity, tests of within-subjects effects used the Huynh-Feldt adjusted degrees of freedom. Results revealed significant main effects of mode, compassionate goals, and self-image goals, and a number of significant interactions. Because the interpretation of lower-order components is qualified by higher-order interactions, we focus on breaking down the highest-order interactions: mode x compassionate goals x self-image goals, mode x valence x compassionate goals, and mode x domain x valence.

Table 3. ANOVA Results for Response Likelihood by Interpersonal Goals, Response Mode, Post Domain and Post Valance.

Interpersonal Goals and Response Mode as Predictors of Response Likelihood

The mode x compassionate goals x self-image goals interaction is shown visually in Figure 1.

Figure 1. Interaction Effects of Response Mode, Compassionate Goals and Self-Image Goals on Response Likelihood.

Note. Error bars denote the standard error of the mean.

Follow up pairwise comparisons indicated that participants in the low self-image goal condition were significantly more likely to respond with a one-click reaction when they held high versus low compassionate goals, M_diff = 0.56, SE = 0.16, p < .001, 95% CI [0.25, 0.86]; however, in the high self-image goal condition, there was no significant effect of compassionate goals, M_diff = 0.15, SE = 0.15, p = .922, 95% CI [−0.29, 0.32]. Additionally, participants in the low compassionate goal condition were significantly more likely to respond with a one-click reaction when they held high versus low self-image goals, M_diff = 0.40, SE = 0.16, p = .010, 95% CI [0.10, 0.71]; however, in the high compassionate goal condition, there was no significant effect of self-image goals, M_diff = −0.14, SE = 0.15, p = .356, 95% CI [−0.45, 0.16].

Similar to responding with a one-click reaction, follow up pairwise comparisons indicated that participants in the low self-image goal condition were significantly more likely to respond with a comment when they held high versus low compassionate goals, M_diff = 0.44, SE = 0.13, p = .001, 95% CI [0.19, 0.69]; however, there was no significant effect of compassionate goals in the high self-image goal condition, M_diff = −0.57, SE = 0.13, p = .653, 95% CI [−0.30, 0.19]. Furthermore, participants in the low compassionate goal condition were significantly more likely to respond with a comment when they held high versus low self-image goals, M_diff = 0.49, SE = 0.13, p < .001, 95% CI [0.24, 0.74]; however, in the high compassionate goal condition, there was no significant effect of self-image goals, M_diff = −0.01, SE = 0.13, p = .965, 95% CI [−0.25, 0.24].

Unlike one-click reactions and comments, there were no significant effects of interpersonal goals on the likelihood of responding with a private message.

Collectively, these results partially supported H1, H2a and supported H2b. Either self-image or compassionate goals at low levels of the opposing goal increased the likelihood of one-click reactions and comments. However, goals did not influence private messages.

Interpersonal Goals, Response Mode and Post Valence as Predictors of Response Likelihood

The mode x valence x compassionate goals interaction is shown visually in Figure 2.

Figure 2. Interaction Effects of Response Mode, Post Valence and Compassionate Goals on Response Likelihood.

Note. Error bars denote the standard error of the mean.

Follow up pairwise comparisons indicated that when participants held high versus low compassionate goals, they were significantly more likely to respond to positive Facebook posts with one-click reactions, M_diff = 0.36, SE = 0.12, p = .003, 95% CI [0.12, 0.59]; however, when responding to negative posts, there was no significant effect of compassionate goals, M_diff = 0.22, SE = 0.12, p = .074, 95% CI [−0.02, 0.46]. Similarly, when participants held high versus low compassionate goals, they were significantly more likely to respond to positive Facebook posts with comments, M_diff = 0.21, SE = 0.10, p = .034, 95% CI [0.17, 0.41]; however, when responding to negative posts, there was no significant effect of compassionate goals, M_diff = 0.17, SE = 0.10, p = .103, 95% CI [−0.03, 0.37]. Additionally, when holding high versus low compassionate goals, participants were more likely to respond to negative Facebook posts with private messages, M_diff = 0.24, SE = 0.10, p = .021, 95% CI [0.36, 0.44]; however, when responding to positive posts, there was no significant effect of compassionate goals, M_diff = 0.08, SE = 0.08, p = .370, 95% CI [−0.09, 0.24].

Collectively, these results partially supported H3a, H3b, H3c, and H2c. One-click reactions and comments were more likely when responding to positive Facebook posts, while private messages were more likely when responding to negative Facebook posts. Although, these effects were found when participants had compassionate goals.

Response Mode, Post Domain and Post Valance as Predictors of Response Likelihood

The mode x domain x valance is shown visually in Figure 3. Differences in post domain were not initially anticipated to influence responding behaviours. However, once analyses were conducted, unexpected findings revealed post domain effects.

Pairwise comparisons revealed that when responding to work-based posts, participants were significantly more likely to respond with a one-click reaction when the post had a positive versus a negative valence, M_diff = 1.31, SE = 0.08, p < .001, 95% CI [1.16, 1.46]. When responding to social-based posts, participants were significantly less likely to respond with a one-click reaction when the post had a positive versus a negative valence, M_diff = −0.74, SE = 0.07, p < .001, 95% CI [−0.89, −0.60]. Similarly, participants were significantly more likely to respond with a comment to a work-based post with a positive versus a negative valence, M_{diff =} 0.90, SE = 0.07, p < .001, 95% CI [0.76, 1.04]. Additionally, participants were significantly less likely to respond with a comment to a social-based post with a positive versus a negative valence, M_diff = −0.85, SE = 0.07, p < .001, 95% CI [−0.98, −0.71]. Lastly, participants were significantly more likely to respond with a private message to a social-based post with a negative versus positive valance, M_diff = 0.91, SE = 0.07, p < .001, 95% CI [0.77, 1.04]. However, there was no significant difference in responding with a private message to a work-based post with a negative versus a positive valence, M_diff = −0.10, SE = 0.06, p = .117, 95% CI [−0.22, 0.03].

Figure 3. Interaction Effects of Response Mode, Post Valance and Post Domain on Response Likelihood.

Note. Error bars denote the standard error of the mean.

Collectively, these results partially supported H3a and H3b and supported H3c. One-click reactions and comments were more likely for positive Facebook posts, but only in the work domain. Responding with a private message was more likely for negative Facebook posts, but only in the social domain. Surprisingly, one-click reactions and comment responses were also more likely for the negative-social post.

Discussion

Facebook can be used as a tool to maintain casual relationships (Bryant & Marmo, 2009). Responding to others’ Facebook posts can signal relationship investment and play a crucial role in relationship outcomes and well-being (Burke & Kraut, 2014, 2016; Ellison et al., 2014). The current study aimed to understand what motivates people to engage in specific responsive behaviours on Facebook. Results showed that the likelihood of responsive behaviours can be influenced by the interpersonal goals that people hold. Specifically, both experimentally induced compassionate and self-image goals predicted the likelihood of responding with one-click reactions and comments, although having both goals together did not further increase responsive behaviours. Additionally, whether a post was positive or negative was also meaningful. Compassionate goals predicted publicly responding (one-click reactions and comments) to positive posts and responding with private messages to negative posts. These findings highlight that in an experimental context, when instructed to hold interpersonal goals and imagine interacting with a casual friend on Facebook, both goals provoke responsiveness, albeit guided by different intentions.

Our findings revealed a significant interaction effect of response mode, compassionate goals, and self-image goals on the likelihood of responding. Either self-image or compassionate goals, at low levels of the opposing goal, increased the likelihood of one-click reactions, partially supporting H1 and H2a. Similarly, supporting H2b, holding high compassionate at low self-image goals, increased the likelihood of responding with comments. Unexpectedly, holding high self-image goals at low compassionate goals also increased the likelihood of responding with comments. Differing from H2c predictions, high compassionate at low self-image goals did not significantly influence the likelihood of responding with a private message.

Although either goal initially appeared to lead to the same responsive behaviours, when considering post-valence, compassionate goals were found to be more influential. A significant compassionate goal, response mode, and post valence interaction effect was observed. Being instructed to hold high versus low compassionate goals increased the likelihood of responding with one-click reactions and comments to positive posts and with private messages to negative posts. These findings partially supported H2c, H3a, H3b, and H3c.

A significant response mode, post valence, and domain interaction effect was also observed on the likelihood of responding. The positive work-related post elicited more one-click reactions and comments than the negative work-related post, partially supporting H3a and H3b. In contrast, the negative social post elicited more private messages than the positive social post, partially supporting H3c. Unexpectedly, the negative social post also elicited more one-click reactions and comments. We next consider these findings in more detail.

Interpersonal Goal Effects on Responding

Participants were more likely to respond to the specified Facebook posts with one-click reactions or comments when they were instructed to hold either a self-image or a compassionate goal, with no additive effect of the goals. In relationship research, responsiveness involves other-focused emotional and behavioural aspects, including expressing attentiveness, empathy, and a willingness to provide support. On the other hand, barriers to responsiveness can arise when people become self-focused, limiting their ability to respond (Reis & Clark, 2013; Reis et al., 2004). Although the interaction effects we observed were similar to previous offline findings (Crocker & Canevello, 2008), they differed in interesting ways. Our results suggest that the effects of self-image or compassionate goals can be influenced by opposing goal levels, highlighting the complex dynamics of interpersonal goals. However, unlike Crocker and Canevello’s (2008) findings considering face-to-face settings, we observed a relatively high level of responding among those instructed to hold high self-image goals, regardless of their instructed compassionate goal level.

Pleasing others and maintaining relationships, motivations associated with self-image and compassionate goals, respectively (Crocker & Canevello, 2012), have both predicted one-click reaction responses on Facebook (S.-Y. Lee et al., 2016). Moreover, responding by leaving a clear visual trace (e.g., a one-click reaction or comment) is the most reliable method to acknowledge posted content (Ellison et al., 2014). Acknowledging and attending to posted content publicly could serve self-image and compassionate goals differently. For example, people prioritising self-image goals may respond with a comment to appear desirable to others and manage their impressions, whereas those prioritising compassionate goals may also comment but to support others’ well-being. The ease and visibility of public responding within SNS contexts, particularly the imagined SNS context of the experiment, may have contributed to the relatively high responsive behaviours observed when holding either interpersonal goal. Furthermore, perhaps being instructed to have both goals led to confusion or ambivalence about how and whether to respond.

It could be that responding in a controlled SNS context was easier than noticing and responding to people’s needs in face-to-face situations. However, unlike the current experiment, it is arguably not necessarily a common practice to be instructed to convey a desirable image of yourself or support others when responding to Facebook posts. These are considerations to keep in mind when interpreting the current study results. Although our approach has experimental merit and provides insight, it lacks ecological validity, and influential aspects present outside the current experimental conditions may lead to varied behaviours and interpretations.

When instructed to be self-focused and motivated by self-image goals, responding to the posts detailed in the experiment with one-click reactions and comments could have contributed to impression management and achieving self-image goals. Similarly, being instructed to be other-focused and motivated by compassionate goals could have contributed to the same behaviours to support others. Or these behaviours could have been an artifact of the experimental conditions. During the experiment, participants were focused on a specific post and deliberated over whether and how to respond, a scenario that may not necessarily reflect how people realistically interact with posts on Facebook. Additionally, we manipulated the goals participants held during this process. In particular, our instructions focused on approaching the goals to ensure clarity (i.e., to convey a desirable image) and did not specify avoidance aspects (i.e., avoid showing weaknesses). Self-image and compassionate goals can be both approach and avoidance (Crocker & Canevello, 2018). Participants were not instructed to consider self-image or other-oriented risks in our experiment, aspects that people holding self-image or compassionate goals might naturally consider. The experimental conditions enabled control, however, they may influence results and not necessarily reflect more naturalistic conditions. Nevertheless, our findings suggest that responsiveness may differ in SNS contexts, particularly when motivated by self-image goals. Overall, these insights deepen our understanding of interpersonal goals and what can drive responsive behaviour online.

Notably, when participants were instructed to hold either self-image or compassionate goals, neither appeared to affect the likelihood of sending a private message across all posts used in the study. This result could be due to the perceived intimacy of sending private messages. Bazarova (2012) explained that Facebook users recognised private versus public disclosures as more intimate. Responding to a Facebook post privately (e.g., sending a private message) versus publicly (e.g., clicking like or commenting) may signal differences in relationship closeness (Donath, 2007; Ellison et al., 2014). In the current study, participants were asked to imagine responding to posts from casual friends to simulate a typical scenario on Facebook. Causal friendships on Facebook are seen as less intimate than close friendships (Bryant & Marmo, 2009, 2012), and this perception may have contributed to participants’ low likelihood of wanting to respond with a private message. Generally, the mean likelihood of responding to each post with a private message indicated that ratings tended to cluster at the lower end of the scale. These relatively low means indicate a possible floor effect, which would have made it difficult to detect differences in private message responses guided by varied interpersonal goals without considering other contextual factors. However, when interacting with casual friends on Facebook, it is acceptable to provide support and respond using a variety of modes (Bryant & Marmo, 2009). In the current study, subsequent investigations indicated that considering contextual specifications, like post valence, could provide further insight.

Post Valence Effects on Responding

When considering post valence, effects of compassionate goals were revealed in the current experiment. Participants instructed to hold compassionate goals indicated to be more likely to respond with one-click reactions and comments to the positive posts detailed in the experiment and with more private messages to the negative posts. Compassionate goals aim to support the well-being of others (Crocker & Canevello, 2008). Providing support to casual friends on Facebook is considered appropriate (Bryant & Marmo, 2009). Thus, our findings suggest that when instructed to hold high versus low compassionate goals, positive posts could be celebrated via public endorsement (one-click reactions and comments), while negative posts may signal a need for support and elicit connection via private messages. These results extend past findings correlating compassionate goals with increased responsiveness offline and responsive behaviours online (Canevello & Crocker, 2010; Tobin et al., 2020) by providing causal evidence, albeit within the specified experimental context.

These results also provide some support and extend past research indicating that positive Facebook posts can elicit more public responsive behaviours and negative posts can elicit more private messages (Y. Lee et al., 2021; Vogel et al., 2018; Zhang et al., 2022). Interestingly, these findings also diverge from previous research that negative Facebook statuses have been perceived as inappropriate to post, and only close friends are likely to respond with a private message to such content (Ziegele & Reinecke, 2017). Patterns of responding may be explained not only by post valence but also by the presence of compassionate interpersonal goals. Notably, these conclusions are presented within the confines of the current study in which compassionate goals were experimentally induced rather than spontaneously occurring and may not necessarily reflect real-world relationships or SNS behaviours. Accordingly, our findings and associated interpretations may not generalise to more naturalistic interactions on SNSs. However, results show that within the current experimental conditions, compassionate goals can inform the likelihood of varied responsive behaviours to hypothetical positive and negative posts by casual friends.

The current study provided additional insights into responsive SNS behaviours. Differences in post domain were intended to provide contextual variations aimed at mimicking typical Facebook posts and were not predicted to impact the likelihood of responding. Unexpectedly, analyses revealed the influence of post domain. Participants reported a likelihood of responding with more one-click reactions and comments when confronted with a positive relative to a negative work-based post, and more private messages when confronted with a negative relative to a positive social-based post. These results initially appeared to be consistent with previous research indicating that positive posts increase the likelihood of responding with a one-click reaction or comment, and negative posts increase the likelihood of responding with a private message (Vogel et al., 2018; Ziegele & Reinecke, 2017). Surprisingly, one-click reactions and comment responses were also more likely when responding to a negative versus a positive social post. Thus, our results appeared to be context dependent and suggest that participants considered other implied nuances in the posts presented beyond valence when deciding how and whether to respond.

The unexpected influence of post domain on responding could be explained by the perceived appropriateness of the posts, as well as the perceived need for support indicated in the negative social post. Supporting Bazarova’s (2012) finding that negative Facebook posts were perceived as less appropriate than positive ones, the positive work post about winning an award at work was rated high on appropriateness in the pilot study (see Table 1). However, unlike Bazarova’s (2012) findings, the negative social post about missing a friend who had passed away was also rated rather high on appropriateness (see Table 1). These perceptions of appropriateness could indicate that people are more accustomed to seeing and responding to such posts, which may have contributed to the responding behaviours observed. Additionally, in the case of responding to the negative social post, perhaps the perceived need for support alongside the appropriateness of specifically posting about missing a friend who has passed away may help explain our findings. Even though none of the post stimuli used in the current experiment explicitly asked for support, certain posts may communicate support needs without specific requests. On SNSs, asking for support can elicit greater responsive behaviours (Wang et al., 2013). It could be that posting about missing a friend who has passed away may imply a greater need for support and elicit increased responsive behaviours by all response modes. Thus, beyond post valence, contextual factors, such as post domain, may influence the likelihood of responding to casual friends on Facebook.

Theoretical Implications

Our findings have theoretical implications and extend knowledge about interpersonal goals and responsive SNS behaviours. The current study answers calls for research seeking to overcome the limitations of correlational designs when examining the influence of interpersonal goals on behaviour (Crocker & Canevello, 2012; Tobin et al., 2020) and explores the dynamics of more distant relationships on SNSs (Roper & Tobin, 2022). To the author’s knowledge, the current study provides the first evidence of the causal effects of experimentally induced interpersonal goals on the likelihood of responsive behaviours within a hypothetical SNS context. Our findings help to clarify the influence of interpersonal goals on responsive behaviours. Specifically, our findings suggest that compassionate and self-image goals may interact differently in online versus offline settings, in that both goals were found to increase responsive behaviours (Crocker & Canevello, 2008).

The current study further contributes to discussions exploring the positive aspects and complexities of SNS use by providing possible explanations for variations in responsive behaviours. Verduyn and colleagues’ (2021) extended active-passive model highlights the well-being benefits of warm, targeted active SNS use (e.g., supportive comments on a post). It was out of the scope of the current study to analyse whether responses were warm (e.g., agreeable) or cold (e.g., quarrelsome; Verduyn et al., 2021). However, at a broad level, our findings suggest that when instructed to hold either compassionate or self-image goals, this can lead to targeted active SNS use (i.e., increased public responsive behaviours to posts). Thus, by exploring how interpersonal goals could affect SNS responding, we have contributed valuable insights into psychological processes that may influence targeted active SNS behaviours, which have the potential to impact well-being. Nevertheless, these findings are limited by the conditions defined in the experiment, and further research is needed to guide more specific inferences regarding how interpersonal goals inform SNS behaviours and impact well-being in more naturalistic settings.

Limitations and Future Research Directions

The current study required participants to imagine holding different interpersonal goals and respond to predetermined posts from casual friends. While this approach enabled control, it limited ecological validity. Previous research has used scenario-based designs where participants were asked to imagine holding certain goals (Bates et al., 2023; Singer & Dobson, 2007). Experimental methods that prompt participants to imagine specific scenarios can simulate real-world situations, allowing researchers to study abstract factors that are difficult to manipulate directly (Lutz et al., 2024; MacGregor et al., 2017; Zhang et al., 2022). For instance, studies have used this approach to examine the causal link between fear of missing out (FOMO) and social media use (David & Roberts, 2023), as well as how interpersonal goals shape social media posts (Tobin et al., 2025). Similarly, instructing participants to imagine holding certain interpersonal goals and controlling what they respond to allowed the current study to draw conclusions about causal effects of these goals on responsive Facebook behaviours. However, current experiment restrictions may have influenced participants’ behaviour and may not necessarily be an accurate representation of interpersonal goal-directed behaviours in a more naturalistic setting. Although the results of the manipulation checks indicate that the manipulation used was effective in changing interpersonal goals, it remains unclear if participants’ responses in the hypothetical scenarios accurately reflect how they may behave when they pursue interpersonal goals in the real world. For example, perhaps the natural consideration of both avoidance as well as approach aspects of interpersonal goals could have influenced behaviours. Future studies should examine whether daily fluctuations in goals driven by intrinsic and contextual factors produce similar effects. Researchers could consider conducting daily diary studies to examine whether the effects generalise with naturally occurring state fluctuations in interpersonal goals.

Furthermore, future experimental studies exploring whether interpersonal goals influence SNS responding behaviours could consider responses to similar posts from close and casual friends. Limiting the study to instructing participants to imagine responding to posts by causal friends may have restricted our ability to detect certain effects and gain deeper insights into the influence of interpersonal goals in SNS contexts. Since Facebook is commonly used as a tool to maintain casual relationships (Bryant & Marmo, 2009), participants were asked to imagine responding to someone they interacted with occasionally. However, participants may have found it challenging to imagine someone they interact with occasionally sharing the posts they were instructed to respond to within the experiment. We also cannot be certain whether participants interpreted the posts as being posted by a casual friend, which could influence the interpretation of results. Moreover, while we anticipated self-image goals to increase responding in some capacity, outcomes may have been different if we had examined responding to close friends. Holding self-image goals might lead to some responsiveness with a casual friend, but like patterns found in offline studies of close relationships (Canevello & Crocker, 2010), they could lead to lower responsiveness when interacting with close friends on Facebook. Similarly, responding to posts by close friends with private messages across all posts could have been detected for compassionate goals. Nevertheless, researchers may face challenges manipulating interpersonal goals for interactions with close friends. Instructing participants to imagine adopting particular goals may be challenging because people may have clear, pre-existing goals for these relationships. Future studies could consider other ways to manipulate interpersonal goals that are more subtle and ecologically valid.

The content of the Facebook post stimuli may also limit our findings’ generalisability. Post stimuli were tested for appropriateness and reflected commonly posted SNS content (Wang et al., 2013). However, the observed responding behaviours in the current study may have reflected responding to implied support-seeking requests (e.g., missing a friend who passed away) or dismissing posts that appeared less appropriate to post about (e.g., feeling exhausted and unproductive at work). Additionally, while the specific negative social post used in this study may explain the increased likelihood of responding with all modalities, this interpretation cannot necessarily be verified without testing a larger sample of posts. It would be beneficial for future experimental studies to instruct participants to respond to multiple positive and negative posts within the same domain to assess the causal determinants of responsive behaviours on SNSs comprehensively. Instructing participants to respond to at least two posts for each condition would increase control. An accumulated response measure could allow greater insight into how interpersonal goals, post valence, and domain influence SNS behaviours.

Conclusion

The current study extends interpersonal goal and SNS behaviour research by contributing to understanding the causal effects of interpersonal goals on Facebook responding behaviours. The results illustrated that in experimental settings, responsive behaviours can be influenced by the interpersonal goals people are instructed to hold and SNS contextual factors. Both experimentally induced self-image and compassionate goals were found to increase the likelihood of responding publicly (i.e., one-click reactions or comments). Compassionate goals also increase the likelihood of responding publicly when confronted with positive posts and privately when confronted with negative posts. Within the confines of the current experiment, when people decide whether to respond to Facebook posts by a casual friend, assigned interpersonal goals were found to be influential, potentially alongside the perceived intimacy and appropriateness of the response modality. All active, responsive behaviours tested were also post valence and context dependent. Overall, our findings provide initial causal evidence that in an experimental context, both induced self-image and compassionate goals can increase the likelihood of responding to SNS posts, contributing insights into the existing interpersonal goal and SNS literature.

Footnotes

¹ Participants completed three attention checks. These checks included responding with very negative to click very negative as your answer to this item, responding with work to click work as your answer for this item and responding with very appropriate to click very appropriate as your answer to this item. To be included in the study, participants needed to pass all three attention checks.

² We initially planned to analyse these responses, however, there were several issues with the quality and amount of data collected. Not all participants provided an appropriate comment or private message response, resulting in a substantial amount of missing data. Additionally, a considerable amount of the responses collected were short or contained a single word, and complexities surrounding the variation in post stimuli presented further challenges for analyses. These issues were not initially anticipated, however, presented considerations for future research.

³ The four attention checks included: computing participant’s age in years and year of birth; answering his son to The man couldn’t lift his son because he was so heavy. Who was heavy?; answering a little to If you are reading this statement, please select “a little” as your answer and assessing how seriously they responded to the experiment questions via their indication on a 5-point scale (1 = not seriously at all and 5 = extremely seriously). A rating of less than three was deemed insufficient attention.

Conflict of Interest

The authors have no conflicts of interest to declare.

Use of AI Services

The authors declare they have not used any AI services to generate any part of the manuscript or data.