1 Introduction
Imagine participating in a survey, where researchers present you with an advertisement for a car. Next, they ask you to envision this car coming alive and as a person. How difficult would it be for you to comply with these instructions? After all, cars are neither alive nor human. Would this difficulty impact your subsequent responses to other survey questions?
In marketing research, anthropomorphism is frequently induced through a cognitive task that instructs participants to imagine a product/brand (e.g.,Aggarwal & McGill, 2012; Chandler & Schwarz, 2010; Chen et al., 2018; Fronczek et al., 2023; Kim & Kramer, 2015; Zhou et al., 2019) or abstract ideas (Chen et al., 2020) as alive and human. Despite its popularity, this cognitive task poses two research problems that have never been empirically addressed, nor solved.
Advertisement
The first problem concerns the conflation of two distinct thought processes: animism (perceiving an object as alive) and anthropomorphism (perceiving an object as a person). This conflation bears significant implications for an experimental manipulation because instructing respondents to perceive an object as alive does not inherently imply that they would consider the object as human or human-like. Yet, no research has addressed the question of how harmful/beneficial distinguishing anthropomorphism from animism can be for marketing research, and what consequences one might expect after disentangling manipulations of anthropomorphism from animism in experiments.
The second research problem pertains to the nature of the cognitive task itself, which is essentially an associative projective technique and some scholars (e.g., Avis, 2011) argue that it does not assess how people ordinarily perceive products or brands. Indeed, without any visual or verbal cues that link a product to human characteristics, envisioning a product as human can become difficult, making the task seem artificial and coerced rather than natural and cognitively fluent. Although respondents are kind to comply with instructions provided by researchers, even assigning human personalities to such unconventional objects as rocks (Avis et al., 2014), this coercion may hinder the actual marketing effects because it disrupts the cognitive ease at which words and associations are retrieved from memory.
In three studies, I address these two research problems and empirically explore how manipulations of anthropomorphism versus animism through cognitive tasks influence task-related performance and subsequent persuasive outcomes. My work contributes to both theory and marketing practice in several ways. First, I provide empirical evidence that tasks manipulating animism and anthropomorphism elicit distinct cognitive processes. Second, I demonstrate that animism leads to different task performance than anthropomorphism, as evidenced by the length of product-related descriptions participants spontaneously generate. Third, I show that manipulating anthropomorphism (vs animism) with a cognitive task deflates persuasion. Fourth, I identify a boundary condition for this effect: embedding many human-like features into the product equates the persuasive impact of anthropomorphism with animism. In sum, my findings can help marketing researchers in refining their experimental manipulations and enable them uncover marketing effects that might otherwise remain unnoticed.
2 Theoretical background
2.1 Manipulating anthropomorphism
Anthropomorphism is a frequently examined phenomenon in marketing research, which has already brought a few meta-analyses (Blut et al., 2021; Velasco et al., 2021) along with numerous empirical (e.g., Ding & Xu, 2023; S. Y. Kim et al., 2019; Kumar et al., 2022) and conceptual studies (Hermann, 2022). The essence of anthropomorphism is to see a human in objects or abstract ideas. It involves the perception of traits that are typical for homo sapiens, but not for other species, such as morphological structure (e.g., a face) and speech. Thus, the existing manipulations of anthropomorphism often rely on either visual or verbal cues (MacInnis & Folkes, 2017) and researchers either induce anthropomorphism through physical features that visually resemble human characteristics (e.g., a face) or through language (e.g., a first-person perspective, "I am Lucy").
Advertisement
Despite the prevalence of these two approaches, there exists a third method for manipulating anthropomorphism. It entails a cognitive task that instructs participants to imagine a target object as alive and human, subsequently prompting them to describe its physical appearance and demeanor. It is based on the assumption that no visual or verbal human-like cues are needed to activate anthropomorphism because this mode of thinking can be applied to “anything that acts with apparent independence” (Epley et al., 2007, p. 865). Although this task has raised doubts concerning its coercive nature (Avis, 2011), it has excessively been employed in various marketing studies (e.g., Aggarwal & McGill, 2012; Chandler & Schwarz, 2010; Chen et al., 2018, 2020; Fronczek et al., 2023; H. C. Kim & Kramer, 2015; Zhou et al., 2019).
2.2 Manipulating animism
Animism denotes the extent to which we see a target object as alive, and there are initial indications that animism and anthropomorphism should be treated as distinct phenomena (e.g., Blut et al., 2021, include anthropomorphism and “animacy” as two separate serial mediators into their meta-analytical model). The distinction between animism and anthropomorphism is commonly drawn in rhetoric, biology, psychology or religious thought (Epley et al., 2008; Guthrie, 1995; McDonald & Stuart-Hamilton, 2000; Piaget, 1992), but marketing researchers use them interchangeably or often conflate both concepts (e.g., Aggarwal & McGill, 2007; Chandler & Schwarz, 2010; Kwak et al., 2020). Theoretically, it makes sense to combine anthropomorphism with animism because humans constitute a living species. However, being alive does not necessarily imply humanness (e.g., plants and animals are alive, but they are taxonomically different from homo sapiens). In fact, animism and animalistic traits are often used to denote dehumanization (see political research: Cassese, 2021).
There has been relatively little research conducted specifically on animism, and manipulating animistic mode of thinking still remains an underexplored area. From the limited pool of experiments available, one may only infer that movement can successfully induce animism, regardless of the object's visual appearance. For example, people have been observed to attribute life to moving dots and rectangles (Tremoulet & Feldman, 2000), animated visuals (Santos et al., 2008), moving sounds (Nielsen et al., 2015), and moving products (Karpinska-Krakowiak & Eisend, 2021). Yet, there have been no studies that directly manipulate animism versus anthropomorphism and compare their resulting effects.
2.3 Anthropomorphism, animism and elaboration
Manipulating anthropomorphism (or animism) through a cognitive task requires participants to engage in extensive cognitive processes (Epley et al., 2007, p. 865). First, participants must recall what it means to be human (or alive) and retrieve relevant knowledge about human (or living) traits and attributes. Next, they apply this knowledge to a target object, imagining it as human (or alive), and eventually, they articulate their thoughts in words. For example, when tasked with envisioning an advertised sports car as a living being, participants would begin by drawing from their existing knowledge of living agents, such as cheetahs, lions or horses. They would then project this understanding onto the sports car, perceiving it as if it were endowed with life, perhaps imagining it as a racing horse or a cheetah sprinting across the savannah. Ultimately, they must express these thoughts in writing. In essence, this cognitive exercise involves transferring knowledge from familiar entities, like humans, animals or plants, onto the target objects, such as advertised products, and it demands significant cognitive effort and workload.
How participants approach and perform these cognitive tasks should vary based on their knowledge of living organisms versus humans. When they make inferences about living entities, they can more easily draw from the bigger pool of examples (e.g., various species of animals and plants) to attribute life-like qualities to a target object. Thus, animism should be inherently more accessible than anthropomorphism because it extends to a broader range of entities, including plants, animals, and even microorganisms (Guthrie, 1995). Animism is also believed to have evolved as an adaptive defense mechanism to cope with unpredictable environments (Guthrie, 1995; Haselton & Nettle, 2006) and such universality of animism should facilitate animistic perceptions. Conversely, anthropomorphism may be less accessible and harder to elicit via a cognitive task than animism because it pertains to a single and narrower category—humans. The cognitive leap required to envision a target object as human may be more difficult to perform because there are fewer human-specific traits available to assign to the object. For example, participants may struggle to describe a car with human qualities if they do not have a wide range of human attributes to draw upon.
I expect that manipulation of animism (vs anthropomorphism) via a cognitive task should result in different performance outcomes. Specifically, animism should lead to greater elaboration, as measured by the volume of word generation, because participants can draw from a wider array of examples and broader knowledge associated with the more general category of “living organisms” compared to the narrower category of “humans”. Therefore, animism should yield more elaborate descriptions, while anthropomorphism should produce fewer words. This expectation aligns with past research indicating that individuals tend to generate longer texts on topics that are fluent and accessible, with familiarity facilitating elaboration and resulting in higher quantity of text (Li & Yu, 2024; for an overview see McNamara et al., 2015, p. 39). The effect of enhanced text quantity for an easier and more cognitively accessible animism than anthropomorphism should be particularly observable in the absence of various constraints, such as time limits or unclear instructions. Past research (Alter et al., 2007) suggests that blurred or degraded fonts, which impede fluency, can enhance focus and activate analytical reasoning. However, when participants are free from such distractions, they should be more likely to experience a sense of spontaneity, enabling more a natural flow of associations and greater elaboration, as reflected in word count.
2.4 Anthropomorphism, animism, elaboration and persuasion
When anthropomorphism (or animism) is manipulated through a cognitive task, different levels of elaboration may lead to varying persuasive effects in experimental settings. Specifically, if manipulating animism results in a greater number of words, it may prompt participants to respond positively to the target object because they have dedicated more labor into the descriptions. Prior research (e.g., Norton et al., 2012) shows that individuals tend to value objects more when they invest effort in creating them, as it increases their sense of connection and ownership. This effect is not strictly dependent on the amount of physical effort but arises from the process of creation (Marsh et al., 2018). Thus, when participants generate more words while construing mental images of an advertised product, this elaboration may foster positive reactions toward it, potentially enhancing purchase intentions and buying behavior.
An alternative explanation to the positive effect on persuasion may stem from the fluency experienced by participants. A task requiring participants to envision an object as alive may feel easier to perform than a task to envision an object as a human, due to broader and more varied associations within the category of living entities. Past research indicates a strong relationship between the ease of cognitive processing and persuasion. For example, easy-to-process messages enhance purchase intentions (Gomez et al., 2017) and are considered truer (Dechêne et al., 2010). Likewise, fluently processed products receive more favorable evaluations (Labroo et al., 2008; Lee & Labroo, 2004).
In sum, manipulation of animism should yield stronger persuasive effects than manipulation of anthropomorphism in experiments using cognitive tasks. I hypothesize that elaboration (as measured by the volume of generated words) should mediate these effects. Since purchase intentions typically precede and increase the likelihood of subsequent choice behavior (an assumption grounded in the theories of behavioral intention, such as the theory of planned behavior; Ajzen, 1991), I anticipate a serial mediation: animism should produce more extensive elaboration (i.e., more words) than anthropomorphism, leading to higher purchase intentions and, subsequently, greater choice behavior.
3 Overview of studies
Study 1 was designed to test the impact of manipulated animism (vs anthropomorphism) on elaboration and persuasion. Study 2 replicated the findings from Study 1, using a different product category as the stimulus, and including a control group for better comparisons. Study 3 examined the number of human-like cues as a boundary condition for the effects of manipulated animism (vs anthropomorphism) on elaboration and persuasion.
4 Study 11
Study 1 was an online one-factor experiment with a between-subjects design. I manipulated animism (vs anthropomorphism) with a cognitive task adapted from Aggarwal and McGill (2012) and others (e.g., Chandler & Schwarz, 2010; Chen et al., 2018, 2020; Fronczek et al., 2023; H. C. Kim & Kramer, 2015; Zhou et al., 2019).
4.1 Procedure and stimuli
Participants were recruited on Prolific (n = 200, Mage = 38.57, SD = 11.15; 20–65 years of age; 116 women, 80 men; UK and US citizens; native English speakers). They were randomly assigned to one of two experimental conditions. I selected a car as the focal category for testing because this product has been frequently used in empirical studies on anthropomorphism (Aggarwal & McGill, 2012). I created an online advertisement for an autonomous vehicle (Appendix A1) and instructed participants to imagine and describe it as a living being (vs a person) in writing. Following the manipulation, participants responded to questions measuring the dependent variables (purchase intentions, choice behavior), a potential mediator (i.e., processing fluency—an individual's perception of how easy a cognitive task feels) and manipulation checks.
4.2 Measures
Purchase intentions for the autonomous car were measured with a three-item seven-point scale adapted from Bruner (2009, p. 722; α = 0.94). Furthermore, participants were presented with a choice scenario in which they could click to either learn more about the car (coded as 1) or skip the advertisement (coded as 0). This captured participants' choice behavior in response to the advertisement and resulting from the manipulation (Appendix A2). Next, participants rated the fluency of imagining the car as a living being (vs a person) on dimensions such as difficulty, clarity, effort, fluency, and comprehensibility (α = 0.90; Graf et al., 2018). Eventually, I checked whether I successfully induced animism (α = 0.80) and anthropomorphism (α = 0.77) using six seven-point questions (three questions per cognitive process) adapted from Puzakova and Kwak (2017), Hur et al. (2015), Kim and McGill (2011). The measure of anthropomorphism (animism) asked about the extent to which an advertised product was perceived as if it were a person (alive), reminded of some human-like (living) qualities, looked like a person (a living being).
4.3 Results
I conducted a repeated-measure analysis of variance (ANOVA) to check manipulation of animism and anthropomorphism. Participants who were subject to manipulation of animism perceived the advertised car as more animate (M = 3.68, SD = 1.51) than anthropomorphic (M = 2.30, SD = 1.26; F(1,99) = 123.40, p < 0.001). Participants who were subject to manipulation of anthropomorphism perceived the car as more anthropomorphic (M = 3.62, SD = 1.42) than animate (M = 2.91, SD = 1.52; F(1,99) = 57.97, p < 0.001). These results suggest that the intended cognitive processes were successfully elicited.
The Kruskal–Wallis test revealed a statistically significant difference between both experimental conditions in the number of generated words (H = 7.23, df = 1, p = 0.007). Specifically, participants who were primed to anthropomorphize the car generated a lower number of words in their descriptions (Mdn = 17.50) compared to those who were primed to animate the car (Mdn = 20). These results remain consistent even when extreme outliers (i.e., exceeding 3 times the interquartile range above the third quartile or below the first quartile) are removed.
I conducted ANOVA, chi-square test and logistic regression to examine the main effects of manipulated animism (versus anthropomorphism) on persuasion (Table 1). Specifically, manipulation of anthropomorphism generated lower purchase intentions than manipulation of animism (Manthropomorphism = 2.54 vs Manimism = 2.98) and lower choice behavior (B = −0.78, SE = 0.34, W = 5.24, p = 0.02). However, both cognitive tasks were evaluated as similarly fluent.
Table 1
Kruskal–Wallis, ANOVA and chi-square results (Study 1)
Animism (n = 100) | Anthropomorphism (n = 100) | ||
|---|---|---|---|
Elaboration (median number of generated words [SD]) | 20(14.51) | 17.50(11.75) | H = 7.23, df = 1, p = .007 |
Fluency (mean [SD]) | 4.02(1.46) | 4.18(1.42) | F(1,198) = .58, p = .44 |
Purchase intentions (mean [SD]) | 2.98(1.72) | 2.54(1.36) | F(1,198) = 4.11, p = .04 |
Choice behavior (% of “Learn more” clicks) | 31% | 17% | χ2(1,200) = 5.37, p = .02 |
To test the mechanism behind the observed effects, I conducted a regression analysis using the SPSS PROCESS macro (model 6; Hayes, 2018) with manipulation of animism (coded as 1) versus anthropomorphism (coded as 0) as the independent variable. Elaboration (measured with word count) was entered as the first mediator, fluency as the second, purchase intentions as the third and choice served as the dependent variable (Appendix A3). Manipulation of animism led participants to generate more words than manipulation of anthropomorphism, resulting in higher perceived fluency. This fluency then increased purchase intentions, which raised the likelihood of clicking to learn more about the advertised car than skipping it. The indirect effect, including all three mediators, was statistically significant and positive (B = 0.007; bootSE = 0.006, bootLLCI < 0.001, bootULCI = 0.01 at 90% CI).
In sum, I found that animism generated more elaborate descriptions, which fully mediated the effects of manipulated animism (vs anthropomorphism) on choice behavior. Given the lack of direct effects on fluency, it may not be the primary mechanism underlying the observed outcomes. Therefore, I excluded this variable from subsequent studies.
5 Study 22
Study 2 sought to replicate the effects from Study 1 using a different product category (a smart speaker) and advertisement that would communicate more human-like activities embedded in the product (i.e., speaking and autonomy). It was an online between-subjects experiment (manipulation of animism vs anthropomorphism vs control).
5.1 Procedure, stimuli and measures
Participants were recruited on Prolific (n = 330, Mage = 41.02, SD = 13.03; 18–74 years of age; 195 women, 130 men; citizens of Ireland, UK and US; native English speakers). They were randomly assigned to one of three experimental conditions. The procedure, instructions and the measures were similar to Study 1 (Appendix B1-2).
5.2 Results
A repeated-measures ANOVA showed that participants in the animism condition perceived the speaker as more animate (M = 4.25, SD = 1.98) than anthropomorphic (M = 3.62, SD = 2.13; F(1,106) = 13.40, p < 0.001). However, those in the anthropomorphism condition saw the speaker as less anthropomorphic (M = 3.62, SD = 1.99) than animate (M = 4.01, SD = 1.86; F(1,107) = 12.60, p < 0.001). This suggests that animism manipulation is more potent, while anthropomorphism is challenging to evoke effectively. In the control group, participants rated the speaker as more animate (M = 1.89, SD = 1.28) than anthropomorphic (M = 1.75, SD = 1.12; F(1,114) = 4.05, p = 0.04), with both scores well below the midpoint indicating rather a non-animate and non-anthropomorphic perceptions. Given this imperfect anthropomorphism manipulation, I tested perceived anthropomorphism and animism as covariates in subsequent analyses.
The Kruskal–Wallis test revealed a statistically significant difference between the study conditions in the number of generated words (Table 2). Participants in a control condition produced the lowest number of words (Mdn = 13), followed by participants who were primed to anthropomorphize the product (Mdn = 15.50). Those who were instructed to animate the speaker provided the highest number of words in their descriptions (Mdn = 19). I removed two extreme outliers and the pattern of results remained the same.
Table 2
Kruskal–Wallis, ANOVA and chi-square results (Study 2)
Animism (n = 107) | Anthropomorphism (n = 108) | Control (n = 115) | ||
|---|---|---|---|---|
Elaboration (median number of generated words [SD]) | 19(14.78)a | 15.50(11.35)b | 13(7.54)b | H = 19.96, df = 2, p < .001 |
Purchase intentions (mean [SD]) | 3.69(1.68)a | 3.54(1.65)a | 2.97(1.74)b | F(2,327) = 5.66, p = .004 |
Choice behavior (% of “Learn more” clicks) | 24%a | 12%b | 23%a | X2(2,330) = 6.38, p = .04 |
I conducted ANOVA, logistic regression and chi-square test to inspect the effects of manipulation type on persuasion. Both animism (M = 3.69, SD = 1.68) and anthropomorphism (M = 3.54, SD = 1.65) produced higher purchase intentions than the control group (M = 2.97, SD = 1.74). Although animism showed numerically higher purchase intentions than anthropomorphism, this difference was not statistically significant. However, manipulation of animism significantly affected choice behavior: 24% of participants in the animism condition clicked “Learn more” compared to 12% in the anthropomorphism condition (B = 0.85, SE = 0.37, W = 5.25, p = 0.02; Table 2). To explore these findings further, I conducted a mediation analysis using PROCESS macro. Two models were structured: animism coded as 1 and anthropomorphism as 0 in the first, and control as −1 in the second model (Appendix B3-4). In both, elaboration (i.e., number of generated words) served as the first mediator, purchase intentions as the second mediator, while choice behavior was the dependent variable. Consistent with Study 1, manipulation of animism prompted more words than manipulation of anthropomorphism and the control condition, boosting purchase intentions and click likelihood. My analysis revealed significant indirect-only mediation (B = 0.04; bootSE = 0.02, bootLLCI = 0.01, bootULCI = 0.10 at 95% CI). I reran the mediation analysis with perceived animism and anthropomorphism as covariates, but the results remained consistent.
Studies 1–2 demonstrate that animism generates more elaborate descriptions, facilitating persuasion. However, unlike Study 1, animism's effect on persuasion in Study 2 was mixed, directly influencing choice behavior but not purchase intentions. This discrepancy may stem from varying levels of human-like cues across the studies. Study 1 focused solely on “autonomy”, while Study 2 included both “autonomy” and “speaking”, potentially complicating the distinction between anthropomorphism and animism. To address these issues, Study 3 manipulated the quantity of human-like cues.
6 Study 33
Study 3 was a 2 (manipulation of animism vs anthropomorphism) by 2 (human-like cues: many vs none) between-subject experiment.
6.1 Procedure, stimuli and measures
I recruited participants on Prolific (n = 440, Mage = 40.59, SD = 13.76; 18–77 years of age; 242 women, 191 men; citizens of Ireland, UK and US; native English speakers; pet owners). They were randomly assigned to one of four experimental conditions. The procedure, instructions and measures were the same as in previous studies (Appendix C1-2). Manipulation of human-like cues was assessed with a four-item scale (α = 0.90).
As a stimulus, I created an online advertisement for a pet feeder. In a condition involving human-like cues, the description of the product outlined activities typical for humans and human behavior, while in a condition involving no human-like cues, the description of the product was devoid of any features typical for humans (Appendix C1).
6.2 Results
I used ANOVA to test the manipulation of human-like cues. Participants recognized more human-like cues in the advertisement describing many human-like features (M = 4.19, SD = 1.56) compared to the advertisement without such description (M = 3.06, SD = 1.53; F(1,436) = 57.71, p < 0.001). A repeated-measure ANOVA reveled a significant main effect of the animism (vs anthropomorphism) manipulation (F(1,436) = 47.39, p < 0.001); no interaction effect with human-like cues was observed. Participants in the animism condition perceived the advertised feeder as more animate (M = 4.34, SD = 2.03) than anthropomorphic (M = 3.35, SD = 2.07; F(1,212) = 67.86, p < 0.001). However, participants in the anthropomorphism condition perceived the feeder as similarly anthropomorphic (M = 3.86, SD = 2.09) and animate (M = 3.78, SD = 2.02; F(1,226) = 1.85, p = 0.17). These results indicate successful manipulation of human-like cues and animism. Although the difference between anthropomorphic and animistic perceptions among participants who were subject to the manipulation of anthropomorphism was not statistically significant, the means in this condition differed in the expected direction, suggesting a trend that supports further analysis.
The Kruskal–Wallis test revealed a statistically significant difference in the number of generated words between study conditions (H = 7.85, df = 1, p = 0.005). Participants primed to anthropomorphize the product generated fewer words (Mdn = 19) than those instructed to animate the product (Mdn = 22). Removing eight extreme outliers, did not change the results. Furthermore, I found a significant interaction effect between manipulation type and the use of human-like cues on word count (B = −0.13, SE = 0.04, Wald χ2 = 11.32, df = 1, p < 0.001). Specifically, without human-like cues, animism led to higher word count (Mdn = 21) than anthropomorphism (Mdn = 17; H = 9.85, df = 1, p = 0.002). However, with multiple human-like cues, the number of words generated in the anthropomorphism condition increased, making both manipulations comparably effective (Mdn_manipulation_of_anthropomorphism = 19 vs Mdn_manipulation_of_animism = 22; H = 1.97, df = 1, p = 0.16).
As neither ANOVA nor logistic regression revealed any significant main effects of manipulated animism (vs anthropomorphism) on persuasion (ps ≥ 0.07), I conducted a moderated mediation analysis (model 85; Hayes, 2018) to examine the full process. Manipulation of animism (vs anthropomorphism) was the independent variable, while manipulation of human-like cues served as a moderator; elaboration (measured by word count) and purchase intentions were serial mediators, with choice behavior as the dependent variable. I found a similar pattern of results as in Studies 1–2 (Table 3). Specifically, animism encouraged participants to generate more words, leading to increased purchase intentions and higher likelihood of clicking the ad. This path was positive and statistically significant when the advertisement lacked human-like cues (B = 0.04, bootSE = 0.02, bootLLCI = 0.009, bootULCI = 0.10 at 95% CI) but not when it featured many human-like cues (B = 0.02, bootSE = 0.02, bootLLCI = −0.001, bootULCI = 0.06).
Table 3
Moderated mediation results (Study 3; n = 432; without outliers; model 85)
Elaboration (number of generated words) | Purchase intentions | Choice behavior | |||||||
|---|---|---|---|---|---|---|---|---|---|
B | SE | t | B | SE | t | B | SE | Z | |
Manipulation of animism (1) vs anthropomorphism (0) | 5.43 | 1.59 | 3.40*** | -.33 | .27 | −1.24 | -.02 | .33 | -.06 |
Humanlike cues: yes (1) vs no (0) | 2.50 | 1.57 | 1.59 | -.42 | .26 | −1.60 | -.001 | .32 | -.004 |
Manipulation of animism (1) vs anthropomorphism (0) x Humanlike cues: yes (1) vs no (0) | −2.82 | 2.25 | −1.25 | -.11 | .37 | -.31 | .27 | .46 | .58 |
Elaboration (number of generated words) | .01 | .008 | 2.01* | .01 | .009 | 1.68 | |||
Purchase intentions | .55 | .06 | 8.00*** | ||||||
Constant | 18.79 | 1.12 | 16.77*** | 3.79 | .24 | 15.62*** | −3.53 | .44 | −8.01*** |
7 Discussion
By employing separate manipulations for animism and anthropomorphism, I empirically disentangle these constructs and demonstrate their distinct effects. In three experiments, I show that tasks manipulating animism and anthropomorphism can elicit different cognitive processes, which has further consequences for persuasion.
I uncover that imagining a product as human (anthropomorphism) is more challenging than envisioning a product as alive (animism), as measured by the number of generated associations. This can lead to suboptimal outcomes in marketing research if anthropomorphism is manipulated without consideration for its cognitive load. Researchers should be cautious when using cognitive tasks to manipulate anthropomorphism, as this may obscure persuasive results. My findings suggest that some effects observed in past research may have been less negative (e.g., Fronczek et al., 2023) or turn significant (Zhou et al., 2019) if the manipulation had involved purely animism.
The effects of manipulated animism versus anthropomorphism on elaboration and persuasion appear contingent on the quantity of human-like cues. While it has been theorized that anthropomorphic thinking can be activated without any visual or verbal human-like cues (Epley et al., 2007, p. 865), my research demonstrates that a significant number of human-like cues is required to elicit anthropomorphism and equalize its effects with those of animism. In Study 1, I introduced just one cue; in Study 2, two cues were used; in both cases, animism appeared more persuasive than anthropomorphism. In Study 3, however, anthropomorphism approached the persuasion levels of animism when multiple human-like cues were integrated into manipulation. This finding suggests that too few human-like features make animism more readily elicited than anthropomorphism.
In sum, my work shows the need for greater precision in experimental manipulations of anthropomorphism. My findings can help researchers detect subtle yet significant effects that might otherwise be overlooked. Moreover, my research emphasizes the importance of the way the cognitive tasks are used, cautioning against conflating instructions to imagine objects as “alive” and “human”. Such conflation may increase the risk of activating animism instead of anthropomorphism, leading to persuasive effects that would not typically be observed with clear manipulation of anthropomorphism without multiple human-like cues.
Acknowledgements
I would like to thank the editors and the anonymous reviewers for their very encouraging, constructive and insightful feedback throughout the revision process. I also thank my friend for a lively debate about the applicability of academic research—a conversation that nearly ended in flying coffee cups and plates. This paper officially puts our discussion to rest, and I hope my findings will help researchers dodge research pitfalls and uncover those elusive effects that sometimes hide in plain sight.
Declarations
Ethics approval
Not applicable.
Informed consent
The participants in all the studies provided informed consent.
Conflicts of interest
No financial or non-financial interests to disclose.
Research involving human participants and/or animals
The studies reported in this manuscript involved Human Participants.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.