Abstract
Bridging the gap between the mental accounting and identities/roles literatures, the present research examines how the extent to which an individual's life roles (e.g., “employee,” “spouse”) are integrated (i.e., have more flexible and permeable psychological boundaries between them) moderates the fungibility of mentally accounted funds. Specifically, individuals with more integrated roles are more able to circumvent the constraints typically imposed by mental budgeting and earmarking; therefore, they are more likely to use funds that are allocated or budgeted for the purposes of one role to service the needs/wants of another role. This holds regardless of whether funds have been (1) allocated to a broader role-specific mental account for future expenditures or (2) earmarked for a specific purchase. The authors find evidence that the effect of role integration arises because those with more integrated roles believe that making purchases for one role using funds allocated or budgeted for the other role is more justifiable.
Mental accounting (Thaler 1980, 1985) has received a great deal of attention in the marketing literature, and for good reason. After all, despite having some nonnormative consequences (Heath and Soll 1996) and work-arounds (Cheema and Soman 2006), mental accounting is, at its core, a self-control mechanism (Shefrin and Thaler 1992) that aids in goal pursuit (Brendl, Markman, and Higgins 1998) and can improve consumer welfare (Soman and Cheema 2011). It is, therefore, no surprise that the lion's share of attention has been placed on the consequences of mental accounting, including its impact on preferences (Reinholtz, Bartels, and Parker 2015), consumption (Soman and Gourville 2001), saving (Thaler 1990), and sunk-cost effects (Heath 1995), among others. Naturally, there has been concurrent interest regarding how the manipulation of components of the mental accounting process, including the ambiguity (Cheema and Soman 2006) and temporal framing (Atlas and Bartels 2018; Gourville 1998) of expenses, moderates its consequences.
The research presented in this article builds on this prior work and contributes to a burgeoning literature examining the impact of exogenous factors on mental accounting effects (Abeler and Marklein 2017; Hossain 2018; Shah, Shafir, and Mullainathan 2015). Specifically, we bridge the gap between the identity/role and mental accounting literatures, and we find that greater role integration—the extent to which the boundaries between life roles (e.g., “spouse,” “executive”) are flexible and permeable (Ashforth, Kreiner, and Fugate 2000; Richter 1990)—increases the likelihood of spending funds allocated or budgeted for the purposes of one role to service the needs/wants of the other role. People with more integrated roles experience a greater degree of connectedness (Rothbard, Phillips, and Dumas 2005) between those roles and, as we subsequently argue in detail and find evidence for, believe that making purchases for one role using funds allocated or budgeted for the other role is more justifiable. In short, funds allocated or budgeted for the purposes of two life roles become more fungible and interchangeable as those two roles become more integrated.
Imagine a college student who runs for their university track team and regularly incurs expenses related to their student role (e.g., books, school supplies) and their athlete role (e.g., shoes, training gear). If they need to buy a new pair of running shoes, but already earmarked the money for school supplies, prior mental accounting literature suggests that they are not likely to purchase the shoes (Thaler 1980, 1985). We propose, however, that the likelihood of this expenditure depends on the extent to which their athlete and student roles are integrated. If they perceive the boundaries between their roles to be flexible and permeable, then it should be easier for them to justify the expense (e.g., “New track shoes will help me keep my scholarship to continue my education.”). So, the mentally accounted funds become more fungible, and the likelihood that they purchase the shoes using the funds set aside for school supplies increases.
We believe there is great value in examining the intersection of the identity/role and mental accounting literatures for several reasons. At the most basic level, despite the intense scrutiny placed on these two broad theoretical domains within marketing academia, the influence that they may have on each other has not been considered. But, more importantly, there is good reason to expect that the way people think about themselves (i.e., how they perceive, structure, and manage their roles and identities) can and should influence how they think about and manage their resources. After all, both are influenced by fundamental categorical principles and reasoning. One pertains to how we compartmentalize and label the different parts of our lives, whereas the other pertains to how we do the same with our various resources, including money. So, it seems only prudent to examine how these important theoretical domains are related, and this is a first step in that direction.
In what follows, we unpack the logic underlying the predicted effect of role integration on mental accounting. We first focus on role integration, an unexamined topic in the marketing literature. We then briefly discuss the core tenets of mental accounting, which are extensively covered in the marketing literature, as mentioned. Then, we integrate these two literatures to derive our predictions, present a series of studies supporting our contentions, and close with a discussion of the contributions and limitations of this work.
Theoretical Development
Role Integration
Contemporary self-theory posits that the self is composed of multiple roles rather than a single, unitary self (Higgins, Klein, and Strauman 1985; Linville 1987; Stryker 1968). For example, one may think of themselves in terms of their various roles as a manager, spouse, sibling, parent, and tennis player. These role-based self-concepts, which are synonymous with definitions of identity (Reed et al. 2012; Stryker and Serpe 1994), are assumed to be structures in a larger associative network of multiple roles (Campbell, Assanand, and Di Paula 2000; Gramzow et al. 2000; Reed and Forehand 2016; Reed et al. 2012). Each role, in turn, comprises specific functions, societal standards, and norms, which are largely exogenously determined and imposed on the individual (although some may be established and self-imposed by the individual). Roles are concurrently composed of values, goals, and beliefs, which are largely internal and specific to the individual. Collectively, these various facets of roles guide and constrain the consumer's behavior within a given social category and context (Linville 1987; McConnell 2011; Stryker 1968).
Some research examines how multiple roles impact consumers’ decisions, but research primarily focuses on the consequences of role conflict (Benet-Martínez et al. 2002; LeBoeuf, Shafir, and Bayuk 2010; Saint Clair and Forehand 2020). By contrast, the current work focuses on a structural feature of the self-concept: the extent to which individuals have more integrated versus segmented life roles (Ashforth, Kreiner, and Fugate 2000; Richter 1990).
To understand the role integration construct, we draw on boundary theory, which examines how individuals create and maintain “boundaries” or “mental fences” between roles as a means of simplifying and ordering the environment around them (Ashforth, Kreiner, and Fugate 2000; Zerubavel 1991). These psychological boundaries are real in the sense that individuals perceive them as such and act as though they are real (Weick 1979).
Flexibility and permeability are two key structural dimensions of role boundaries. Flexibility is the degree to which the boundary is spatially and temporally pliable (Hall and Richter 1988). Permeability is the degree to which an individual can be physically located in one role's domain but psychologically and/or behaviorally involved in another role (Pleck 1977; Richter 1990). So, an individual can erect strong psychological boundaries between two or more roles, enforcing a high degree of distinctiveness and separation between them, or allow these boundaries to be more flexible and permeable (Edwards and Rothbard 2000; Rau and Hyland 2002); the more (less) flexible and permeable the boundaries are, the greater the role integration (segmentation). Consequently, greater role integration is also associated with a stronger belief that one can determine when and where role-based activities take place (Matthews and Barnes-Farrell 2010; Winkel and Clayton 2010), which in turn increases the likelihood and frequency of transitioning between life roles. It should, therefore, be unsurprising that researchers find that greater role integration (1) begets a greater chance of an individual withdrawing from one role and stepping into another to meet the demands of that role (Winkel and Clayton 2010) and (2) attenuates interrole conflict because the individual is more equipped to make role transitions as necessary (Ashforth, Kreiner, and Fugate 2000).
Of course, role integration is not an either–or construct. Instead, the integration of two roles can lie at any point along a continuum between purely integrated and purely segmented. Moreover, the extent to which an individual's roles are integrated at a particular time is influenced by many factors, including life and family status (Kossek and Ozeki 1999; Richter 1990) and, of course, the extent to which the individual wishes to integrate the roles (Greenhaus and Singh 2003; Hartman 1997; Kossek, Lautsch, and Eaton 2006; Nippert-Eng 1996a, b). Thus, role integration can vary both chronically between individuals (i.e., behave as a trait) and transiently within individuals (i.e., behave as a state), as is the case with most individual differences (Mischel 1983) such as regulatory focus (Pham and Higgins 2005). Although the specific antecedents are beyond the scope of the current work, some individuals will have a greater propensity than others to integrate their life roles in general (i.e., a chronic tendency to integrate that does not relate to any set of specific roles). More critical for the current work, the extent to which an individual believes any two roles are integrated can vary over time and be influenced by both volitional internal factors (e.g., the desire to integrate/segment those roles) and uncontrollable external factors (e.g., external expectations or cues that evoke memories and evidence of the integration of those two roles over time). We return to this point in the “General Discussion” section.
It is important to emphasize that even as two roles become more or less integrated, they will remain distinct in terms of what is expected of the individual and the functions those roles serve (e.g., one's work-life role will continue to be a source of income, health care, and professional achievement, whereas one's home-life role will continue to be a source of love, affection, and social connectedness). This is because the functions, standards, and norms of roles are largely (though not necessarily wholly) exogenously determined and, thus, independent of how integrated or segmented one perceives their roles. So, an individual may not erect strong psychological boundaries between socially defined life roles (i.e., increase the integration of those roles) despite explicitly recognizing and accepting that those roles have dissimilar expectations and functions. Yet, the functions, standards, and norms associated with a given set of roles can, and often do, influence the likelihood of those roles being integrated (Ditzfeld and Showers 2013; Linville 1987).
To date, research has primarily examined how role integration impacts employee well-being (Peng, Ilies, and Dimotakis 2011), satisfaction and commitment (Kahn et al. 1964; Rothbard, Phillips, and Dumas 2005), and productivity (Kossek and Ozeki 1999), as well as the effectiveness of organizational policies such as on-site day care facilities (Hall and Richter 1988), work–family balance programs (Hall and Richter 1988), and telecommuting (Kossek, Lautsch, and Eaton 2006). To the best of our knowledge, no research has linked role integration to broader consumption-relevant behaviors, much less mental accounting. Thus, it is not clear if or why role integration could impact the fungibility of mentally accounted funds.
Mental Accounting
The concept of mental accounting posited by Thaler (1980, 1985) helps explain why spending, consumption, and investment decisions regularly deviate from normative microeconomic predictions. The most fundamental takeaway from the mental accounting literature is that funds that have been budgeted to a specific account (Heath and Soll 1996) or earmarked for specific purposes or purchases (Soman and Cheema 2011) become nonfungible. For instance, once a consumer allocates a given dollar to a specific account (e.g., food) or specific purchase (e.g., a special dinner), that dollar is significantly less likely to be spent on purchases that do not correspond with that focal account or intended purchase (e.g., going to a movie).
Yet, mental accounting is not a black-and-white process, and consumers often behave as if mental accounts and potential purchases are malleable. This is, in part, because mental accounting relies on the basic principles of mental categorization (Henderson and Peterson 1992; Reinholtz, Bartels, and Parker 2015), and many mental accounts are vaguely defined at best (e.g., “entertainment”), sometimes intentionally, which renders the assignment of expenses to mental accounts nondeterministic. Moreover, various expenses may be motivationally framed to make them better fit one account or the other (e.g., a dine-in movie can be classified as “food” or “entertainment”). Doing so makes incurring such expenses more (or less) justifiable, thereby increasing (or decreasing) the likelihood of incurring the expense (Cheema and Soman 2006).
So, it is empirically established that factors pertaining to the mental accounting process itself, such as the vagueness of mental accounts and the ambiguity and desirability of potential expenses, moderate the apparent fungibility of mentally accounted funds. Far less is known about how factors pertaining to the individual (i.e., individual differences) influence the mental accounting process and, more specifically, the fungibility of mentally accounted funds. Exceptions are limited to evidence that consumers with weaker mathematical abilities (Abeler and Marklein 2017), fewer resource constraints (Shah, Shafir, and Mullainathan 2015), and more analytical (vs. holistic) cognitive styles (Hossain 2018) tend to be more influenced by mental accounting and earmarking. Notably, each of these factors, much like role integration, can persist chronically (e.g., an inherent mathematical inability or systemic poverty) or be more transient, changing over time or context (e.g., thinking holistically for certain purchases but analytically for others). We ask whether the extent to which one's life roles are integrated—another individual-specific factor—also moderates the fungibility of mentally accounted funds.
How Role Integration Can Impact Mental Accounting
As discussed, people have multiple roles that collectively define who they are. These roles are integral to their daily lives, and each role has certain unique goals (e.g., get promoted at work, be a good parent, complete a marathon). Because mental accounting is often guided by one's goals (Brendl, Markman, and Higgins 1998), it stands to reason that people will often earmark funds and form mental accounts in the service of role-specific goals. For instance, an office manager who wishes to be professional and productive at work may have role-aligned mental accounts for work clothing and lunches in the office cafeteria, and they may save and earmark funds to attend a professional development course. This individual may also have monetary budgets for their other roles (e.g., parent, friend, philanthropist, runner). All of this is to say that a meaningful portion of individuals’ goals pertain to or are driven by the roles they play in life, and accordingly, a meaningful portion of their mental accounting activity should also correspond to those roles.
Although the idea that people are likely to budget and earmark funds for role-specific purposes is intuitively appealing, we empirically verified this assumption with two pilot studies (we describe the full procedures of both pilot studies in the Web Appendix). In the first pilot study, 100 participants, recruited from CloudResearch, first listed purchases they made for which they had previously set aside money. Two participants did not list any purchases, leaving 98 usable responses for analysis (53 male, 44 female, 1 other; Mage = 38.44 years). On average, participants listed 5.28 purchases (SD = 2.15). Participants then read a description of life roles and indicated whether each purchase they listed was associated with one of their life roles. On average, participants indicated that 2.88 (SD = 2.17), or 51.7%, of their purchases were associated with one of their life roles. In the second pilot study, we conducted in-depth interviews with six members of a community band, during which we established that they considered “band member” and “musician” personal life roles. Three of these six individuals recalled specific instances in which they set aside funds specifically for purchases for their band member/musician roles. The remaining three individuals indicated that they had not specifically budgeted for such purchases because their financial circumstances did not require it. All six individuals indicated that they regularly incur expenses for their band member/musician roles. In summary, across the two pilot studies, we found compelling evidence that a substantial portion, but by no means all, of people's mental accounting occurs in the service of specific life roles.
With those results in mind, it is important to emphasize that mental accounting is a form of self-control by which the consumer protects specific resources intended for a given purpose from being spent on other tempting opportunities (e.g., other products or activities) by labeling those resources and creating preset rules regarding how they can or should be spent. These preset rules impose a psychological-cost threshold that must be overcome for the individual to violate them (Cheema and Soman 2006). This psychological-cost threshold is commonly associated with how easily justified a purchase is on the basis of the preset rules for the money that is to be spent. To wit, researchers argue that psychological costs in the form of the pain of choosing (Kivetz and Simonson 2002) and regret, both anticipated and experienced (Zeelenberg and Pieters 2007), are greater when purchases are less justifiable. If the individual hopes to avoid these affective consequences, they must find a way to justify the purchase, which is a process that also imposes psychological costs because it requires the individual to generate arguments and reasons for making the purchase. These arguments can adopt the approach of either demonstrating that the purchase does, in fact, conform to the preset rules or undermining the validity or reasonableness of those preset rules. In summary, less justifiable purchases impose greater psychological costs on the individual, be they affective penalties if the purchase is not justified or cognitive penalties if the individual makes the effort to justify the purchase to avoid the affective consequences.
As discussed previously, individuals regularly take advantage of desirable but ambiguous expenses that allow for malleable expense posting and account construction, which makes those purchases more justifiable (Cheema and Soman 2006) by enabling the individual to reason that the desirable purchase meets their preset spending rules. Similarly, we propose that the more integrated two roles are, the more justifiable making purchases for one role using funds earmarked or budgeted for the other role will appear.
Making purchases for one role using funds earmarked or budgeted for another role should be considered more justifiable when the individual perceives themselves to be more the same person (vs. distinct people) across the two roles with common (vs. distinct) goals. Work on role integration suggests that integrated roles share personality traits, dispositions, and values (Frone, Russell, and Cooper 1992; Morf 1989; Staines 1980; Zedeck 1992). Other research suggests that people believe their integrated roles have general mood, behavior, and skill similarities (Ashforth, Kreiner, and Fugate 2000; Rothbard, Phillips, and Dumas 2005). As such, greater role integration leads to a high degree of congruence and feeling of connectedness between roles (Rothbard, Phillips, and Dumas 2005). Consequently, the more integrated two roles are, the more the individual should believe they are able to justify making purchases for one role using funds allocated or budgeted for the other role. For example, a working mom with highly integrated roles may believe it is justifiable to use money earmarked for an activity with her family to attend a work seminar because success at work serves her family values and goals in the form of providing funds necessary to give her son opportunities, an education, and so on. When roles are more segmented and less connected, consumers should be less able to justify the purchase because fewer commonalities, including goals and values, are shared across the roles. In summary, consumers with more integrated (vs. more segmented) roles should be significantly more likely to spend resources earmarked for the purposes of one role to service the needs/wants of another role because they find such purchases more justifiable.
Summary
We argue that greater role integration makes purchases for one role using funds allocated or budgeted for the other role more justifiable and thus likely. In other words, it renders resources allocated or budgeted for the purposes of each role more fungible. We test our predictions across five studies. Study 1 finds that chronic role integration influences the likelihood of consumers using funds allocated and budgeted for one role to purchase a product that will benefit another role. Study 2 finds that manipulated, as opposed to measured, role integration yields the same effect, allaying concerns pertaining to potential confounds driving the results in Study 1. Study 3 finds that the ease of justifying purchases mediates the link between role integration and purchase likelihoods. Study 4 moderates ease of justification to provide further evidence for the proposed psychological process underlying these results. In addition, Studies 3 and 4 employ an experimental design characteristic that helps cast doubt on an important alternative explanation. Specifically, both studies examine how people choose to spend funds that have been earmarked for a specific purchase and find that role integration increases the likelihood that those funds are spent on the needs/wants of the role for which they were not originally earmarked. Because these results are obtained when participants are considering money earmarked for a specific purchase, as opposed to allocated to a spending account (e.g., work clothing, entertainment), it is unlikely that the documented effect arises because people with more integrated roles construct or maintain broader, less well-defined, or more malleable mental accounts than those with less integrated (more segmented) roles. Finally, Study 5 demonstrates that the effect replicates in the context of an incentive-compatible choice that forces trade-offs (i.e., sacrificing benefits intended for one role in the service of another). Study 5 also addresses an important ecological validity concern about Studies 3 and 4 by holding the focal roles constant and asking participants if those roles are integrated or segmented only after they respond to the focal dependent measures in the study. We discuss this point further in the “General Discussion” section.
For all our studies, sample sizes were determined prior to collecting the data. Data for all studies are available at https://osf.io/89v2r/?view_only=639294eca8e34130b380e4f283b7c31f. We report complete stimuli and details of all the measures captured in each study in the Web Appendix. We also note that we employed a quality control procedure at the beginning of Studies 1, 2, 3, and 4: we redirected all participants who failed this procedure to the end of the survey, thus they did not complete any portion of the study. Further details on this quality control procedure, including the number of people who failed, can be found in the Web Appendix.
Study 1: Chronic Role Integration and Purchase Decisions
Study 1 examines the impact of role integration on purchase decisions involving mentally accounted funds. This study placed participants in a scenario in which they were confronted with a purchase that would benefit one role but require the use of funds budgeted for the purposes of another role. Choosing to make such a cross-account purchase would indicate that one's mentally accounted funds are at least somewhat fungible.
Method
Two hundred one Amazon Mechanical Turk (MTurk) participants (Mage = 38.83 years, 47.26% male) were given descriptions of work-life roles (“those we play [i.e., the person we are] when we are at work or are engaged in work-related activities”) and home-life roles (“those we play [i.e., the person we are] when we are either at home or engaged in home-related activities”) (for full details, see the Web Appendix), and they were asked to imagine that for the last few months they had been maintaining two budgets (a home-expense budget and a work-expense budget) to support their home-life and work-life roles. All participants then imagined they were out shopping when they came across a desktop mini fridge that they really liked, which was of the type they had been thinking of buying for their workplace. They also read that they would use this fridge only at their workplace.
We randomly assigned participants to the within-budget or cross-budget condition. Participants in the cross-budget condition read that they already exhausted their work-expense budget for the month (i.e., the budget from which the fridge purchase should ostensibly be made) but there were sufficient surplus funds in their home-expense account to cover the purchase, if they so desired. Thus, if these participants wanted to purchase the mini fridge for their workplace, they would need to use money from their home-expense budget (i.e., make a cross-budget purchase). Participants in the within-budget condition read that they had adequate funds in their work-expense budget to make the purchase. All participants then responded to the question, “How likely are you to buy the desktop fridge?” (1 = “very unlikely,” and 9 = “very likely”).
To conclude the study, participants read a brief description of the differences between integrated and segmented life roles (for full details, see the Web Appendix). The description of integrated roles stated, “For some people, the roles (work life and home life) we play are related and influence each other. In fact, they often overlap in time and location. For instance, when we are at work, we are still a son, daughter, brother, sister, niece, nephew, friend, and so on.” The description of segmented roles stated, “For some people, the roles (work life and home life) we play are clearly separate and distinct. In fact, they rarely overlap in time and location. For instance, when we are at work, we don’t think of our role of being a son, daughter, brother, sister, niece, nephew, friend, and so on.” Participants then indicated whether they considered themselves to be a person with either integrated or segmented work-life and home-life roles (i.e., they self-categorized their chronic role integration tendencies for these specific roles). Although it is true that role integration lies along a continuum, we expected that participants would be able to categorize their own chronic role integration tendencies meaningfully.
We predicted that those with chronically integrated work-life and home-life roles would treat the money allocated to their role-aligned budgets as more fungible, and thus they would be significantly more likely to purchase the workplace mini fridge using their home-expense budget (i.e., make a cross-budget purchase) than participants with segmented roles. By contrast, role integration was not expected to influence purchase likelihoods in the within-budget condition in which participants could simply use their work-expense budget to purchase the workplace mini fridge (i.e., the expense and the budget from which the resources would be drawn were related).
Results and Discussion
A 2 (roles: integrated vs. segmented) × 2 (budget: within vs. cross) analysis of variance (ANOVA) with purchase likelihood as the dependent variable revealed a nonsignificant main effect of role integration (Mintegrated = 5.23, SD = 2.61 vs. Msegmented = 4.58, SD = 2.86; F(1, 197) = 1.26, p = .26, ηp2 = .006) but a significant main effect of budget (Mcross-budget = 3.87, SD = 2.53 vs. Mwithin-budget = 5.93, SD = 2.58; F(1, 197) = 31.82, p < .0001, ηp2 = .139). This main effect is consistent with typical mental accounting effects; participants were significantly more likely to purchase the mini fridge for their workplace when there were funds available (vs. not) in their work-expense budget (i.e., the budget associated with the purchase).
More critically, there was a significant interaction of role integration with budget (F(1, 197) = 7.33, p = .007, ηp2 = .036; Figure 1). Planned contrasts revealed the expected pattern of results. In the cross-budget condition, participants with integrated (vs. segmented) roles were significantly more likely to purchase the workplace mini fridge using funds allocated to their home-expense budget (Mintegrated = 4.64, SD = 2.58, N = 44 vs. Msegmented = 3.27, SD = 2.34, N = 56; F(1, 197) = 7.31, p = .007, d = .56). By contrast, role integration did not have a significant effect in the within-budget condition (Mintegrated = 5.68, SD = 2.56, N = 57 vs. Msegmented = 6.25, SD = 2.59, N = 44; F(1, 197) = 1.26, p = .26, d = .22).
The Effect of Chronic Role Integration and Budget Condition on Purchase Likelihoods (Study 1).
As documented in the Web Appendix, there was a marginally significant effect of budget condition on participants’ role integration responses. Thus, it is possible that the two independent variables in this study were not fully orthogonal. Although it strikes us as unlikely that the hypothetical budget manipulation used in this study would meaningfully alter how participants perceived their life roles, the following studies address this weakness in several ways. First, role integration is manipulated, as opposed to measured, in two different ways in Studies 2, 3, and 4. Second, participants choose their own life roles in Studies 3 and 4, as opposed to considering the broader work-life and home-life roles used in Study 1. Third, role integration is measured again in Study 5, but there is no manipulation of mental accounts beforehand. The focal result of Study 1 is replicated across all these methods, thereby undermining the conclusion that these results were a result of a lack of orthogonality.
In summary, Study 1 found that role integration increased purchase likelihoods when the purchase would require using funds budgeted for one role (i.e., allocated to one's home-expense budget) to benefit the other role (i.e., one's work-life role). By contrast, as expected, role integration did not impact purchase likelihoods when the purchase would be made using funds allocated for the role that would benefit from the purchase (i.e., a purchase made from a work-expense budget on a product for the workplace). However, as mentioned, there are always inherent issues with confounds when independent variables are measured as opposed to manipulated, including the potential for self-generated validity, so we manipulate role integration in Study 2.
Study 2: Transient Role Integration and Purchase Decisions
The goals of Study 2 were twofold. First, we wanted to replicate the findings of Study 1 by manipulating, as opposed to measuring, role integration to reduce potentially inherent confounds between chronic role integration and other individual differences, as well as potential self-generated validity effects. Second, we examined ease of justification as the mediator between role integration and purchase likelihoods.
Method
Two hundred one MTurk participants (demographic information was not collected) read the same brief descriptions of work-life and home-life roles used in the previous study. We randomly assigned participants to one of the two role integration conditions (integrated vs. segmented). Adapting a procedure from Reyt and Wiesenfeld (2015), we manipulated role integration by first giving participants a brief description of what it means to have either integrated or segmented roles (for full details, see the Web Appendix). In the integrated (segmented) condition, participants read that “a person with integrated (segmented) work-life and home-life roles will strive hard to keep these roles integrated (separate) … such a person would answer (not take) home calls while at work or discuss work at home with family members and vice versa … [and] would generally encourage (discourage) any overlap between roles.” Participants then described how their own work-life and home-life roles were either integrated or segmented. We validated this manipulation in a study reported in the Web Appendix. Subsequently, as in Study 1, all participants then imagined that they maintained two budgets: a work-expense budget and a home-expense budget. On the same screen, all participants saw examples of typical items purchased with each budget. They then imagined that they were considering purchasing a computer monitor they would use “exclusively” in the service of one role (either their work-life or home-life role). Next, they read that the budget from which the computer monitor purchase should ostensibly be made (either the work-expense or home-expense budget) was depleted for the month. Thus, they would have to use some discretionary funds from their other role-aligned budget, which was not depleted, to make the purchase. We counterbalanced where the product would be used and, necessarily, from which budget the purchase would be made. Half the participants read that the product would be used exclusively at work but the purchase would have to be made from their home-expense budget, and half the participants read the opposite. Subsequently, all participants responded to the question, “How likely are you to purchase the computer monitor for your office (home) using money budgeted for home-life (work-life) expenses?” (1 = “very unlikely,” and 9 = “very likely”). All participants also indicated their agreement with two statements: “making the above purchase would be easy to justify to myself” and “making the above purchase would feel right to myself” (1 = “strongly disagree,” and 9 = “strongly agree”). We averaged these two statements to form our measure of ease of justification (r = .92).
Results and Discussion
The budget the money for the purchase would come from, which was counterbalanced across participants, had no significant main effects on participants’ purchase intentions, and it did not significantly interact with the role integration factor. Thus, we collapse across this distinction in the following analyses.
A one-way ANOVA revealed that participants in the integrated (vs. segmented) roles condition were significantly more likely to purchase the computer monitor using funds earmarked for the other role (Mintegrated = 4.79, SD = 2.63, N = 99 vs. Msegmented = 3.41, SD = 2.51, N = 102; F(1, 199) = 14.39, p < .001, d = .54). We found a similar pattern for ease of justification; participants in the integrated (vs. segmented) roles condition reported greater ease of justification (Mintegrated = 4.82, SD = 2.72 vs. Msegmented = 3.92, SD = 2.60; F(1, 199) = 5.77, p = .017, d = .34).
To test ease of justification as the process underlying the effect of role integration on purchase likelihoods, we used PROCESS model 4 (Hayes 2013) with 5,000 bootstrap samples (role integration was coded as integrated [1] vs. segmented [2]). The results revealed that the predicted indirect effect of role integration on purchase likelihoods through ease of justification was significant (b = −.76, SE = .32, 95% confidence interval [CI95] = [ − 1.3865, −.1387]). Collectively, Studies 1 and 2 found that consumers’ chronic and transient role integration tendencies influence the fungibility of funds in their role-aligned mental accounts; the more integrated the roles, the more fungible the funds. In addition, Study 2 found evidence that the influence of role integration on purchase likelihoods was mediated by the ease of justifying the purchase, as predicted.
Although Study 2 provides initial evidence for ease of justification as the process underlying the effect of role integration on purchase likelihoods, another feasible explanation exists. It is possible that individuals with greater role integration maintain broader mental accounts. In other words, rather than having a work-aligned mental account and a home-aligned mental account, perhaps individuals with greater role integration have a combined work and home account. Accordingly, in Studies 3 and 4 we examine another established mental accounting effect documented in the literature: the fungibility of funds that have been earmarked for a specific purchase (Soman and Cheema 2011). Examining our effect in this context relaxes the assumption that people maintain formal budgets related to their roles, and it enables us to determine if the results we have found thus far are the result of differences in the structure of such mental accounts across integrated and segmented individuals.
Study 3: Replicating the Effect in the Context of Specific Earmarking
There were three goals for Study 3. The first was to examine if the pattern of results obtained in Studies 1 and 2 would be replicated in the absence of broad mental accounting budgets. It is possible that many people maintain broad mental budgets for home- and work-related expenses, but it is by no means a given. So, although our predictions were fully supported in Studies 1 and 2, the results of both studies could be considered somewhat artificial and lacking in external validity because all participants were asked to imagine that they maintained home- and work-expense budgets. To quell such concerns, Study 3 asked participants to imagine that they had set aside (i.e., earmarked) money for a specific purchase only, which is a common practice.
The second goal was to replicate the results of Studies 1 and 2 using roles and products chosen by the participants (vs. chosen by us, as in Studies 1 and 2). Doing so enables us to (1) generalize our results to a broader range of roles and products and (2) continue to reduce concerns regarding the artificiality and external validity of Studies 1 and 2. The third and final goal of this study was to examine, once again, whether ease of justifying the purchase mediates the link between role integration and purchase likelihoods.
Method
One hundred nineteen MTurk participants (Mage = 38.36 years, 52.94% male) first read a brief description of life roles and several examples of common roles most people have (e.g., spouse, parent, manager, athlete, musician) (for complete details and instructions, see the Web Appendix). Next, as in Study 2, we randomly assigned participants to one of the two role integration conditions (integrated vs. segmented). Participants in the integrated (segmented) roles condition read a description of integrated (segmented) roles similar to the manipulation used in Study 2). They were then asked to indicate two roles in their life (hereinafter, Role A and Role B) that they considered integrated (segmented). Subsequently, participants in the integrated condition described how they kept their two roles integrated, blending “certain elements and facets of one role with the other role.” Participants in the segmented condition were asked to describe how they kept their two roles segmented, keeping “elements and facet of one role totally separated from the other role.” Then, participants were asked to name one product they planned to purchase for Role A (hereinafter, Product A) and one product they planned to purchase for Role B (hereinafter, Product B). The survey software piped these roles and products into the text throughout the remainder of the study. So, each participant in Study 3 considered roles they actually maintained in their personal lives that were either integrated or segmented (manipulated between subjects) and two products (one associated with each role) that they actually intended to purchase in the future.
Participants then imagined that they had set aside enough money to buy Product A but came across Product B while shopping. They read that they could purchase Product B but they would have to use the money set aside for Product A to do so, and they indicated how likely they would be to purchase Product B (1 = “very unlikely,” and 9 = “very likely”). They also indicated how easily justifiable purchasing Product B would be using the same two measures of ease of justification from Study 2 (r = .78).
Results and Discussion
A one-way ANOVA revealed that participants in the integrated roles condition were significantly more likely to purchase a product associated with one role (Product B) using funds earmarked for a product associated with another role (Product A) (Mintegrated = 5.83, SD = 2.42, N = 60 vs. Msegmented = 4.51, SD = 2.7, N = 59, 1; F(1, 117) = 7.94, p = .006, d = .52). A second one-way ANOVA used ease of justification as the dependent variable and found that participants in the integrated roles condition believed purchasing Product B was more easily justified (Mintegrated = 6.84, SD = 1.99 vs. Msegmented = 5.56, SD = 2.70; F(1, 117) = 8.69, p = .004, d = .54).
Using Hayes’s (2013) PROCESS macro (Model 4, 5,000 bootstrap samples, role integration coded as integrated [1] vs. segmented [2]), we again found that ease of justification significantly mediated the effect of role integration on purchase likelihoods (b = −.82, SE = .30, CI95 = [−1.4278, −.2624]). In summary, we once again find that people with more integrated (segmented) roles are more (less) likely to use funds budgeted for the purposes of one role to support the needs/wants of the other role. Moreover, this was found to hold when the funds under consideration were earmarked for a specific purpose (as opposed to allocated to a more general spending budget) and when participants chose their own existing life roles and intended purchases.
Study 4: Moderating Ease of Justification
Studies 2 and 3 found evidence that the ease of justifying the purchase mediates the link between role integration and purchase likelihoods using the conventional measured mediator approach. Study 4 was designed to test moderation of this psychological process by directly manipulating the ease of justifying the purchase, with the expectation that this would moderate the effect of role integration on purchase likelihoods. As mentioned, mental accounting is, at its core, a self-control mechanism (Shefrin and Thaler 1992) that aids in goal pursuit (Brendl, Markman, and Higgins 1998). Importantly, it has been found that perceived progress toward the focal goal makes pursuing alternative goals more justifiable (Louro, Pieters, and Zeelenberg 2007). Relevant to the current focus, this implies that people should believe using earmarked funds for something other than the intended purchase is more easily justified if the intended purchase can be achieved via other means (e.g., received as a gift). From this premise, we predicted that the effect of role integration on purchase likelihoods would be attenuated when the earmarked purpose would likely be fulfilled via alternative means. Study 4 tested this prediction. This study was preregistered (https://aspredicted.org/7tb4f.pdf).
Method
Two hundred eighty-two MTurk workers participated (demographic information was not collected) in Study 4. The study utilized a 2 (roles: integrated vs. segmented) × 2 (alternative means: no vs. yes) between-subjects design (for full details, see the Web Appendix). We manipulated role integration exactly as in Study 3; participants chose two roles of their own (Roles A and B) that they considered either integrated or segmented (manipulated between subjects), and they explained why they considered those roles to be integrated or segmented. In addition, as in Study 3, participants indicated one product they planned to purchase for Role A and one product they planned to purchase for Role B (Products A and B). The survey software piped these roles and products into the text throughout the remainder of the study. Participants imagined that they had set aside enough money to buy Product A but came across Product B while shopping. They read that they could purchase Product B, but they would have to use the money set aside for Product A to do so.
At this point, participants in the alternative means condition imagined their birthday was fast approaching, and they had overheard conversations among their relatives indicating that they would receive Product A as a gift (i.e., they could satisfy the goal of acquiring Product A via means other than the money they had earmarked to purchase it). Participants in the no-alternative-means condition were given no additional information about the likelihood of acquiring Product A via other means, and thus these integrated and segmented roles conditions were identical to those in Study 3. Participants in all four conditions then indicated how likely they would be to purchase Product B (1 = “very unlikely,” and 9 = “very likely”). Finally, participants indicated how easy it would be to justify purchasing Product B using the same two measures of ease of justification from Studies 2 and 3 (r = .90).We predicted that the likelihood of purchasing Product B with the funds earmarked for Product A would be high regardless of the extent to which Roles A and B were integrated in the alternative means conditions because Product A would likely be acquired via other means, which makes the purchase of Product B more easily justified. By contrast, we predicted that the ease of justifying the purchase of Product B with funds earmarked for Product A and the likelihood of making the purchase would be greater for those with integrated (vs. segmented) roles in the no-alternative-means condition.
Results and Discussion
A two-way ANOVA using the likelihood of purchasing Product B as the dependent variable revealed a main effect of having alternative means to acquire Product A (Mno = 4.64, SD = 2.64 vs. Myes = 5.67, SD = 2.51; F(1, 278) = 11.93, p < .001, ηp2 = .041). This result reflects the fact that people should be more likely to purchase Product B with funds earmarked for Product A when they believe they will acquire Product A via alternative means. There was no significant impact of role integration on the likelihood of purchasing Product B (Mintegrated = 5.25, SD = 2.46 vs. Msegmented = 5.06, SD = 2.78; F < 1, p = .507, ηp2 = .002). Most importantly, the expected interaction of role integration (integrated vs. segmented) with the availability of alternative means of acquiring Product A (no vs. yes) was significant (F(1, 278) = 7.32, p = .007, ηp2 = .026; Figure 2). Planned contrasts found the expected pattern. Replicating the previous studies, when participants had no apparent alternative means for acquiring Product A (i.e., the no-alternative-means conditions), they were significantly more likely to purchase Product B in the integrated roles condition (M = 5.14, SD = 2.47, N = 72) than in the segmented roles condition (M = 4.12, SD = 2.73, N = 69; F(1, 278) = 5.68, p = .018, d = .40). By contrast, when participants were aware that they would likely acquire Product A via other means (i.e., as a birthday gift, the alternative means conditions), role integration did not significantly affect the likelihood of purchasing Product B (Mintegrated = 5.37, SD = 2.46, N = 71 vs. Msegmented = 5.99, SD = 2.53, N = 70; F(1, 278) = 2.08, p = .15, d = .24).
The Effect of Role Integration Condition and Alternative Means of Purchase on Purchase Likelihoods (Study 4).
A second two-way ANOVA found a significant main effect of the availability of alternative means of acquiring Product A (no vs. yes) on the ease of justifying purchasing Product B (Mno = 5.80, SD = 2.61 vs. Myes = 6.47, SD = 2.17; F(1, 278) = 5.83, p = .016, ηp2 = .021; Figure 3). These results show that knowing Product A will likely be acquired via other means makes it easier to justify purchasing Product B with funds originally earmarked for purchasing Product A. There was no impact of role integration on justification for the purchase of Product B (Mintegrated = 6.29, SD = 2.28 vs. Msegmented = 5.98, SD = 2.55; F(1, 278) = 1.26, p = .263, ηp2 = .005). Most importantly, the expected interaction of role integration (integrated vs. segmented) with the availability of alternative means of acquiring Product A (no vs. yes) was significant (F(1, 278) = 6.25, p = .013, ηp2 = .022). Planned contrasts found the expected pattern. When participants had no apparent alternative means for acquiring Product A (i.e., the no-alternative-means conditions), they reported significantly greater justification for purchasing Product B with money originally earmarked for acquiring Product A in the integrated roles condition (Mintegrated = 6.30, SD = 2.33 vs. Msegmented = 5.28, SD = 2.78; F(1, 278) = 6.56, p = .011, d = .40). By contrast, when participants were aware that they would likely acquire Product A via other means (i.e., as a birthday gift, the alternative means conditions), role integration did not significantly affect justification (Mintegrated = 6.28, SD = 2.24 vs. Msegmented = 6.66, SD = 2.09; F < 1, p = .33, d = .18).
The Effect of Role Integration Condition and Alternative Means of Purchase on Purchase Justification (Study 4).
Accordingly, we conducted a moderated mediation using Hayes’s (2013) PROCESS macro (Model 7, 95% confidence interval, 5,000 bootstrap samples, role integration coded as integrated [1] vs. segmented [2], alternative means coded as no [1] vs. yes [2]) to determine if the ease with which participants believed they could justify purchasing Product B mediated the combined effect of role integration (integrated vs. segmented) and the availability of alternative means of acquiring Product A (no vs. yes) on purchase likelihoods. This analysis revealed that the moderated mediation was significant (b = .97, SE = .40, CI95 = [.2270, 1.7720]). More importantly, and as we predicted, the indirect effect of role integration on purchase likelihoods via ease of justification was only significant when there were no alternative means of acquiring Product A (b = −.70, SE = .31, CI95 = [ −1.3256, −.1233]). When there were alternative means of acquiring Product A, which makes the purchase more easily justified regardless of role integration, there was not a significant indirect effect (b = .27, SE = .25, CI95 = [−.2164, .7620]).
In summary, when the ease of justifying the purchase of Product B was increased by indicating there were other means by which Product A would be acquired, role integration did not impact purchase likelihoods. By contrast, and replicating the previous studies, role integration again drove purchase likelihoods when there were no alternative means for acquiring Product A. So, once again, we find evidence that the influence of role integration on purchase likelihoods is driven by how easily justified the purchase is.
Study 5: Replicating the Effect with Real Choices
Study 5 had two main goals. The first was to examine if the pattern of results obtained so far can be replicated when participants make real-world choices (i.e., choices that would affect the type and amount of financial compensation they would receive), and thus Study 5 used an incentive-compatible design. The second was to replicate the result of Studies 3 and 4 while holding the focal roles constant and asking participants if those roles are integrated or segmented only after they respond to the focal dependent measures in the study. This study was preregistered (https://aspredicted.org/976fx.pdf).
Method
Participation in this study was restricted to Prolific participants who indicated in the Prolific screening questions that they (1) were currently a student and (2) had played a musical instrument for more than two years. Their responses to these Prolific screening questions were reconfirmed at the beginning of our study, and if a participant indicated that they were neither a student nor a musician, they were not allowed to participate in this study.
We chose these two screening questions because they maximized the likelihood that each participant would consider “student” and “musician” personal life roles. We selected the student and musician roles because they are commonly associated with internal aspirations and external expectations, which are the defining characteristics of life roles. Although these two characteristics tend to be readily apparent for student roles, it was only through in-depth interviews with community band members in our second pilot study that we realized musicians consider music an integral part of their lives and often socialize around this activity in various musical ensembles, which is where external expectations tend to arise.
The study began by giving 152 prescreened participants (Mage = 23.36 years, 37.5% male) a brief description of life roles and several examples of common roles most people have, as in Studies 3 and 4 (for complete details and instructions, see the Web Appendix). Participants then read that it is common to have specific needs and wants for our various life roles, which require purchasing specific goods or services. Next, participants indicated the name of one retailer from which they could make online purchases pertaining to their student role, and they provided the URL for that retailer's website. They then repeated this task for their musician role. The survey software piped the names of the two retailers into the text of the remainder of the study. Next, participants read that, as a thanks for participating in the study, they would be given ten virtual raffle entries, and each raffle entry would give them a 1-in-500 chance of winning a $50 gift card to the retailer they listed for their student role. The drawing would occur within a week of the study's completion, and the winner would be contacted to claim their prize.
To induce participants to earmark these funds explicitly for purchases for their student roles, they were asked to specify the item(s) they would purchase for their student role with the $50 gift card. A follow-up question asked them to indicate why they would purchase that (those) items(s) and how such item(s) would help satisfy a need or want specific to their student-life role. Then, participants were reminded (1) that they had ten virtual raffle entries (i.e., ten chances to win a $50 store gift card) and (2) of what they wanted to purchase with the $50 store gift card.
At this point, the focal dependent measure was administered. Participants read that there was a second raffle for which they could use none, some, or all ten virtual raffle entries. This alternative raffle was for a $55 gift card to the retailer associated with their musician role. So, participants could trade off chances to win a $50 gift card to the retailer they chose for their student role, which was already earmarked explicitly for purchases for their student role, to gain chances to win a $55 gift card to the retailer they chose for their musician role. Thus, our focal dependent measure was the number of virtual raffle entries, out of ten, that participants chose to allocate to the raffle for the $55 gift card to the retailer they chose for their musician role.
Once the participants indicated the number of virtual raffle entries they would allocate to each raffle, they read the same brief descriptions of the differences between integrated and segmented life roles as in Study 1 and then indicated whether they considered themselves to be a person with either integrated or segmented student and musician roles (i.e., participants self-categorized their chronic role integration tendencies for these specific roles). We expected that those who considered their student and musician roles to be segmented (integrated) would allocate significantly fewer (more) virtual raffle entries to the raffle for the $55 gift card to the retailer they chose for their musician role, even though doing so would give them a chance to win a greater financial reward.
Results and Discussion
A one-way ANOVA found that participants with more segmented (vs. integrated) student and musician roles allocated fewer of the ten virtual raffle entries, which they previously earmarked for their student role, to the musician-retailer raffle (Msegmented = 2.75, SD = 3.10, N = 76 vs. Mintegrated = 3.91, SD = 3.33, N = 76; F(1, 151) = 4.93, p = .028, d = .36), thus replicating the previously documented effect with consequential choices.
General Discussion
Prior research has typically focused on how the malleability (Cheema and Soman 2006) and framing (Atlas and Bartels 2018; Gourville 1998) of expenses and accounts moderates the consequences of mental accounting on consumers’ spending behaviors. The current research takes a different tack by focusing on how an idiosyncratic characteristic of the accountant (i.e., the person forming the mental accounts, making the purchase decisions, and tracking the expenses) can also moderate the consequences of mental accounting. Specifically, we found that consumers are significantly more likely to use funds earmarked for the purposes of one life role (e.g., money allocated for work-life expenses such as professional apparel) to satisfy the needs/wants of another life role (e.g., purchasing a computer monitor for use at home) when those two roles are more integrated (vs. segmented). We contend that this occurs because greater role integration—the extent to which the boundaries between life roles (e.g., “spouse” and “executive”) are flexible and permeable, thereby affording greater connectivity and a sense of being more the “same person” with shared goals across the roles—renders such purchases more easily justifiable. Five studies support our contentions.
Using a self-reported measure of role integration, Study 1 demonstrated that those with more integrated (vs. segmented) roles are more (less) likely to spend resources earmarked for the purposes of one role on the needs/wants of the other role. Building on the premise that role integration can vary both chronically across individuals and transiently within individuals (i.e., one can believe the two life roles are more or less integrated at different points in time), Study 2 replicated this basic effect after manipulating the perceived integration of two life roles between randomly assigned participants. In so doing, Study 2 casts doubt on any explanation of the results that suggests that there are other inherent differences (i.e., confounds) that exist between those with more or less integrated roles. Whereas Studies 1 and 2 focused on the general and fairly ubiquitous home-life and work-life roles, Studies 3 and 4 allowed participants to choose any life roles and products, thereby generalizing the results to a wide range of roles and products. Finally, Study 5 demonstrated the basic effect arises when participants are making consequential (i.e., incentive-compatible) choices, thereby demonstrating the robustness and generalizability of our results. In addition, Studies 2 (mediation), 3 (mediation), and 4 (moderation) found consistent evidence that people with more integrated roles are more likely to spend funds earmarked for one role on the needs/wants of the other role because they consider such purchases more easily justified.
Our experimental designs cast doubt on several alternative accounts as well. First, as mentioned previously, the results generalize across both measured and manipulated role integration, and thus it is unlikely the results were caused by other co-occurring individual differences. Second, Studies 3, 4, and 5 found that the effect arises even when the funds under consideration were earmarked for a specific purpose/purchase, as opposed to allocated to a broad spending budget such as entertainment. Because the funds were associated with a specific purpose/purchase, it is unlikely that the results were the consequence of those with greater role integration forming broader, more inclusive, or more malleable mental accounts. Choosing to spend money allocated for one purchase on a different purchase is a decision that does not require one to assess how well each purchase fits the intended criteria of a general spending account. We additionally present a study in the Web Appendix in which we find no significant relationship between role integration and the breadth of mental accounts that people construct and maintain. Although null results are far from conclusive evidence, the combined results of Study 3, Study 4, Study 5, and the Web Appendix study converge on the conclusion that role integration does not reliably impact consumers’ spending decisions by changing how they construct or maintain mental accounts. Finally, because Studies 3 and 4 used more common purchase-specific cases of mental accounting and allowed participants to choose the roles and products, the results are unlikely a mere artifact of contrived and potentially artificial experimental contexts. In summary, the evidence strongly supports the contention that role integration increases the likelihood of spending funds earmarked for one role to satisfy the needs/wants of another role because such purchases are considered more justifiable.
Given that this research is the first to examine the intersection of the identity/role and mental accounting literatures, the opportunities for future research are abundant. First, although we examined role integration as an individual difference variable, it is possible that certain roles are more likely to be integrated than others (e.g., spouse and parent). Future research could, therefore, examine whether certain role-aligned accounts have greater cross-account fungibility. Second, whereas mental accounting typically focuses on monetary funds, many similar principles can be applied to other resources, such as time (Rajagopal and Rha 2009). It could be worthwhile to explore the effects of role integration on the budgeting of temporal resources. Third, it is important to examine how role conflict could moderate our effects. If two roles are in conflict, people may be more likely to view a benefit for one role as a cost to the other. Fourth, one could view spending money from a mental account corresponding with one role to service the needs/wants of the other role as taking a risk. It could, therefore, be interesting to examine role integration in the context of risk taking. Finally, could products be marketed as serving multiple roles? How could consumers’ role integration affect the way they classify an expenditure that serves more than one role?
Future research could also examine important downstream implications of our findings. For example, what implications do these findings have for consumer spending tendencies more broadly? Is it possible that role integration, by increasing the fungibility of funds, increases spending more generally or increases spending via credit? As an initial examination of the latter question, we measured role integration and propensity to overspend on credit cards (Sotiropoulos and d’Astous 2013) among 171 MTurk workers and found a significant positive correlation (p = .03). Although these results are preliminary, they suggest that there may be interesting downstream implications on consequential financial behaviors.
The findings also have implications for managers. Although it may appear that little can be done regarding consumers’ tendencies to integrate or segment their life roles, this is not the case. We offer three suggestions for the manager who is concerned with the impact of role integration on sales of their product or service. First, consumers with integrated roles may be an identifiable and targetable consumer segment in specific circumstances, particularly when considering home-life and work-life role integration. For instance, consumers who telecommute for work or use on-site daycare services likely have a higher degree of integration between their work-life and home-life roles. These individuals may be less constrained in their spending than consumers with more segmented roles, at least for products that correspond with those roles. Alternatively, the firm may influence consumers’ transient role integration perceptions via the messages in its advertising, direct marketing, or personal selling channels. Lastly, it is interesting to think about a firm potentially positioning and framing its product or service as benefiting multiple roles. Although it seems likely that consumers will prefer products with many benefits, products that additionally benefit multiple roles may be more appealing to those with both integrated and segmented roles. In other words, touting the many work-life benefits of a given product or service is worthwhile, but there may be an additional bump provided by touting how the product benefits the customer's work and home life. For example, positioning a casual shirt, which is typically associated with home life, as a “casual Friday” office wear option could increase the likelihood that people with segmented work-life and home-life roles purchase it.
All articles have limitations, and this article is no different. First, although we demonstrated that the documented effect generalizes (1) to measured and manipulated role integration and (2) across many life roles and products, all the studies in this article were controlled experiments. Of course, we chose this path to maximize the internal validity of our studies. However, as we discuss subsequently, this leaves opportunities for future work to examine the boundaries of this relationship on multiple dimensions. Another potential limitation is that concurrently measuring or manipulating role integration and examining purchase decisions may have generated demand effects. Although that is possible, our stimulus and design sampling across the five studies, including one study with an incentive-compatible design, reasonably reduce the likelihood of pervasive demand effects.
Future research could examine the impact of role integration on spending behaviors in less controlled but more ecologically valid contexts, perhaps using an ethnographic observation approach. In addition, it could be worthwhile to identify real-world factors that influence the extent to which consumers integrate various roles. For instance, we mentioned that role integration has been used primarily to explain the effectiveness of organizational policies, such as on-site day care facilities (Hall and Richter 1988), on employee satisfaction, commitment (Rothbard, Phillips, and Dumas 2005), and productivity (Kossek and Ozeki 1999). In these cases, the general proposition is that organizational policies should match the consumers’ chronic role integration tendencies. However, as we have shown, exogenous factors can also influence consumers’ transient role integration tendencies. This highlights an important question: To what extent can role integration measured or observed at one time predict spending behavior at a later time? The answer is far from simple and depends on the amount of variance in role integration between a given set of roles that is explained by the individual's chronic role integration versus the various internal motivations and external influences that weigh on a person at any given time. Assessing the stability of role integration (broadly or between specific roles) is outside the scope of the current work, but it will certainly speak to the value of role integration as a tool for marketers.
More broadly, there is a need for real-world evidence of the phenomenon we documented, beyond experiments that are incentive compatible but heavily controlled, such as Study 5. Such evidence may come in the form of field studies, analysis of archival data, or observations of actual shopping data. Of course, the challenges and complexities of examining the consequences of phenomena such as role integration in the real world are substantial. But previous research on role integration in real-world contexts can motivate and guide the way. In summary, this article examines the intersection of two previously disparate theoretical domains. Although this provides insight for both domains, our results also open many questions. We hope future research investigates these and other questions.
In closing, it is worth noting that we conducted a significant portion of the current research during the height of the COVID-19 pandemic, which forced the collision of numerous roles in most people's lives. For many, our work lives have intruded on our personal lives and vice versa. Likewise, our social lives have invaded our family lives and, again, vice versa. It is interesting to consider what consequences the COVID-19 pandemic has had on people's preferences and tendencies for integrating life roles. Although we have no direct evidence, we suspect that the consequences are more nuanced than one might suspect. Certainly, the co-occurrence of two roles drives them toward integration in many ways. However, if such integration is unwanted or problematic, it motivates the individual to segment those roles even further. That our results hold now does not speak to the general ubiquity of role integration or segmentation, and future researchers may also consider this a fruitful path for future research.
Supplemental Material
sj-pdf-1-mrj-10.1177_00222437221112058 - Supplemental material for Role Integration Increases the Fungibility of Mentally Accounted Funds
Supplemental material, sj-pdf-1-mrj-10.1177_00222437221112058 for Role Integration Increases the Fungibility of Mentally Accounted Funds by Iman Paul, Jeffrey R. Parker and Sara Loughran Dommer in Journal of Marketing Research
Footnotes
Acknowledgments
The authors would like to thank Abigail Sussman for her comments on an earlier draft of this article.
Associate Editor
Manoj Thomas
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
