Religious Priming Increases Help if Requesters Dress Poorly: Effects of Dress, Priming, and Task Difficulty on Helping Behaviors

The present study investigated requester–helper interaction, helper personal disposition, and their interactions in effecting helping behavior. In a 2 × 2 × 2 experimental design, participants were asked to help with an effortful or effortless task by a well-dressed or poorly dressed help requester after being primed with religious or neutral words. Results showed that well-dressed requesters, religious primes, and effortless tasks independently increased helping behavior, and helping with effortful task requested by poorly dressed requesters particularly increased after religious primes. These findings provide an integrated understanding of the micro and meso causes of helping behavior.


Introduction
According to a multilevel perspective, prosocial behaviors are explained by causes in three levels, namely the micro level, involving a helper's personal disposition, the meso level, involving helper-recipient interaction, and the macro level, involving group and societal contexts (Penner, Dovidio, Piliavin, & Schroeder, 2005).However, studies of prosocial behaviors have focused mostly on one level and have rarely provided an integrative understanding of interactions across different levels.The present study explored the interplay between the micro level and the meso level by identifying one cause from each level and studying the interaction between the two causes.This study also explored strategies that can increase help in a one-time help request; macro-level causes were thus not included in this study because they involve organizational and societal contexts that are unchangeable in a short time.Whether a person desires to help is determined by various factors, including personal tendencies such as primed mindset, empathetic affects, other positive emotions, and agreeable personality (Ashton, Paunonen, Helmes, & Douglas, 1998;Davis, Luce, & Kraus, 1994;Eisenberg, Fabes, Guthrie, & Reiser, 2000;Abbate, Boca, Spadaro, & Romano, 2014), as well as requester-helper interaction such as the relationship between the two, the appearance of requesters, and the helper's emotional arousal stimulated by the requester (Ainscough & Motley, 2000;Batson & Moran, 1999).Priming was considered as a micro-level cause in this study because other micro-level causes are personal traits that cannot be manipulated or studied for short-term or one-time effect.The clothing of a help requester was considered as a meso-level cause because clothing substantially affects the appearance of the requester and easily and immediately influences a potential helper's decision to help (Long, Mueller, Wyers, Khong, & Jones, 1996).

Priming of Religion
People exposed to particular primes show increased prosocial behaviors including helping strangers.For example, after exposure to words related to help in a sentence construction task, participants became more likely to assist another person whom they barely knew with dropping pens, compared with participants who were exposed to only neutral words (Macrae & Johnson, 1998).Priming concepts of friendship elicited increased helping behaviors in participants toward strangers, particularly when a stranger resembled a friend of the participants (Berk & Andersen, 2000;Fitzsimmons & Bargh, 2003).Implicit bystander effects showed that participants primed with the scenario of being alone were more likely to offer help than those primed with the scenario of being in a crowd (Abbate et al., 2014;Garcia, Weaver, Moskowitz, & Darley, 2002).
Among various information that can be primed in people to effectively increase helping behaviors, religion-related information was adopted in this study for three reasons.First, as micro-level causes of prosocial behaviors that represent personal disposition (Penner et al., 2005), faiths are implantable in people throughout their lifetime, in contrast to other primes such as scenarios of crowdedness and friendship reminders (Batson, Schoenrade, & Ventis, 1993).In accordance with the purpose of this study, which was to seek ways to promote help toward strangers, information that can be primed in people in a short time and that are crucial to people's exposure to religious beliefs in the long term were applicable.Second, people with religious beliefs have high social desirability that prepares them to access prosocial thoughts and engage in prosocial behaviors (Trimble, 1997).Compared with other primes such as friendship-related information that concerns a relatively small group of recipients, religious primes have wider social implication because they enhance coalition and reciprocity of large groups (Kirkpatrick, 2005).Third, because religious primes are not immediately related to help, they can more likely prevent participants from guessing the linkage between primes and the invitation to help than can primes that are directly about help.Although religious primes are indirect primes for help, they are effective in eliciting help even if the priming information is presented briefly and unconsciously (Pichon, Boccato, & Saroglou, 2007).
People with religious beliefs are generally helpful although they have preferences and egoistic purposes in helping (Batson et al., 1989;Saroglou, Pichon, Trompette, Verschueren, & Dernelle, 2005).They tend to help religious charities rather than nonreligious groups (Hunsberger & Platonow, 1986) and the native homeless rather than the immigrant (Pichon & Saroglou, 2009); people intrinsically oriented to religion prefer to participate in volunteer work that is planned rather than to offer help in response to unexpected requests (Hansen, Vandenberg, & Patterson, 1995).Various reasons why religious people help have been identified.Religious thoughts evoke feelings of siblinghood and extend biological kinship to societal kinship, thereby promoting prosocial behaviors in the extended kinship (Batson et al., 1993).Responsibilities to fulfill God's mission (Einolf, 2011) and feelings of being watched by God (Milinski & Rockenbach, 2007) are also reasons for prosocial behaviors.Moreover, rather than focusing on the need of the help requester, religious people are driven to help because of social rewards such as enhanced reputation and self-rewards such as positive mood (Batson & Flory, 1990).
Various studies have reported that priming people with religion-related information increased helping behaviors.Religious primes, even without people's awareness, activate the cognitive network related to religious concepts such as extended brotherhood, mission of God, a just world that repays virtuous deeds, and others that reduce one's threshold of offering help.After subliminal exposure to religion-related words, participants took away more pamphlets about volunteer work than did participants exposed to only neutral words (Pichon et al., 2007).Participants primed with God-related ideas in a sentence construction task with religious words donated more money than those primed with neutral information (Shariff & Norenzayan, 2007).The priming effect of religion was effective even when the religion of the primes was different from the religion that the participants identified with (Clobert & Saroglou, 2013).Therefore, we hypothesized that religious priming, compared with neutral priming, elicits more help in participants.

Dress
Compared to distressed facial expression that requires careful observation and strong empathy to identify, clothing is obvious and has direct and instant influence on a potential helper's perception of the requester.Well-dressed requesters are more likely to receive help than are poorly dressed requesters.For instance, a woman dressed neatly received more money than a woman dressed sloppily while asking for a dime to make a phone call in an airport (Kleinke, 1977).Well-dressed people calling for help by a roadside were more likely to stop passing cars compared with poorly dressed people (Mallozzi, McDermort, & Kayson, 1990).People dressed in a business suit rather than a work uniform received more help while searching for lost money (Bickman, 1971).
A few reasons explain why well-dressed people are well received.First, the perceived attractiveness of well-dressed requesters is higher than that of poorly dressed requesters (Benson, Karabenick, & Lerner, 1976).People like the attractive more than the unattractive and hence are more willing to help attractive people.Moreover, attractive requesters induce positive mood rather than disgust in potential helpers, and the positive mood causes positive thoughts and intention of behaviors such as help (Rind, 1997).Association between positive moods such as happiness and prosocial behaviors was identified (Schaller & Cialdini, 1990).Second, desirable attire, similar to luxury accessories, is a sign of high status (Solomon & Herman, 1977), and people prefer to help the high-rather than the low-status because the high-status are believed to be more capable of repaying a favor (Flynn, 2003).The expectation for reciprocity may not be conscious when people offer help; however, potential benefits gained from helping affect one's willingness to help (Lu & Chang, 2011).Therefore, among other signs of high status, such as the possession of an expensive car and decent occupation (Goodman & Gareis, 1993;Solomon & Herman, 1977), dressing well is effective in inducing help.Third, the well-dressed are perceived as more trustworthy than the poorly dressed are.People hesitate to offer help partly because of uncomfortable feelings toward the requester (Long et al., 1996), particularly when the requester appears to be dishonest.For example, when requesters ask for money, people assume that the well-dressed requesters are truly in need otherwise they would not seek help, whereas the poorly dressed requesters are habitual beggars who are not in urgent need (Kleinke, 1977).Therefore, we hypothesized that people are more likely to help when the help requester is well-dressed rather than poorly dressed.

Why Interaction between Priming and Attire?
Studying ways to increase the likelihood of helping behaviors is crucial in the first encounter between a potential helper and requester to ensure that people in urgent need are likely to receive help from the society.According to the multilevel analysis of help (Penner et al., 2005), various reasons from the micro and meso levels affect one's decision on helping.We studied two reasons that could be manipulated, one from each level, and their interaction to seek ways to increase assistance between strangers.Personal traits and characteristics such as agreeableness are the micro-level reasons for help; however, one's predisposition to help and habitual prosociality cannot be cultivated during a short encounter between helper and requester.One micro-level reason that can function within a short time and immediately affect social behaviors is priming, which activates cognitive networks that are associated with one's personal traits, attitudes, and knowledge (Greenwald et al., 2002).Within a short time, either explicitly or even implicitly, primes dispose a person to particular behavioral tendencies according to the purpose of priming (e.g., Oda, Niwa, Honma, & Hiraishi, 2011;Shariff & Norenzayan, 2007).Therefore, priming was adopted as a micro-level variable of reasons for help in this study.
The clothing of the requester is one meso-level reason of help because it affects the helper-requester interaction.Clothing is a potential cue that indicates the requester's social status (Goodman & Gareis, 1993) and it effectively affects a potential helper's perception of the requester's reliability and neediness (Long et al., 1996).Although clothing is obvious in interpersonal perception, it is less likely to change.A person of low social status may find difficulty in dressing well to elicit help.With this constraint, requesters must seek ways to increase the likelihood of receiving help.One feasible way is priming that enhances behavioral prosociality in a potential helper.Priming probably means little when a requester is well dressed; however, when a requester is poorly dressed, a supplementary way to elicit help is crucial.We hypothesized that the attire of the requester interacts with religious priming on the helper.The likelihood of help was higher in religious priming than in neutral priming conditions when the requester dressed poorly, whereas the likelihood of help was equal in religious priming and neutral priming conditions when the requester dressed well.

Overview of the Present Study
We studied whether people help and the willingness to help in a field setting.When people behave altruistically, they calculate the costs and benefits of helping.For the same helping task, helping a stranger is more costly compared with helping friends and blood relatives because one's future interaction with a stranger is less and the chance of receiving reciprocation is low.If no return benefit is expected from a stranger, the cost of help is critical in one's decision to help (Stewart-Williams, 2007).In real life, a help request can be costless or costly.Costless help requires a short time, little physical effort, and little financial cost, whereas costly help requires a longer time, more physical effort, and more financial expenditure.Thus, the characteristics of help, represented by difficulties of the help task, were also included in this study.
To explore the effects of clothing, primes, and task difficulties on helping behaviors in a real-life situation, participants resting alone in a shopping mall were invited to help an investigator who dressed well in half of the cases and dressed poorly in the other half of the cases.The research sites were rest areas in a shopping mall because people there were assumed to be relatively available and less occupied with personal business compared with people who were walking or standing elsewhere.The research sites in this study were similar to that of airport waiting areas in which other field studies of helping behaviors have been conducted (Fitzsimmons & Bargh, 2003;Kleinke, 1977).After being primed with neutral words or religion-related words, participants were requested to help the investigator with a task that should be completed at the same venue (effortless) or at another venue that required walking for a few minutes (effortful).Positive and negative answers to the help request and the extent to which a participant was willing to help were recorded.We had four hypotheses: (a) more help occurs when a requester dresses well than poorly; (b) participants primed with religion-related words are more likely to help than those primed with neutral words are; (c) participants are more likely to help with an effortless than an effortful task; (d) the effect of religious primes was stronger when the help requester dressed poorly than well.

Participants
In a rest area of a shopping mall, a male investigator recruited participants by randomly approaching potential participants and asking them to complete a word game that took three minutes.If the investigator was declined, he tried another target in another rest area in the same shopping mall.Participants were included if they agreed to proceed with the word game.Recruitment ended when both poorly and well-dressed conditions each had 80 participants.The mean age of the 160 participants (37 males) was 40.64 with a standard deviation of 12.01.

Design
This study was a 2 (dress: poorly vs. well) × 2 (prime: religious vs. neutral) × 2 (task difficulty: effortful vs. effortless) between-subjects design.Participants were randomly assigned into one of the experimental conditions, with 20 participants being in each condition.

Materials and Procedure
Priming.The word game that participants were invited to complete was the prime.It was a phrase construction task in which participants had to find four characters corresponding to a Chinese phrase out of five Chinese characters placed in a random sequence.After constructing a phrase, participants had to read it out to the investigator.This task of phrase construction from scrambled characters was adapted from previous works that supported the link between religious priming and prosocial behaviors (Ahmed & Salas, 2011;Shariff & Norenzayan, 2007).Under the neutral priming condition, participants had to complete ten neutral phrases that were plain descriptions of natural phenomena (e.g., warmth in winters and coolness in summers; in Chinese: Dong Nuan Xia Liang; participants had to construct this phrase from five scrambled characters: Dong Li Liang Nuan Xia).Under the religious priming condition, participants had to complete seven phrases related to religion and virtuousness (e.g., God loves the world; in Chinese: Shen Ai Shi Ren; participants had to construct this phrase from five scrambled characters: Ai Zu Ren Shi Shen) and three neutral phrases.The mixture of seven religion-related and three neutral phrases in the religious priming conditions was designed to prevent participants from guessing the purpose of the study.
Under the poorly dressed condition, the male investigator wore a wrinkled old shirt, a pair of dirty jeans, and a pair of slippers.Under the well-dressed condition, the investigator wore a business suit with a tie and a pair of leather shoes.He approached a potential target in a rest area of a shopping mall and invited him or her to complete a word game.Upon agreeing to complete the game, the participant was presented with a sheet on which scrambled characters of ten phrases were written.After reading aloud all four-character phrases that the participant had identified, he or she was asked to help the investigator with an additional phrase construction task similar to the previous one.In the effortless condition, the participant was able to complete the task at the same venue, whereas in the effortful condition, the participant was asked to walk to another rest area at the other end of the same shopping mall because the question sheet was at that investigation site.The hassle of walking and the longer time needed to complete the task were assumed to be more costly for participants compared to the task in the effortless condition.Whether a participant wanted to help with the additional task was recorded.If the participant agreed to help, he or she was further asked to indicate his or her willingness to help on a 6-point Likert scale with 1 representing least willing to help and 6 representing very much willing to help.
Participants did not have to actually complete the additional phrase construction task.As a manipulation check, they were asked whether they knew about the linkage between the word game and the subsequent help task.Participants were then debriefed with a consent form for the study and a gift as reward, and their age and religious belief (Buddhism, Christianity, Islam, Others, Atheism, or Agnosticism) were recorded at the end of the study.

Results
In order to recruit 80 participants for each condition, 221 and 131 people were approached by the investigator under poorly and well-dressed conditions, respectively.That is, 141 and 51 people declined the first approach from the investigator.The head count showed that the investigator received more rejection when he was poorly dressed than well dressed (χ 2 (1) = 20.52,p < 0.001.).
The numbers of participants who agreed to help and who rejected to help with the additional task in different conditions are shown in Table 1.The results showed that more participants agreed to help in the religious priming condition than in the neutral priming condition, χ 2 (1) = 14.63, p < 0.001.More participants helped in the well-dressed condition compared with the poorly dressed condition, χ 2 (1) = 3.66, p = 0.06, and more participants helped in the effortless task condition than in the effortful task condition, χ 2 (1) = 5.71, p < 0.05.Neither gender differences nor religious belief differences were correlated with agreeing to and rejecting to help, χ 2 (1) = 0.85, p = 0.36 and χ 2 (2) = 2.00, p = 0.37, respectively.These results showed that the experimental conditions affected whether participants helped or not, whereas demographic variables had no effect on whether they helped.
Thirteen out of 160 participants were aware of the linkage between the priming task and willingness to help.They were hence excluded from further analysis because the awareness of the priming task may have affected the response of willingness to help.The willingness to help of participants who rejected to help was coded as 0.

Discussion
Participants' willingness to help under various conditions of priming, clothing of the help requester, and task difficulties was examined in this study.Results showed that religious primes compared with neutral primes, a well-dressed requester compared with a poorly dressed requester, and an effortless task compared with an effortful task increase the likelihood of help.Most participants agreed to help with an additional task; this was not an unexpected result because the participants had already accepted the invitation to complete a word construction task, and hence they were more likely to help than were people who had declined the invitation from the investigator at the beginning.Among those who had agreed to help, the participants varied in willingness to help under different experimental conditions, with willingness to help being highest in the condition with religious primes, decent attire, and an effortless task, and lowest in the condition with neutral primes, poor attire, and an effortful task.
When the help requested was effortless to the participants, both the decent attire of the requester and the religious primes were effective in increasing help.However, when the help requested was effortful, the effect of religious primes on increasing help was not significant if the requester dressed well, but was substantial if the requester dressed poorly.These results were partly consistent with our hypothesis of an interaction between priming and attire, showing that the interaction occurred but only in the effortful help task condition.A possible interpretation is that cost-benefit calculation substantially affects one's decision to help.Costless help with little effort is affordable and thus people are less concerned about reciprocation (Kruger, 2003;Stewart-Williams, 2007).Decisions of effortless help are more likely to be affected by the instant mood (e.g., Isen & Simmonds, 1978), primed mental states (e.g., Scaffidi Abbate et al., 2014;Shariff & Norenzayan, 2007), and helper-requester interaction (e.g., Stewart-Williams, 2007).By contrast, when the help is effortful, people are more likely to calculate the chance of receiving return benefits.Willingness to help is assumed to be low if the helper expects no return from the help requester, regardless of whether the requester dresses well or poorly and whether helpers are primed with religious or neutral information.When making helping decisions, people also take into consideration their own helping conditions.For example, individuals who are concerned about their own economic conditions were less likely to offer others financial aid (Long et al., 1996).Therefore, primes on potential helpers and the attire of help requesters can affect the willingness to help, but only to a particular extent if the requested help is costly.
Religious primes on participants were effective particularly when the help requester was poorly dressed.These results suggested that people with constraints of appearance and status can adopt other strategies to increase help, such as activating the cognitive network associated with religion by priming potential helpers.In this study, religious words were used as primes in a word construction task.Although only two to three minutes were needed to complete the task, the time was considered long in the applied settings.Future studies can explore more feasible ways to prime religious mental states in real-life situations.For example, the primes can be posters of religious symbols such as portraits of Gods and the Cross, or accessories of religious figures and articles worn by help requesters.Symbols of eyes such as an eye-like painting (Oda et al., 2011), a pattern resembling a face with a pair of eyes (Haley & Fessler, 2005), and the eyes of a robot (Burnham & Hare, 2007) have increased cooperation and prosocial behaviors and reduced undesirable social behaviors.Eye symbols activate people's awareness of reputation because eyes resemble the presence of audiences, and also the awareness of being watched by the supernatural (Milinski & Rockenbach, 2007).Religious symbols, in addition to activating feelings of being watched, also activate feelings of extended siblinghood, which enhance prosocial and altruistic behaviors (Batson et al., 1993;Einolf, 2011), because people feel positively towards kin and prefer to help kin more than strangers (Lu & Chang, 2009).
This study contributed evidence to the integrative understanding of interplay between micro-level and meso-level causes of helping behavior, with priming representing the micro level because it affects personal disposition, and attire of the requester representing the meso level because it affects interpersonal perception in requester-helper dyads.The two variables were manipulatable in the experimental design, which explored ways for increasing help.Future studies can examine other variables, either unchangeable or manipulatable, from the micro and meso levels and explore their interactions.For example, the mood and emotionality of the helper are micro-level variables; dispositional positive emotionality is positively correlated with help, and induction of positive mood increases help (Eisenberg et al., 2000;Rind, 1997).Shared social identity between helper and requester is a meso-level variable that affects the likelihood of help.Help is positively correlated with closeness in relationship, and priming common membership of a group increases help between different group members (Gaertner et al., 2000).Studies on possible interaction between emotion and group identity in affecting helping behaviors will improve the understanding of micro and meso factors that increase help.
This study had several limitations.First, the investigator was a man, and the reaction of potential female helpers and male helpers toward a male requester can be different.However, effects of requester's gender on eliciting help were mixed and sometimes men and women equally received help (Long et al., 1996;Mallozzi et al., 1990).
To maintain consistency of the help requester, this study included only one investigator.Second, there were more female than male participants in this study because the research site was in a shopping mall.However, no conclusive findings showed whether men or women were more likely to help (Eagly & Crowley, 1986).The disproportion of gender may have little effect in biasing the results.Third, few participants had religious beliefs in this study.Including more religious participants ensures that the proportion of theists and atheists is even.Nonetheless, religious priming was effective in both theists and atheists (Shariff & Norenzayan, 2007), and cross-religion priming in which the primed religion and participants' religion were different was also effective (Clobert & Saroglou, 2013).Whether participants had religious beliefs should not affect the results of this study.Despite these limitations, this study provided an integrative understanding of the effect of priming on helpers and the attire of requesters on increasing helping behaviors.

Figure 1 .
Figure 1.Willingness to help as a function of priming, dress, and task difficulty (error bars indicate ± 1 standard errors)

Table 1 .
The numbers of participants who agreed and who rejected to help in different conditions