When door holding harms: gender and the consequences of non-normative help

Abstract

This work explored the potential negative consequences of unexpected help. A behavioral observation and a survey study found that men are unlikely to have the door held open for them in a chivalrous manner, whereby they walk through the door before the person helping them does. In an experimental field study, passersby were randomly assigned to experience this type of door-holding help or not. Males who had the door held for them in this manner by a male confederate reported lower self-esteem and self-efficacy than males who did not have the door held for them. Females were unaffected by door-holding condition. These results demonstrate negative consequences of seemingly innocuous but unexpected helping behavior that violates gender norms.

Keywords::

• helping
• Threat to Self-Esteem
• expectations
• gender
• gender norm violation

Although receiving help from others is often associated with positive outcomes, responses to being helped are not always positive. Indeed, the Threat to Self-Esteem Model of reactions to help suggests that receiving help can be associated with both positive and negative outcomes (Fisher, Nadler, & Witcher-Alagna, 1982; Nadler & Fisher, 1986). This work builds on this model by investigating the consequences of unexpected help in an ecologically valid field experiment. Specifically, this work provides evidence that men are less likely than women to experience and expect chivalrous door-holding help, and that when men do experience this help it is associated with decreases in self-esteem and self-efficacy. Thus, this work demonstrates the potential influence of brief, seemingly inconsequential social behaviors on our beliefs about ourselves.

The Threat to Self-Esteem Model is a comprehensive model of reactions to help that integrates previous research on responses to receiving aid. The central idea is that the relative degree to which help is perceived as self-supporting versus self-threatening influences recipients' reactions to receiving help (Dovidio, Piliavin, Schroeder, & Penner, 2006). When help is perceived as self-supporting, the recipient displays positive responses including positive affect. When help is perceived as self-threatening, the recipient displays negative responses such as decreased self-esteem.

A variety of situational and individual factors may lead to help being perceived as threatening, including the degree to which the help challenges freedoms or suggests the recipient is inferior or too dependent. For example, receiving aid from a similar versus a dissimilar other leads to decreases in self-esteem (Fisher & Nadler, 1974). These results may be due to social comparison processes and the greater salience of inferiority cues when one receives aid from a similar other (Festinger, 1954). Help may also be more threatening on a highly ego-relevant task (Nadler, 1987). In addition, when affirmative action programs are perceived as indicating the relative inferiority of the beneficiaries of these programs, these beneficiaries may experience decreased self-esteem (Turner & Pratkanis, 1994).

Receiving help may also be associated with decreases in self-esteem when the help is non-normative (Tessler & Schwartz, 1972). Researchers have theorized that non-normative help may be threatening because it is inconsistent with one's self-concept (Fisher et al., 1982). Thus, unexpected help may also be perceived as threatening, and thus decrease self-esteem. Research suggests that unexpected help is associated with increased ambivalence, and that receiving unexpected help without anticipating a reciprocity opportunity is associated with fewer prosocial responses (El-Alayli & Messe, 2004). These researchers propose that receiving unexpected help may be associated with increased psychological reactance, or negative feelings that one's freedoms are being constrained (Brehm, 1966; Brehm & Cole, 1966). This work sought to explore the negative consequences of helping in the context of door-holding behavior, as door holding is a simple and pervasive social behavior that is associated with normative gendered expectations.

Door-holding behavior

Work in a variety of fields has investigated gendered door-holding behaviors (Goldman, Florez, & Fuller, 1981; Parker, 1988; Walum, 1974). This work suggests that it is normative for men to hold doors open for women. For example, Renne and Allen (1976) found that men were more likely than women to hold doors open for strangers. In addition, women were more likely than men to have doors opened for them, and men were generally unlikely to have doors held for them by men or women. More recent work suggests that door holding is still a gendered behavior in which men are likely to help women (Yoder, Hogue, Newman, Metz, & LaVigne, 2002). However, this finding only held true in a dating, but not in an everyday, context. Thus, it is important to note that the social norms surrounding door-holding behavior are gendered, but not static.

Although there is a small body of research focused on the degree to which door-holding behavior is gendered, there is little research focused on recipients' reactions to door-holding behavior and the ways in which experiencing this help may influence them. Notably, Ventimiglia (1982) did explore gratitude responses to door-holding behavior, and found the highest levels of gratitude in traditional pairs involving male helpers and female recipients.

The present study

The consequences of unexpected help were investigated in the context of door-holding behavior, as experiencing door-holding help may be expected or not depending on one's gender. However, given that the most recent work on door holding suggests that it is not as gendered as it once was (Yoder et al., 2002), we explored a specific type of door-holding behavior that is associated with chivalry: the door holder opens the door, lets the recipient of the help enter first, and then follows behind him or her. This type of chivalrous door holding may be associated with traditional Western dating scripts (Laner & Ventrone, 2000; Rose & Frieze, 1993). However, as previous empirical work on the gendered nature of door holding has not explored different types of door-holding behaviors, we investigated the gendered nature of this specific type of chivalrous door holding in a behavioral observation study (Study 1A) and a survey study (Study 1B), before exploring the effects of experiencing this helping behavior in an experimental field study (Study 2). We predicted that men will be less likely than women to have someone hold the door for them and let them enter first (Hypothesis 1).

Study 1A

Method

Research assistants sat unobtrusively on a bench at two different academic building entrances on a large university campus. They indicated the following each time an instance of door holding occurred: the gender of the door holder, the gender of the person(s) helped, and how the door was held. Both building entrances had two doors directly next to each other that opened outward, such that individuals entering the building had to pull the door towards them to enter. Two hundred and seventy-six door-holding instances, involving a total of 573 people, were observed. As the gendered nature of receiving door-holding help was of primary interest, 11 instances in which a mixed-gendered group of people received help were excluded from the analyses, leaving a total of 265 instances.

Results

In the majority of observed door-holding instances (62.2%), the door was held such that the door holder walked through the door first, but held the door back for someone following him or her. The second most frequently observed type of door holding (17.6%) was our specific chivalrous behavior of interest in which the door holder opened the door and let the person following enter the building before him or her. For the primary analyses, the manner in which the door was held was coded dichotomously: the chivalrous door-holding type of interest versus any other type of door holding. This dichotomization was done for concision purposes; notably, the below analyses remain very similar when more differentiated types of door holding are included.

In order to explore whether the frequencies of door-holding behavior depended on helper gender (male vs. female), helped gender (male vs. female), and helping behavior (chivalrous vs. other), we conducted a log-linear analysis with all three of these factors. A backward elimination process was used, suggesting that only the helped gender × helping behavior interaction and main effects should be retained. This final model fit the data well, as the χ² test of model fit was non-significant, χ²(1, N = 265) = 1.15, ns. Main effects of helper gender [χ²(1, N = 265) = 7.68, p = .006] and helping behavior [χ²(1, N = 265) = 122.76, p < .001] were obtained. These effects suggest that men are more likely than women to engage in the social behavior of door holding (155 vs. 110 instances), and that chivalrous door holding is less typical than the combination of all other types of door holding (46 vs. 219 instances). However, these main effects were qualified by a significant interaction between the helped gender and the helping behavior [χ²(1, N = 265) = 11.44, p = .001], indicating that men are less likely than women to experience chivalrous door holding (12 vs. 34 instances, respectively) but that these gender differences do not extend to other types of door-holding experiences (116 vs. 103 instances, respectively). These results provide support for Hypothesis 1 that men are less likely than women to have someone hold the door for them and let them enter first. No other effects reached significance.

Discussion

The data from Study 1A suggest that door holding in which the person holding the door lets the person being helped enter first is common, although not the most common type of door-holding behavior. The results also provide evidence that men are less likely than women to experience this chivalrous type of door holding, although they are not less likely than women to experience other types of door-holding behavior. These results provide support for Hypothesis 1 and suggest that the manner in which a door is held is an important factor to consider when exploring expectations and the gendered nature of door holding. However, given that this work explores the consequences of unexpected help that violates gender norms, and that people's beliefs are not always consistent with reality, we sought supplemental evidence for the gendered nature of chivalrous door-holding behavior from a survey study that directly tapped participants' expectations.

Study 1B

Method

One hundred and sixty-six introductory psychology students (112 male) participated in exchange for partial course credit. The design was a 2 (helper gender: male vs. female) × 2 (helped gender: male vs. female) × 2 (helping behavior: general vs. chivalrous) repeated measures design. Participants were first asked to indicate, “How typical is it to see a male/female hold the door for a male/female?” Participants were then asked to indicate “How typical is it to see a male/female step in front of a male/female and hold the door for him/her such that the male/female being helped would walk through first?” These eight ratings were made on a scale from 1 (not at all typical) to 7 (extremely typical). The first four items assessed participants' expectations of door-holding behavior generally, whereas the second four items assessed participants' expectations of the specific chivalrous door-holding behavior of interest.

Results

A repeated measures analysis of variance including all three factors was conducted on typicality ratings. A significant main effect of helped gender was obtained such that participants rated situations involving helped females (M = 4.73) as more typical than situations involving helped males (M = 3.45), F(1, 165) = 321.75, p < .001. A significant main effect of helper gender was obtained such that participants rated situations involving male helpers (M = 4.68) as more typical than situations involving female helpers (M = 3.50), F(1, 165) = 130.99, p < .001. A significant main effect of helping behavior was also obtained such that participants rated situations involving general door-holding behavior (M = 4.62) as more typical than situations involving chivalrous door-holding behavior (M = 3.56), F(1, 165) = 147.35, p < .001.

These main effects were qualified by a significant interaction between the helped gender and the helping behavior, F(1, 165) = 17.05, p < .001 (see Figure 1). Situations involving general door-holding behavior were rated as more typical when a female was helped (M = 5.16) than when a male was helped (M = 4.08), t(165) = − 13.96, p < .001. Similarly, situations involving chivalrous door-holding behavior were rated as more typical when a female was helped (M = 4.31) than when a male was helped (M = 2.81), although this mean difference is more pronounced, t(165) = − 15.24, p < .001. This pattern of results provides support for Hypothesis 1 that men are expected to be less likely than women to experience chivalrous door-holding behavior.1

Figure 1 The effects of helped gender and helping behavior on typicality ratings. Error bars represent one standard error.

Discussion

The results of Study 1B demonstrate that people believe men are less likely than women to have the door held open for them such that the person holding the door lets them enter first, supporting Hypothesis 1. Notably, Study 1B suggests that people expect door holding to be more gendered than Study 1A suggests it actually is. Indeed, participants expected men to be less likely than women to receive any type of door-holding behavior, although chivalrous door holding was rated as particularly gendered. In sum, both Study 1A and Study 1B provide novel evidence that the manner in which doors are held is a significant factor to consider when exploring expectations and the gendered nature of door holding.

Study 2

Given that men are less likely to experience and expect to receive the specific type of chivalrous door-holding help explored in Studies 1A and 1B, an experimental field study was conducted to explore the consequences of receiving this unexpected helping behavior that violates gender norms. The Threat to Self-Esteem Model of reactions to help suggests that receiving threatening or low-normative help can result in negative consequences including decreased self-esteem (Fisher et al., 1982). Thus, receiving the specific type of chivalrous door-holding help explored in Studies 1A and 1B was expected to have negative consequences for men but not for women. Specifically, men who experience this chivalrous door-holding help were expected to have lower self-esteem than men who did not experience this help, but women's self-esteem was expected to be unaffected by chivalrous door-holding help (Hypothesis 2). Receiving unexpected help that violates gender norms was also predicted to have negative consequences for self-efficacy, or beliefs about what one can accomplish. Men who experience chivalrous help were predicted to have lower self-efficacy than men who did not experience this help, but women's self-efficacy was predicted to be unaffected by this door-holding help (Hypothesis 3).

Methods

Participants and design

Two hundred and twenty-one people (122 female) entering one of two academic buildings on a large university campus served as the participants. The design was a 2 (participant gender: male vs. female) × 2 (helping condition: no door held vs. door held) between-subjects design.

Procedure

The study took place in the same two building entryways observed in Study 1A. These entranceways had two doors directly next to each other, both of which opened outwards. When a person approached the building and was walking alone and not engaged in simultaneous behavior (e.g., not on the phone), they were treated as a participant and randomly assigned to a helping condition. A male confederate would walk next to the person and either hold (or not) the door open for them depending on condition. In the door-held condition, the confederate took a step in front of the participant, opened the door, and let the participant walk through the door first. In the door not held condition, the confederate walked next to the person such that the confederate and the participant each entered the building relatively simultaneously, but through the two different doors that were right next to each other. Thus, in the door not held condition, the participant opened the door for him or herself.

Once participants entered the building, a female experimenter approached them with a clipboard and asked if they had time to complete a short survey. Participants were asked to indicate their agreement with three items from Rosenberg's (1965) Self-esteem Scale: “On the whole, I am satisfied with myself,” “I feel that I have a number of good qualities,” and “I feel that I'm a person of worth, at least on an equal plane with others.” Participants were also asked to indicate their agreement with three personal control items from the Spheres of Control Scale (Paulhus, 1983; Paulhus & Van Selst, 1990): “Almost anything is possible for me if I really want it,” “I can usually achieve what I want if I work hard for it,” and “I can learn almost anything if I set my mind to it.” All items were answered on a scale from 1 (strongly disagree) to 10 (strongly agree). These items were averaged to form composites of self-esteem (Cronbach's α = .80) and self-efficacy (Cronbach's α = .75), respectively. Participants also indicated how they felt on a scale from 1 (very negative) to 10 (very positive). Mood was measured so that any potential mood effects could be controlled for in the primary analyses. The experimenter recorded participant gender and condition regardless of whether the participant was able to complete the survey.

Results

Twenty-five participants were dropped from the analyses due to procedural issues (e.g., opened their own door in the door-held condition, waited to have the door opened for them in the no-door-holding condition), leaving a final sample of 196 participants (112 female).

Participation

Participation in the survey was a dichotomous measure, indicating whether or not the participant agreed to answer the survey. The majority of participants (143, 73%) completed the survey. A logistic regression was conducted predicting participation using subject gender, helping condition, and their interaction as predictors. No significant effects were obtained. Thus, participation rate did not differ by gender or door-holding condition.

Primary analysis strategy

2 (subject gender: male vs. female) × 2 (helping condition: no door held vs. door held) between-subjects analyses of variance were conducted on self-esteem, self-efficacy, and mood.

Self-esteem

A significant main effect of participant gender was obtained such that males reported lower self-esteem (M = 7.88) than females (M = 8.68), F(1, 139) = 17.17, p < .001. A marginal main effect of helping condition was obtained such that those in the door-held condition reported lower self-esteem (M = 8.10) than those in the no-door-held condition (M = 8.47), F(1, 139) = 3.77, p = .054. However, these main effects were qualified by a significant interaction between the participant gender and the helping condition, F(1, 139) = 8.42, p = .004, providing support for Hypothesis 2 (see Figure 2). Male participants reported lower self-esteem in the door-held condition (M = 7.42) than the door not held condition (M = 8.35), t(65) = − 3.19, p = .002. Female participants' self-esteem was unaffected by door-holding condition, t(74) = .73, ns.

Figure 2 The effects of participant gender and helping condition on self-esteem. Error bars represent one standard error.

Self-efficacy

A marginal main effect of participant gender was obtained such that males reported lower self-efficacy (M = 8.16) than females (M = 8.61), F(1, 139) = 3.85, p = .052. No main effect of helping condition was obtained. However, a marginal interaction between the participant gender and the helping condition was obtained, F(1, 139) = 3.70, p = .056, providing support for Hypothesis 3 (see Figure 3). Male participants reported lower self-efficacy in the door-held condition (M = 7.80) than the door not held condition (M = 8.53), t(65) = − 2.22, p = .030. Female participants' self-efficacy was unaffected by door-holding condition, t(74) = .45, ns.

Figure 3 The effects of participant gender and helping condition on self-efficacy. Error bars represent one standard error.

Mood

A significant main effect of participant gender was obtained such that females reported more positive mood (M = 7.85) than males (M = 7.12), F(1, 139) = 5.41, p = .021. No other effects reached significance. Thus, this pattern of findings suggests that the obtained self-esteem and self-efficacy results cannot simply be attributed to differences in mood.

Supplemental analyses

In addition, the male confederate recorded participants' responses to him in the door-holding condition. Each participant's response was coded into one of three categories: a verbal response of “thank you,” a smile but no verbal response, or no response at all. The frequencies of these responses across male and female participants do significantly differ from what would be expected by chance, χ²(2, N = 93) = 8.85, p = .012. The majority of men responded by saying thank you (56%), few men responded with only a smile (3%), and a good proportion of men did not respond (41%). A greater proportion of women responded by saying thank you (77%) or smiling (9%), and fewer women did not respond (14%). These results suggest that women are more likely than men to respond to this chivalrous door-holding behavior by saying “thank you” or smiling, but men are more likely than women not to respond to this behavior.

These results are consistent with the earlier work by Ventimiglia (1982) suggesting that traditional pairings of female recipients of male help are associated with the most positive responses. These findings may also be interpreted as showing support for the role of expectancy in our results. Indeed, when a behavior is not expected, we may react negatively or not know how to best respond to it, which could explain male participants' lack of response in the door-holding condition. Thus, these findings are also consistent with previous research suggesting that receiving unexpected help when one does not anticipate an opportunity to reciprocate is associated with less prosocial responding (El-Alayli & Messe, 2004).

Discussion

The results from Study 2 provide support for Hypotheses 2 and 3. Male participants who had the door held for them by a male confederate reported lower self-esteem and self-efficacy than those who did not have the door held for them. Female participants' self-esteem and self-efficacy were unaffected by door-holding condition. These effects were obtained in the field using a brief helping behavior manipulation without extensive social interaction. Thus, these results demonstrate that experiencing brief, simple, but unexpected helping behaviors that violate gender norms can have negative consequences.

Conclusions

In examining the consequences of unexpected help, this work first found support for the idea that some types of door holding are still gendered in nature. In particular, a behavioral observation study (Study 1A) and a survey study (Study 1B) suggest that men are less likely than women to experience door holding in which they walk through the door prior to the person offering the door-holding help. Previous studies have not investigated the manner in which doors are held, and these results suggest that the manner in which a door is held is an important factor to consider when investigating the consequences of receiving door-holding help.

This work also finds that brief, seemingly inconsequential but unexpected helping behavior that violates gender norms can have unforeseen negative consequences. An experimental field study (Study 2) obtained a significant interaction between participant gender and helping condition on self-efficacy and self-esteem. Male participants who had a door held for them by a male confederate reported lower self-esteem and self-efficacy than male participants in the no-door-held condition and female participants regardless of helping condition. These negative effects of unexpected help are consistent with prior research suggesting that threatening or non-normative help may have negative consequences (El-Alayli & Messe, 2004; Fisher et al., 1982). The lack of a helping condition effect on female self-esteem and self-efficacy is also consistent with prior research that suggests that receiving expected, normative help may not elicit as much of a reaction (Morse, 1972; Morse & Gergen, 1971).

It is important to note that this work has both strengths and weaknesses. The studies have relatively high external validity, as the findings were demonstrated in real-world everyday occurrences. However, field experiments do have limitations on the number of questions that can reasonably be asked of participants. Follow-up experimental lab work may explore additional mechanisms for the current results. For example, the degree to which men feel threatened, belittled, or uncomfortable may mediate the effects of receiving norm-violating help on self-esteem and self-efficacy. Future work may also explore the attributions that recipients of help make for the helper's chivalrous door-holding behavior. Certain types of attributions, such as those that explain the behavior as due to the recipient's need for help, may be particularly threatening for men, challenging the prescription that men are independent and self-sufficient (e.g., Prentice & Carranza, 2002). Indeed, these types of attributions may prove necessary for the obtained effects, suggesting that not all unexpected help may lead to negative consequences. Follow-up work may also explore whether the obtained negative consequences of receiving unexpected, non-normative help have downstream behavioral implications, such as affecting the likelihood that individuals will subsequently offer to help others. Future research may also explore the role of expectations in responses to other common forms of helping behavior.

In conclusion, this work demonstrates that simple but unexpected helping behaviors as fleeting and seemingly innocuous as door holding can have unforeseen negative consequences. Thus, this work contributes to a growing literature on the consequences of helping for the recipients of help (Dovidio et al., 2006), as well as a growing literature on the influence of seemingly inconsequential everyday social behaviors (e.g., Wesselmann, Cardoso, Slater, & Williams, 2012).

Notes

^1. Significant interactions were also obtained between the helper gender and the helped gender (F(1, 165) = 97.20, p < .001), and between the helper gender and the helping behavior (F(1, 165) = 5.76, p = .018). These effects are not discussed in detail for the sake of concision, as they are not germane to the main hypotheses. The three-way interaction between the helper gender, helped gender, and helping behavior was not significant.