The effect of the word “Loving” on compliance to a fundraising request: evidence from a French field study


Nicolas Guéguen, Université de Bretagne-Sud, Rue de la loi, Vannes 56000, France.



In a field setting, students (N = 1800) on different campus locations were asked by a young female confederate for a donation during a health fundraising solicitation (French Téléthon). The solicitor wore a white tee-shirt with different inscriptions: no inscription, “Loving = Helping”, and “Donating = Helping”. Results found that, compared with the no inscription condition, the number of donators increased when the tee-shirt “Loving = Helping” was worn by the solicitor-confederate and decreased when wearing the tee-shirt “Donating = Helping”. The activation spreading theory is used to explain these results. Practical interest for health fundraising organizations is developed. Copyright © 2011 John Wiley & Sons, Ltd.


Why would it be important to know the processes and the techniques that lead people to give to charitable causes? Obviously, the answer is of great practical importance for organizations that rely on fundraising as a source of income. Charitable donations form a substantial portion of income for many nonprofit organizations across the world (Anheier and Salamon, 2006). Thus, knowing the processes and techniques that are predictive of donation become essential for such organizations.

A recent paper of Bekkers and Wiepking (2010), reviewed more than 500 papers around the central question of why people donate to charitable organizations. They have identified a number of papers that focused on solicitation techniques. It seems that the way potential donors are solicited determines the effectiveness of the fundraising campaign.

Fundraising is a significant means for nonprofit organizations to obtain money for their operations. In France, since 1987, an annual Téléthon, for the muscular dystrophy charity in France, L'Association française contre les myopathies (French Association against Muscular Dystrophy), is held each year during the first weekend of December. Several events are organized all around the country. Donations are requested in order to finance medical research programs on several genetic muscular diseases. A host of volunteers solicited people in various locations to give money during this event. This Téléthon is very popular in France, and the amount of donation increases each year. However, it's the only national charitable event that exists in this country. Most of the time, in France, charitable organizations use personalized mail, telephone contact or TV advertising in order to obtain money for their operation (Mermet, 2010). Since 1990, a 5% increase in donations per year has been observed in France whereas the economic growth was around 3% per year during the same period (Malet, 2007). This author also found that if an increase in the number of donators was found during the last decade, only one out of four adults gives money to a charitable organization per year. Thus Mallet (2007) argues that there is a need to develop marketing fundraising strategies to incite French people to donate more frequently.

To gain compliance with a fundraising request, many techniques exist in social psychology literature (see Pratkanis, 2007 for a review). A large part of these techniques use sequential requests: the foot-in-the-door technique (Freedman and Fraser, 1966), the door-in-the-face technique (Cialdini et al., 1975, the low-ball tactic (Cialdini et al., 1978), the lure (Joule et al., 1989), and the 1-in-5 prize tactic (Horvitz and Pratkanis, 2002). Some of them use ingratiation techniques such as flattery (Dunyon et al., 2010), incidental similarity (Burger et al., 2004) or mimicry (Van Baaren et al., 2004). Others use nonverbal behaviors displayed by the solicitor such as tactile contact (Kleinke, 1977), smiling (Solomon, et al.1981) or gazing (Kleinke, 1980). The objective of the experiment presented hereafter was to test the efficiency of a word-priming technique on receptivity to a fund request made during the 2010 annual French Muscular Dystrophy Fundraising Campaign.

Priming and influence behavior

Previous research on priming has demonstrated that the activation of a concept or mental representation can exert an influence on subsequent information processing or behavior. Bargh et al. (1996) found that participants primed with words related to the elderly stereotype (e.g., traditional, retired) walked more slowly than those of a control group when leaving the experiment; also, participants primed with the concept of rudeness interrupted the experimenter more quickly and frequently than did participants primed with the concept of politeness. After exposure to sentences describing stereotyped behavior of dependence (e.g., can't make decisions), participants rated a female target who performed identical behavior as a male target as more dependent; after exposure to stereotyped behavior of aggression (e.g., threatens other people), participants rated a male target as more aggressive than a female target (Banaji et al., 1993). Moreover, impression formation was unrelated to the explicit memory for primes. Priming effects can take place when subliminal priming techniques are employed (i.e., extremely brief exposure to stimuli). Using this technique, Zemack-Rugar et al. (2007) showed that individuals subliminally primed with guilt adjectives were more helpful than those primed with sadness adjectives. Regardless of the duration of exposure to the stimulus, automatic or implicit social cognition occurs when information processing is made without the person's awareness, intention, possibility of control, or effort (Bargh, 1994). To explain by which means mental representations can shape social behavior, theorists have hypothesized mental structures consisting of interconnected information or attributes. The main assumption is the spreading of activation, the activation of one concept being assumed to spread along a network of meaningfully associated information. Activating the concept of gender, for example, would activate the implicit knowledge structure of gender-linked traits, stereotypes, and norms for behavior.

Surrounding cues and behavioral influence

Exposition to sentences, words and figures can also be analyzed as environmental cues that can affect the behavior of the individual. Berkowitz and LePage (1967) found that in the presence of a weapon, a participant in a laboratory administered more electric shocks to a confederate than without the presence of the weapon. In a restaurant, Jacob et al. (2011) displayed various figurative cues related to the sea (a boat or the figurine of a sailor, a napkin with a picture of a boat, and poetry related to the sea) or not (control condition). The results showed that figurative cues related to the sea increased the consumption of fish dishes. In a second experiment, they found the same effect with the word “fish” or fish drawings displayed in several places in the dining room of the restaurant. Research also found that behavior is affected by pictures present in the immediate environment of an individual. Feinberg (1986) showed that in the presence of credit card cues, an individual was more likely to give money to a charitable cause than individuals who were not exposed to the credit card cues. In a similar way, McCall and Belmont (1996) found that a tip tray stamped with a credit card insignia led patrons in a restaurant to leave bigger tips for the waiters or waitresses than when no insignia was primed on the tip tray. Using donation boxes located in various stores, Perrine and Heather (2000) found that a fundraising request for the profit of an organization for animal welfare increased when color pictures of puppies were displayed on the donation boxes than when they were not.

From a theoretical perspective, the behavioral consequences of environmental cues are not explained by a priming effect as before. Scientists in marketing used the theoretical model of the stimulus-organism-response (SOR) to explain these behavioral effects (Mehrabian and Russell, 1974, Donovan and Rossiter, 1982). In the context where a verbal cue is a single word, this word is the stimulus (S), which causes a participant's evaluation (O) and causes some behavioral response (R). Two different evaluations are possible: positive or negative. Consequently, two behavioral responses are activated: approach or avoidance. Approach behaviors are seen as positive responses to someone or to an environment, such as a desire to interact with the person or to explore the environment, whereas negative responses include not wanting to interact with the person or to spend time exploring the environment. Therefore, it could be possible that a single word could have the property of activating a positive response, which in return leads the person, who is exposed to this word, to initiate contact with somebody or to be motivated to perform further behavior. The words “love” or “loving”, which are important for the human culture, could be interesting concepts to use for priming or as a cue in people's environment in order to influence their behaviors.

Love and helping behavior

Despite these advances in the field of priming, automatic social cognition or the role of environmental cues on people's behaviors, few attempts have been made to test the influence of an automatic activation of the cognition of love on social behavior. Lamy et al. (2008) in a natural setting, interviewed participants in the street and asked to recall a memory involving love, or in the control condition, a piece of music they loved. Then they met another confederate who asked them for money. Results showed that inducing the idea of love had a significant positive effect on compliance to the request for male passersby who were asked for help (giving some money “to take the bus”) by a female confederate but not for female passersby. These results were confirmed in a more recent study (Fischer-Lokou et al., 2009) where passersby were asked for directions. It was found that participants induced to recall a memory of love agreed more favorably to give directions and spent more time giving directions to the requester than participants in the control condition who were induced to retrieve a piece of music they loved. By using the same methodology for inducing the idea of love versus the idea of music, Lamy et al. (2009) observed participants' reactions toward a confederate who inadvertently lost a stack of compact discs when they were near each other. The results demonstrated that participants were more helpful when they were induced to recall a memory involving love.

In these studies, participants were asked to recall a memory of love that counted a lot for them, before their helpfulness was tested. However, research found that such behavioral effects can be reached in the absence of any direct reminiscence of love or conscious awareness of love scripts, by means of an automatic activation of love.

In a recent experiment of Lamy et al. (2010), male passersby were asked by a female confederate to indicate the direction of Saint-Valentine street (Saint-Martin street in the control group). Thirty meters ahead, the participant encountered another female confederate who asked for help, pretending a group of four disreputable-looking male confederates had taken her mobile telephone and refused to give it back. It was found that participants primed with the cognition of Valentine helped the female confederate get her mobile back more frequently than those primed with the cognition of Martin.

The objective of the present experiment was to study the impact of a new, but more basic method that induced love cognition and to evaluate its effect on a new dependent variable associated with pro-social behavior. We decided to test the effect of the single presence of the word “loving” on people's behavior. In the earlier mentioned studies, possible confound effects are associated with helping behavior. Indeed, in these experiments, it was found that the love-inducing variable led men to help more favorably a female confederate than a male confederate or that such a love inducing variable was not efficient with female participants. Thus perhaps, the love-inducing method led to further interest for romantic relations for men than for women, which, in return, led only men to help more favorably female confederates. Indeed, participants were asked to recall a memory of a previous love episode that occurred in their own lives. Such reminiscence could increase motivation for further romantic relations. In the same way, “Valentine”, which was used as the priming concept in Lamy et al. (2010) study, is clearly associated with romantic relationships given the fact that Saint-Valentine's Day is the annual commemoration held on 14 February celebrating love and affection between intimate companions. Both types of priming information are clearly associated with romantic and intimate relationships, but not with general affection toward people. Thus, it could be interesting to test the single effect of the word “Loving” without any other reminiscence on people's altruistic behavior.

In a recent paper by Bekkers and Wiepking (2010), more than 500 papers were examined to evaluate why people donate money to charitable organizations. They identified eight mechanisms as the most important forces that lead to charitable giving. One of them assumes that awareness of need is a prerequisite for philanthropy, and, generally, studies have found that perceived need is positively related to donation. Thus, a message for a fundraising solicitation that has associated helping with loving could increase the perception that the need to obtain money is greater. Indeed, “Love” is an important concept for human beings and is deeply related not only to romantic relationships but also to support, solidarity, and community relationships. Therefore, a request for help associated with this concept probably leads people exposed to such information to infer that the request for help is needed for the community.

In this study, volunteers during a fundraising campaign wore a tee-shirt on which the word “loving” was written. It was hypothesized that such an inscription would be associated with more compliance to the altruistic request than in the control condition.



The participants were 900 men and 900 women students who were solicited in various places of the Campus of Université de Bretagne-Sud located on the French West Atlantic Coast in Brittany.


For our confederates who solicited students on the campus, a white short-sleeved tee-shirt was worn that used several inscriptions on the front of the tee-shirt. One tee-shirt had no inscription, on the second one “Aimer = Aider” (“Loving = Helping) was written as is shown in Figure1. On a third tee shirt, the inscription “Donner = Aider” (Donating = Helping) appeared.

Figure 1.

Graphical representation of the Tee-shirt used in the “Loving = Helping” condition.


Fifteen young female confederates (M = 20.3 years, SD = 1.17) and 15 young male confederates (M = 20.4, SD = 1.39) were recruited in a social psychology course on a voluntary basis to act as confederates in the study during the week where the French Association against Muscular Dystrophy collects its funds (first week of December 2010). Of course, the association agreed to all of the volunteers, and the funds collected in this experiment were completely transferred to the funding of the association. Each confederate wore a tee-shirt, and the rest of his/her clothing appearance was nearly the same among female and male confederates: black trousers and a second black long sleeved tee-shirt under the white tee-shirt. Each confederate wore an official badge of the Association against Muscular Dystrophy displayed for each volunteer during this 2010 fundraising campaign.

The experiment was conducted on different campus locations in two towns in Brittany near the West Atlantic coast of France (Lorient and Vannes). The participants were solicited when entering different areas of the campus: libraries, cafeteria, building departments…

Each confederate was instructed to approach a student entering one of the locations listed earlier by saying. “Hello, I am a volunteer for the Association against Muscular Dystrophy. As you certainly know, this week we are collecting money to support medical research. Would you accept to give us something? This campaign is carried out all over the country so even a little donation will help us.” Then the confederate then estimated if the participant agreed or not to his/her request. In the case of a negative answer, the confederate thanked the participant. In the case of a positive answer, he/she waited for the participant to give him/her the money. The confederate estimated the amount and thanked the participant.

Each confederate was instructed to test 10 men and 10 women in each experimental condition. He/she was instructed to change tee-shirts after testing five men and five women. The order of the tee-shirts worn by the confederate was randomly determined. To prevent possible variation in the confederate behavior according to the experimental conditions, the confederates were not informed about the experimental hypothesis. Indeed, a pre-test was conducted in the street in order to train the confederate to become familiar with the request and to verify that they acted in the same way in the three experimental conditions.


The dependent variables used in this experiment were the number of participants who complied with the confederate's request (Table1) and the amount of money left by the participants who accepted to make a donation (Table2).

Table 1. Number of participants having complied with the request in the experimental conditions according to gender of the participant and of the confederate
Experimental conditionMale participantsFemale participantsTotal
No inscription   
Male confederates32.0% (48/150)36.0% (54/150)34.0%
Female confederates31.3% (47/150)38.7% (58/150)35.0%
“Loving = Helping”   
Male confederates42.7% (64/150)46.7% (70/150)44.7%
Female confederates44.0% (66/150)42.0% (63/150)43.0%
“Donating = Helping   
Male confederates26.0% (39/150)31.3% (47/150)28.7%
Female confederates23.3% (35/150)28.7% (43/150)26.0%
Table 2. Average amount of money (SD in parentheses) left by compliant participants in the experimental conditions according to gender of the participant and of the confederate
Experimental conditionMale participantsFemale participantsTotal
No inscription   
Male confederates1.94 (1.36)1.99 (1.46)1.97 (1.40)
Female confederates2.06 (0.95)2.09 (1.36)2.07 (1.19)
Total2.00 (1.17)2.04 (1.40)2.02 (1.30)
“Loving = Helping”   
Male confederates2.46 (2.17)2.66 (1.97)2.56 (2.06)
Female confederates2.68 (1.72)2.67 (1.70)2.67 (1.70)
Total2.57 (1.95)2.66 (1.84)2.62 (1.89)
“Donating = Helping   
Male confederates1.63 (0.98)1.67 (1.42)1.65 (1.24)
Female confederates1.87 (0.93)1.97 (0.89)1.92 (0.91)
Total1.76 (0.96)1.81 (1.20)1.79 (1.09)

With the number of participants who comply with the confederate request, a 3 (experimental conditions) × 2 (confederate gender) × 2 (participant gender) Loglinear analysis test was performed. A main effect of the experimental conditions was found (χ²(2, N = 1800) = 36.01, p < 0.001, r = 0.14). Additional comparisons revealed that the “no inscription” response condition was statistically different from both the “Loving = Helping” condition (34.5 vs. 43.8%, χ²(1, N = 1200) = 10.97, p = 0.001, r = 0.09) and the “Donating = Helping” condition (34.5 vs 27.3%, χ²(1, N = 1200) = 7.21, p = 0.007, r = 0.08). Comparison between the “Loving = Helping” condition and the “Donating = Helping” condition was significant (43.8 vs 27.3%, χ²(1, N = 1200) = 35.63, p < 0.001, r = 0.17). Neither the main effect of participant gender (χ²(1, N = 1800) = 3.16, p = 0.08, r = 0.04) nor the main effect of confederate gender (χ²(1, N = 1800) = 0.24, p = 0.62, r = 0.01) was significant. We found no significant interaction between participant gender and/or confederate gender and experiment conditions (p > .20).

With the amount of money left by compliant participants, a 3 (experimental conditions) × 2 (confederate gender) × 2 (participant gender) analysis of variance was performed. A main effect of the experimental condition was found (F(2, 634) = 17.50, p < 0.001, ηp² = 0.05). The post hoc test revealed that the “Loving-helping” condition was statistically different from the “no inscription” condition (2.62 vs 2.02, LSD test, p < 0.001) and the “Donating = Helping” condition (2.62 vs 1.79, LSD test, p < 0.001), whereas no statistical difference was found between the “Donating = Helping” condition and the “no inscription” condition (1.79 vs 2.02, LSD test, p = 0.14). Neither the main effect of participant gender (F(1, 634) = 0.31, p = 0.58, ηp² = 0.00) nor the main effect of confederate gender (F(1, 634) = 1.72, p = 0.19, ηp² = 0.00) was significant. We found no significant interaction between participant gender or confederate gender and experiment conditions.


The results found in this experiment carried out in a field setting confirmed our hypothesis. The word “Loving” was associated with greater altruistic behavior. Interestingly, a twofold effect was found in this condition: more participants have complied with the fund request, and compliant participants in this condition gave a higher amount of money to the confederate.

Such results confirm previous research that showed an influence of an automatic activation of the cognition of love on helping behavior (Fischer-Lokou et al., 2009; Lamy et al., 2008; 2009). However, in the latter studies, activation of such cognitions was done with the help of reminiscence of previous romantic events in the participants' lives. In a recent experiment of Lamy et al. (2010), the single name of a street was used to activate the same cognition. Passersby in the street were asked by a confederate to indicate the direction of Saint Valentine Street, which, in return, led them to help a second female confederate more easily than passersby who were previously asked to indicate the direction of Saint-Martin street. In these studies, it was found each time that love activation increased helping behavior only when male participants have the opportunity to offer some help to a female confederate. Such cognition had no effect when considering female participants and/or when help was solicited by a male confederate. In our experiment, we did not find this gender effect. In all the conditions and in the “Loving = Helping” condition, we found that both male and female participants comply in the same way to the helping request even when the request was addressed either by a male or by a female confederate. This no-gender effect could be explained by the fact that cognition of love used in this experiment is not associated with romance or romantic relationships, but more probably with support, solidarity, and community relationships. Thus, this could probably explain why cognition of love in our experiment led men and women to react in the same way to the request addressed by our male and female confederates. In the studies of Fischer-Lokou et al. (2009) and Lamy, Fischer-Lokou, and Guéguen (2008; 2009; 2010), cognition of love was probably not related with support, solidarity, and community relationship, but more clearly with romance, romantic relationships, courtship or, perhaps, with sex; and that is why only men reacted positively with such a priming effect when a woman asked them for help. Following cultural stereotypes, it was consistently found that men were more eager for courtship relations or sexual intercourse than women (Hatfield, 1983; Sadalla et al., 1987). Such a variation in male and female behavior could explain why, in these studies, male passersby helped more favorably female confederates than female passersby did. Cognition of love in these circumstances could have probably increased men's desire for romantic relationships, courtship or sex, which, in return, led them to help more favorably female confederates. Such cognition was probably not activated with female passersby when help was requested by a male confederate and/or was not appropriated when interacting with a female confederate. In our experiment, different cognitions were activated with the priming associated with love. Such cognitions can be activated in the same way and with the same level with men and women, and that is why the same rate of compliance was found according to participant and/or confederate gender. Primes associated with love are not the same in these studies as in our experiment, and that is why discrepant results were found in helping behavior according to participant gender.

One further explanation for our results is to interpret the effect of the word “loving” as a consequence of a positive mood activated, which in return led people to comply more favorably to the request. Many experimental studies found that positive mood activation, compared with a neutral mood, increases helping behavior. This positive mood can be activated in many different ways and, most of the time, does not require elaborate means: a smile, crossing a beautiful person in the street, pleasant ambient smells in the street… (Harris and Smith, 1975; Levin and Isen, 1975; Weyant, 1978; Bizman et al., 1980; Job, 1987; Baron, 1997; Forgas, 1997; 1998; Rind, 1997). Thus the word “loving” could have the property of activating such a positive mood, which in turn influences helping behavior. Of course, this explanation is interesting and needs to be evaluated in further experiments. However, a recent study (Fisher-Lokou et al., ) that replicated the studies of Lamy, Fischer-Lokou, and Guéguen (2008; 2009; 2010) described earlier found that priming people with the idea of love was not associated with variation in the mood score. Thus, it seems that mood is perhaps not the mediating effect between the word “loving” and the behavioral response of our participants in our experiment.

It was also found in this experiment that the message “donating = helping” was associated with a decrease in the number of participants who complied with the request. This effect could be explained by the reactance theory (Wicklund, 1974; Brehm and Brehm, 1981). This theory assumes that people feel free to do certain things. When these perceived freedoms are threatened, people are motivated to restore them. In this experiment, “donating = helping” was perhaps perceived by the participants as a form of order to comply, which probably activated some reactance and led them to comply less favorably to the request. In French, “donner” means “to donate” but the same verb is also used when you want to order someone to do something. Thus in our experiment, this verb was perhaps interpreted as an order to give money, which in turn led the solicited participants to perceive that his or her freedom of behavior was threatened or restricted. This negative feeling could explain why we found a significant decrease in compliance to the request compared with the control condition. It would be interesting in a future study to evaluate this feeling of freedom associated with the message printed on the confederate's tee-shirt.

Practical applicability for fundraising professionals

This experiment has some practical interest. Fundraising is a significant means used by health or humanitarian organizations to obtain money for their operations. It could be interesting for them to test various primes in their communication that activated cognition associated with support, solidarity, and community relationships. According to the results found in this experiment, it would be interesting to give the tee-shirt “Loving = Helping” to the thousands and thousands of volunteers who offer their time to the Association against Muscular Dystrophy during its annual fundraising campaign. An important increase in the total amount of money collected could be expected. However, high precaution should be taken in the choice of words. In this experiment, we found that “Loving” was associated with an increase in the rate and the amount of donation whereas “Donating” was associated with a decrease. Thus, it appears that it is important to test a word prime on actual behavior. Various pre-tests using an experimental design seem to be a good method to evaluate the efficiency of a prime. Despite such precautions, this study has some interest to philanthropy professionals. This experiment conducted in a field setting demonstrates that simple and low-cost intervention can increase charitable giving, and it is very easy to use the words “Loving = Helping” in various solicitation situations, such as solicitation addressed by mail, electronic mail, face-to-face interaction or printed on a badge worn by a solicitor. This technique is easy to replicate, easy to use, and so easy to adapt to various fundraising solicitations and pro-social requests.

This study also confirms that marketers who work with fundraising organizations could used pictures or physical objects that elicited altruism and donations. Several studies have found that pictures and images in fundraising advertising are associated with more donations (Eayrs and Ellis, 1990; Perrine and Heather, 2000; Burt and Strongman, 2003). These authors reported that images, which elicited the greatest commitment to give money, were those that prompted feelings of guilt, sympathy, and pity. However, further feelings could be activated. Given the results found in this experiment, it could be argued that there is a need to use and to evaluate the role of physical objects associated with feelings of compassion, support or solidarity.


Of course, this study had some methodological limitations. In this experiment, we examined only students' compliance to a fundraising request addressed by a confederate of the same age. Previous studies conducted with the “love”-inducing technique (Fischer-Lokou et al., 2009; Lamy et al., 2008; 2009) found that this technique still remains efficient regardless of the participants' age from 20 to 60. However, these studies did not examine fundraising requests. This study was conducted in France. Thus, the positive effect of the word “Loving” on donation could not be generalized to other countries. Indeed, previous studies that focused on activation of the cognition of love on helping behavior (Fischer-Lokou et al., 2009; Lamy et al., 2008; 2009) were carried out in France. Thus, replications in further countries are now necessary. The sample-sizes were moderate in this experiment, and replication with higher sample-size is required. However, a recent study conducted by Guéguen et al. (2011) found with a sample-size of nearly 15,000 individuals that the locution “Loving = Helping” is associated with a significant increase in a blood donation request. Thus, it seems that the “loving” evocating effect is robust and could be used with different pro-social requests. A second limitation is related with the fact that the three experimental conditions were performed by the confederates. Thus, it is possible that the loving condition yielded more donations because the confederates changed their manner of solicitation. Of course, this argument is allowable. However, the confederates were unaware of the hypothesis made and were trained in a pre-test to ensure that they acted in the same way in the three conditions. Indeed, we found that “Donating = Helping” decreases compliance to the request, but the confederates had no reason to think that this condition would be less efficient than the “Loving = Helping” condition or at least than the control condition. Finally, in the study of Guéguen et al. (2011), cited earlier, where donating blood was the request, a significant effect of the word “Loving” was found even though the experiment was conducted with solicitation addressed by mail. Thus, in our current experiment, it seems that variation could not be explained by variation in the confederates' behavior.