Communicating the cost of your altruism makes you cool—competitive altruism and sexual selection in a real-life charity situation

Maintaining a good reputation is crucial for humans. Altruism, e.g. charity, may serve as a costly signal that enhances reputation based on the real or communicated cost. Fundraising via charity running triggers competitive altruism when potential donors donate in reaction to the reputation increase of the fundraiser. Using real-life data of marathonists and half-marathonists (388 runners) and their 9281 donors, the present research focuses on how the communicated cost and goal of a charity run affected the potential donors. We analysed the introductory texts of the runners presented online according to the cost and the social benefit of the fundraising communicated by them. We have shown that emphasizing more the subjective cost of running and the social benefit of the goal, or writing a longer text, attracted more donors and, even though the average amount of donation per donor did not increase, still lead to a greater amount of donations collected overall by the fundraiser. It was also shown that a higher communicated subjective cost resulted in a higher ratio of opposite-sex donors, both in the case of male and female runners, suggesting that the communication of the cost of an altruistic act might be the object of sexual selection. A good reputation is crucial for humans, as a reputable person enjoys several benefits. One way to maintain a good reputation is to be altruistic, e.g. doing charity. A seemingly high cost and a socially accepted goal may result in a higher reputation. Using data from a charity running community we demonstrate that fundraisers who emphasize their subjective cost (how difficult to run), and emphasize the good goal of the charity, attract more donors, and even though the average amount of each donation does not increase, a higher number of donors results in a greater amount of donations collected overall. Talking about the difficulties of the charity run results in a higher ratio of opposite-sex donors. Our results may be helpful to plan more successful charity events or to make a human community more altruistic and cooperative in general.

One way to enhance reputation is to show signs of altruistic behaviour (Fehr and Fischbacher 2003;McAndrew and Perilloux 2012a, b). Here the definition of Feigin et al. (2014) is used: "human altruism as an intentional and voluntary act performed to benefit another person as the primary motivation and either without a conscious expectation of reward (altruistic approach) or with the conscious or unconscious expectation of reward (pseudo-altruistic approach)". Humans are sensitive to altruistic cues because recognizing potential altruistic partners has its benefits because altruism functions as a reliable signal of important traits (i.e. willingness to cooperate, the possession of abilities, resources, etc.) (Zahavi 1975;Gintis et al. 2001). It is well known that humans actively use their altruistic acts to maintain their reputation and communicate its cost and benefits to their peers (Birkás et al. 2006;Myers and Carpenter 2007;Bereczkei et al. 2010). Depending on which strategy leads to a higher reputation, humans may mask or emphasize the cost of their altruism. For example, in the case of an altruistic act whose cost differs to an extreme degree from the average costs incurred by others' altruistic acts, humans might communicate their actions and hide the cost (Raihani and Smith 2015;Mokos and Scheuring 2019).
Reputation can only be interpreted in comparison with the reputation of peers (Barclay 2015;Milinski 2016). Therefore, humans constantly monitor their social environment and react to others' actions to maintain their reputations. If the reputation of a peer is increasing, others typically react with the desire to restore their reputation relative to this new level. This leads to competitive altruism, in which people compete to be considered more altruistic than others (Roberts 1998;Barclay 2004Barclay , 2011Hardy and van Vugt 2006;Barclay and Willer 2007;Sylwester and Roberts 2010).

Gender differences in maintaining a reputation
Having a good reputation is the subject of sexual selection as well, playing an important role in inter-and intra-sexual competition (Phillips et al. 2008;Arnocky et al. 2017). A good reputation is a desirable trait in a mating partner. It is shown that both males and females prefer altruistic mating partners, and regardless of sex, more altruistic people report a higher number of sex partners than less altruistic ones (Farrelly et al. 2007;Phillips et al. 2008;Barclay 2010;Farrelly 2013;Moore et al. 2013;Oda et al. 2014;Arnocky et al. 2017). Similar to the situation in other species where females have a higher parental investment than males and are therefore more choosy about partners than males, human males experience a stronger sexual selection pressure and signal any traits more intensively to attract mates or/and to overcome same sex competitors compared to females (Bateman 1948;Buss and Schmitt 1993;Shackelford et al. 2005), whereas there is a stronger selection pressure on females to recognize altruistic cues in their potential mating partner (Phillips et al. 2008). Therefore, people behave in different ways when their potential reputation-enhancing altruistic act is observed by a potential mating partner compared to when it is observed by same-sex peers (Iredale et al. 2008). Males are more likely to be altruistic in the presence of females (Tognetti et al. 2012;Bhogal et al. 2016aBhogal et al. , b, 2017Farmer and Farrelly 2021), which could lead to competitive altruism in males, but this effect was reported to be less strong or absent in females (Roberts 1998;Iredale et al. 2008;Barclay 2011Barclay , 2013McAndrew and Perilloux 2012b;van Vugt and Iredale 2013;Raihani and Smith 2015;Arnocky et al. 2017). Even though females are more altruistic, selfless and empathetic than males in everyday situations (Hoffman 1977;Eisenberg and Lennon 1983;Eagly and Crowley 1986;Erdle et al. 1992;Eckel and Grossman 1998;Croson and Buchan 1999;Andreoni and Vesterlund 2001;Skoe et al. 2002;Cox and Deck 2006;Dufwenberg and Muren 2006;Eagly and Koenig 2006;Chaudhuri and Gangadharan 2007;Simmons and Emanuele 2007;Kamas et al. 2008;Schwieren and Sutter 2008) they do not seem to be more altruistic in the presence of males (Iredale et al. 2008). However, other studies have pointed out that young females' altruism may well depend on the presence of other young females (Tognetti et al. 2012). Further, females are more sensitive to damage to their own reputation than males (Benenson et al. 2013;Garbarini et al. 2014;Reynolds et al. 2018). These phenomena could be explained by the different same-sex friendship patterns of males and females. While males form large coalitions where friends help each other to establish a good reputation, females rather associate with one female friend at a time or in small cliques (Savin-Williams 1980;Benenson 1990;Benenson et al. 1997Benenson et al. , 2013Gabriel and Gardner 1999;Markovits et al. 2001;Seeley et al. 2003). These differences are present from early childhood (Savin-Williams 1980;Benenson 1990;Buhrmester and Prager 1995;Verkuyten et al. 1996;Benenson et al. 1997Benenson et al. , 2018Gabriel and Gardner 1999;Markovits et al. 2001;Seeley et al. 2003) and similar to coalition formation in the case of our closest relative, the chimpanzee (de Waal 1984;Langergraber et al. 2009;Surbeck et al. 2017;Benenson 2019). Overall men seem to be more sensitive to how women value their reputation and women seem to be more sensitive to the social reputation of same-gender peers (Kawasaki et al. 2016).

Charity and reputation
Charity is an altruistic act; it can thus be employed for reputation improving (Keser 2003;Barclay 2004Barclay , 2006Barclay , 2010Albert et al. 2007). The cost of charity may easily be measured in time or money. Non-monetary donations are viewed as more morally praiseworthy in WEIRD (western, educated, industrialized, rich and democratic) societies (Johnson and Park 2019), and people are more likely to perform a charitable action that could be deemed costly in terms of the time required, the pain or discomfort incurred, its difficulty or the fact that the action requires special skills (Smith and Bird 2000;Bereczkei et al. 2007;Olivola and Shafir 2013). It may therefore be assumed that non-monetary donations have a stronger reputation-increasing effect compared to donating money, and the communication of the action's cost is likely to affect the reputation change it causes. This hypothesis is supported by the fact that donors communicate the cost of the charitable action according to the reputation change it might cause. For example, when donating money, people tend to show the amount donated only if it is average compared to the standard, and mask the cost of the outstanding (both smaller and larger than average) donations (Peacey and Sanders 2014;Raihani and Smith 2015;Mokos and Scheuring 2019).
Special types of charity situations may lead to competitive altruism. One such situation is a charity running when one person performs a charitable act that is not a monetary donation (running) and asks their friends to donate money concerning this act. Friends feel the urge to donate not only because they were asked to do so and because they support the goal of fundraising, but also (subconsciously) to restore their relative reputation compared to the runner and the peers, so the donors compete with the charity runner and the other donors. The way the runner communicates the cost of their time-and energy-consuming charitable act affects their potential reputation increase, and therefore the reaction of their friends. It might be assumed that the more the charity runner emphasizes the cost of the act to them (e.g. talking about the time spent on training for the race, or how much they don't like running, or how performing such exercise causes them pain, etc.), the more their reputation increases by this act. With an increase in reputation, the peers of the runner feel a stronger urge to donate and to donate a higher amount. Besides this, the donations of peers also affect any single one of their donations: potential donors are more likely to donate if they know others have donated as well (Barclay 2011) and tend to make an average or just slightly above average donation compared to previous donors (Raihani and Smith 2015).
Not only the communicated cost, but naturally the charitable goal may modify the reputation of the altruist (Sargeant 1999;Barclay 2010). A 'good' goal does not seem selfish, but socially oriented. Contrary, when the motivation of the charitable act is unclear or perceived to be in bad faith, it might even reduce the reputation of the actor instead of increasing it. It could therefore be beneficial to donors to communicate the 'good' goal and the credibility of the charity to which they donate.

In this study
In summary, on the basis of direct and indirect experimental results, competitive altruism can act as a reputationmaintaining mechanism. However, the authors are unaware of any real-life data analysis examining the impact of communicating altruistic behaviour.
To address this gap in the literature, in this study a real-life competitive altruistic situation was investigated, via the study of how communicating the cost and the goal of an altruistic act determines the reputation of the altruist, and therefore affects the reaction of their peers. A charity running situation was investigated in which people enter running races and call upon their friends to sponsor them, the donations going to a therapeutic summer camp for seriously ill children.
On the basis of the points raised in the preceding discussion of competitive altruism, the following hypotheses were tested: 1) Runners who emphasize the cost of the run more gain a larger reputation increase, and therefore, (A) have more donors and/or (B) a higher average donation per donor, and thus (C) raise a greater amount of money overall. 2) Runners who emphasize the socially oriented goal of the charity run gain a larger reputation increase, and therefore, (A) have more donors and/or (B) higher average donation per donor, and (C) raise a greater amount overall. 3) (A) More opposite-sex donors make donations to fundraisers with higher subjective costs and/or (B) oppositesex donors donate a higher amount to such fundraisers.
As any effect of the communication of cost and goal could be a by-product of the length of the introductory text, and as runners who write longer introductory texts might be more committed to the fundraising an explorative hypothesis was added: 4) There is a relationship between the length of the introductory texts and (A) the number of donors, (B) the higher average donation per donor, and thus (C) the amount of money raised.

Data collection
Using web scraping technics in R (R Core Team 2021) and the rvest package (Wickham 2021), the public data on the runners and donors were collected from the runners' ED webpage (https:// elmen ykulo nitme ny. hu). The name, profile picture and introduction of the runners, the amount of the donation collected, the distance run and the number of donors and the name and sum donated by the donors (if the donor made it public) were collected. The web scraping was undertaken on 12 December 2020, resulting in the collection of data on donations made between 21 August 2012 and 8 December 2020. All donations are given in Hungarian Forint (HUF, in 2012 the exchange rate was 1 EUR = 290 HUF, in 2019 1 EUR = 360 HUF; the median wage in Hungary in 2019 was 367,800 Ft ≃ 1000 EUR).
To minimize observer bias, blinded methods were used when all behavioural data were recorded and/or analysed. There are numerous unknown factors, such as the fundraising strategy of the runner, the size of the reference group, the runner's personal connection with the goal, the age of the runner and their friends etc. which can modify the success of the fundraising. Even though it would have been desirable to correct for all of these effects, most of these data were not available, only the age of the runner could be extracted. The age of the runner defines the financial background of the runner (older runners have better pay), and therefore the financial opportunities of the friends of the runners (provided most of the friends are mostly the same age as the runner) and so the average amount of the donation. Knowing or estimating the age of the runners makes it possible to exclude this effect from the calculations. Although most of the runners (448) had not shared the information about their age in the introduction, their profile photo was used to estimate their age. To predict the ages of runners, all the images on their profiles were downloaded and face detection was employed, using the open-source DSFD face detector (Li et al. 2019). The facial images were then cropped along the bounding boxes given by the algorithm. This was followed by a manual validation, the aim being to obtain just the facial images of the runners, as several downloaded images had more people in them. A convolutional neural network was trained for the task. The model achieved an MSE score of 50.6 (measured in years) on a validation dataset, which means that the error of any one prediction was 7.1 years on average. This model was used for prediction on the previously cleaned dataset of facial images of the runners. For any person that had more than one facial image in the dataset, the predictions given by the model were averaged. The neural network training details are given in the Supplementary material.
The subjective cost of the running was measured by using the number and the ratio of sentences in the runners' introduction text that referred to any difficulty related to the run. The following were considered 'difficulties': (1) mentions of physical and mental difficulties, (2) this race is the first the fundraiser had ever entered, (3) the distance run was regarded as long by the fundraiser, (4) the fundraiser did not like/enjoy running or exercise, (5) the fundraiser would not have run had it not been for the ED.
Communication of the socially oriented goal was measured with the number and the ratio of positive words in the runners' introductions.
First, the texts were tokenized using magyarlanc 3.0 (Zsibrita et al. 2013), and then the meaningful words were scored according to the sentiment, using two Hungarian sentiment lexicons (Chen and Skiena 2014;Szabó 2014). If a word appeared as a positive word in any of these two lexicons, it received a score of 1. Emojis referring to positive emotions also received a score of 1. The sum of the scores was calculated for the introductions, and the number or the ratio of positive words was used. It was hypothesized that positive words mainly refer to the charitable goal, so runners who use positive words were writing about the therapeutic summer camp, emphasizing the usefulness and social benefits of the donation. This hypothesis was tested on a random sample of introduction texts, and introductions having a higher number of positive words did indeed prove to have a higher number of sentences about the summer camp and the children (see the Supplementary material). The number and the ratio of positive words could thus be used as a measure of how the fundraiser communicated the good goal of their fundraising act.
The number and the ratio of sentences referring to difficulties and positive words were used as alternative variables because it is unknown whether potential donors perceive the cost of running and communicating the socially oriented goal as additive values or rather the proportion of them within the text is estimated.

Descriptive statistics of the data
Data of 477 runners (amongst whom 245 were half-marathonists, 143 were marathonists and 89 run other distances between 2 and 160 km) and 11,089 donors were collected. Unless it is specified otherwise, only the data of marathonists and half-marathonists (388 runners, 212 females and 176 males; 9281 donors) are used in the calculation.

Statistical methods
A series of linear mixed effect analyses were performed using R (R Core Team 2021) and the lme4 package (Bates et al. 2015) to study hypotheses 1, 2 and 4. The donation collected, the average donation or the number of donors was entered as dependent variables. The length of the introductory text (either the number of words or the number of sentences, hypothesis 4) or the communication of subjective cost (either the number or the ratio of sentences referring to difficulty, hypothesis 1) or the communication of the socially oriented goal (either the number or the ratio of the positive words used, hypothesis 2) was set as the independent variable. The age of the runner was used to control the monetary bias of donors of fundraisers of different ages (without any interaction term), as it was hypothesized that friends of a younger fundraiser are also younger and have lower pay than the friends of older fundraisers, while younger fundraisers probably have more friends and the competition between them is stronger (Bhattacharya et al. 2016). The ratio of non-anonymous donors was used to control the size of different donations received by anonymous or non-anonymous donors, as non-anonymous donors tend to donate average donation while anonymous donors donate a below or above-average donation with a higher chance (Mokos and Scheuring 2019). As random effects, there were random intercepts for the distance run, as it was hypothesized that the distance run affects the size of the average donation and amount collected, and fundraisers who run a longer distance receive a higher average donation and overall collect more than fundraisers running a shorter distance. There were random intercepts for the gender of the fundraisers as it was hypothesized that male fundraisers receive higher donations compared to female fundraisers (see supplementary material).
To study the relationship between the ratio of oppositesex donors and the subjective cost (hypothesis 3A) two linear mixed effect analyses were performed. As fixed effects, the subjective cost (either the number or the ratio of the sentences emphasizing difficulty) and the age of the runner (without interaction term) were entered. As random effects, there were intercepts for the distance run and the gender of the runner, as well as by-subjective cost random slopes for the effect of distance run and the gender of the runner. To test whether the length of the introduction affects the ratio of opposite-sex donors, a similar linear mixed model was performed as mentioned above, using the text length measured by the number of sentences as the independent variable.
To study the relationship between the subjective cost of a runner and the donations of opposite-sex donors (hypothesis 3B), two linear mixed effect analyses were performed. As a dependent variable, the amounts donated by non-anonymous opposite-sex donors were set. As fixed effects, the subjective cost (either the number or the ratio of the sentences emphasizing difficulty) and the age of the runner (without interaction term) were entered. As random effects, there were intercepts for the distance run, the gender of the runner and the name of the runner, as well as by-subjective cost random slopes for the effect of distance run, the gender of the runner and the name of the runner. Only the data of opposite-sex donors who published both their names and the amounts were used. Donations higher or equal to 50,000 HUF were deleted as outliers, resulting in data of 1840 donors. To test whether the length of the introduction has an effect on donations given by opposite-sex donors, a similar linear mixed model was performed as mentioned above, using the text length measured by the number of sentences as the independent variable.
In the linear mixed effect analysis, visual inspection of residual plots did not reveal any obvious deviations from homoscedasticity or normality. P values were obtained using likelihood ratio tests of the full model with the effect in question against the model without the effect in question.
To compare the donation and the subjective cost of marathonists and half-marathonists, Mann-Whitney U tests were conducted.
The relationship between the age of the fundraiser and the subjective cost (measured in the number or the ratio of sentences emphasizing difficulties) was studied using Spearman correlations.
The relationship between the age of the fundraiser and the average size of donations was studied using Spearman correlations.
The relationship between the subjective cost and the ratio of the non-anonymous donors was studied using Spearman correlations.

Runners who emphasize the cost of the run more have more donors
Before studying the hypothesis, the question of whether there is a relationship between the subjective feeling of the cost of running and the objective cost of it, namely, the distance run, was tested. This was done by comparing the money-collecting effectiveness of marathonist to half-marathonist. Marathonists collected significantly larger donations (mean = 145,446 HUF, median = 155,000 HUF) than half-marathonists (mean = 123,955 HUF, median = 118,000 HUF) (U = 150,544, p < 0.001). Further, marathonists emphasized the cost of their run significantly more, writing a higher number and a higher ratio of sentences about the cost (Table 1), showing that the subjective cost is consistent with the objective cost (the distance run).
There was a positive relationship between the subjective cost and the number of donors regardless of whether the subjective cost was measured by the number or the ratio of sentences emphasizing the difficulties of the fundraiser (Table 2, Fig. 1).
The subjective cost seemed not to influence the average donation, as neither the number nor the ratio of the sentences emphasizing difficulties showed a relationship with the average donation, demonstrating that the reason for the larger donations collected by fundraisers with higher subjective costs is not the higher average donation but the higher number of donors attracted (Fig. 1).
The subjective cost might influence the amount collected (Table 2). There was an almost significant positive relationship between the number of sentences emphasizing difficulties and the amount collected, while the ratio of sentences did not influence the amount collected. The shown trend could be a side effect caused by the text length, as the number of sentences in the introduction text shows a positive relationship with the donation collected (Table 4).

Runners who emphasize the socially oriented goal of the charity run might have more donors and might collect a greater amount overall
Emphasizing the socially oriented goal more might help to collect a higher amount by attracting more potential donors (Table 3, Fig. 2). The number of positive words showed a positive relationship with the number of donors. However, the ratio of positive words did not influence the number of donors. Table 1 The mean and standard deviation of the number and ratio of the sentences emphasizing the difficulties, the introduction's length written by the fundraisers and the comparisons of the marathonists and half-marathonists using the Mann-Whitney U test Emphasizing the socially oriented goal did not show any relationship with the average donation. Neither the number of positive words nor the ratio of the positive words had any influence on the average donation, demonstrating that the reason for the fact that fundraisers using more positive words collect more overall donations is not the higher average donation but the higher number of attracted donors.
Note, that the text length measured in words positively influenced both the number of donors and the donations Fig. 1 The relationship between the subjective cost of the fundraises and the donations collected (a, b) and the number of donors (c, d), (Hypothesis 1). The subjective cost was measured by the number (a, c) or the ratio (b, d) of the sentences emphasizing difficulties written by the fundraisers in their introductions. HUF Hungarian Forint a b c d   collected, indicating that the relationship between the communicated goal and the number of donors or the donations collected might be a side effect of the length of the text.

Runners who write a longer introductory text have more donors but do not have a higher average donation per donor and still collect a greater amount of money overall
There was a positive relationship between the length of the introductory text and the number of donors regardless of whether the length was measured by the number of words or the number of sentences (Table 4, Fig. 3). There was a positive relationship between the length of the introductory text and the money collected overall regardless of whether the length was measured by the number of words or the number of sentences. No relationship was found between the length of the introductory text and the average donation.

More opposite-sex donors donate to fundraisers with higher subjective costs, but the donated amount of the opposite-sex donors does not vary with the subjective cost of the fundraiser
The subjective cost showed a positive relationship with the ratio of opposite-sex donors ( Table 5). The number of sentences emphasizing difficulties positively affected the ratio of opposite-sex donors, while the ratio of the sentences emphasizing difficulties has an almost significant positive effect on the ratio of opposite-sex donors, so the gender ratio of the donors of a fundraiser with a higher subjective cost is more biassed toward the opposite sex (Fig. 4). Note that in this calculation only non-anonymous donors are included, as there was no information about the gender of anonymous donors.
The subjective cost did not show a relationship with the amounts donated by non-anonymous opposite-sex donors (Table 5).
Note, that there was no relationship between the text length measured by the number of sentences and the ratio of opposite-sex donors or the amount donated by non-anonymous opposite-sex donors.

Discussion
It has been shown that more friends donate to fundraisers who write longer introductory texts. The sizes of the donations remain the same, but the higher number of donors still leads to more money collected overall.
Fundraisers who wrote a longer text might be seen as more enthusiastic or dedicated to their fundraising goal, which could convince more friends to donate. These fundraisers might allocate more energy to advertise their fundraising act, for example, talk more about their activity, share it on social media or organize donation collection Table 5 The relationship between the ratio of opposite-sex donors and the subjective cost, and between the donation size given by opposite-sex donors and the subjective cost studied with linear mixed models (Hypotheses 3)  events. Such activities help to reach more people and could enhance the reputation of the fundraisers which also could lead to a higher number of donors. It is also possible that fundraisers who write a longer text are more extroverted and maintain more active friendships, having a higher number of potential donors. Further, it has been shown that fundraisers communicating higher costs do not receive a higher average donation, but since more donors donate to them, they still tend to collect more donations than fundraisers who are not advertising their subjective cost.
This result confirms that altruism (and so charity) is a sign of cooperativeness and so could trigger an altruistic response from the peers/friends of the fundraiser. Donating in response to the request of a friend also serves as friendship confirmation. The 'louder' the request, the greater the number of friends who respond. The donor-friend also benefits from the donation, as being seen to help also increases reputation (Roberts et al. 2021).
It was hypothesized that in response to a costlier fundraising act donors react with a higher donation for the gain in reputation of the fundraiser. Data did not support this hypothesis, as no relationship was observed between the subjective cost and the average donation. The amount donated is influenced by several factors: the financial situation of the donor, previous donations (people tend to donate an amount in line with the average donation (Martin and Randal 2008;Smith et al. 2015)), the personal relationship between the fundraiser and the donor, social expectations and so on. The lack of relationship between subjective cost and average donation could be caused by several opposing effects that mask the hypothesized correlation. For example, the correlation could be suppressed by the bystander effect: in a situation where help is required, and which is witnessed by many people, responsibility is divided amongst those present, and people are less motivated to help (Fischer et al. 2011); this includes donating less in a charitable situation if other donors are present (Garcia et al. 2002). In the present study, potential donors are aware of the number of previous donors and the amounts, as the previous donations (including anonymous donations) are visible to any potential donor on the webpage. This information leads probably to a norm-following habit, which suppresses the deviation from the previous donations.
We could not exclude that the relationship between the subjective cost and the number of donors is the side effect caused by the relationship between the length of the introduction and the subjective cost as fundraisers writing a longer introduction have more donors and overall collect a higher amount of donation. However, since when the subjective cost is measured by the ratio of the sentences emphasizing difficulties that take the length of the text out of account, the positive correlation remains significant; therefore, the relationship between the number of donors and the subjective cost still remains valid.
Similarly, the positive effect of the subjective cost on the number of donors and the donation collected might be the side effect of the relationship between the subjective cost and the age of the fundraiser. It is reasonable to assume that running is costlier for older fundraisers which is communicated by them, and at the same time, they have friends with better financial backgrounds than younger fundraisers. So, it could be assumed that age positively correlates with the communicated cost and the amount of collected money at the same time. And while there was an almost significant weak positive relationship between the age of the fundraisers and the average size of the donations (ρ = 0.09, p = 0.08) indicating older fundraisers received higher donations, rather surprisingly we found that younger fundraisers tend to communicate a higher subjective cost than older ones, regardless of whether the subjective cost is measured by the number of sentences (ρ = − 0.118, p = 0.016) or the ratio of the sentences emphasizing difficulties (ρ = − 0.114, p = 0.019).
As pointed out, contrary to the hypothesis adopted in this research, fundraisers using more positive words in their introduction, words which emphasize the socially accepted goal of the fundraising, do not receive a higher average donation. However, the results support the hypothesis that as more donors donate to them, they still collect an overall higher amount than fundraisers who do not emphasize the good goal.
People prefer to donate to charities that they regard as meaningful (Sargeant 1999), and helping those in need, especially helping children, is one of the charitable goals with the widest degree of social acceptance (Élménykülönítmény has collected 287,336,932 HUF, approx. 806,100 EUR since it opened in 2011, an outstanding result in Hungary, demonstrating its goal is socially accepted). What is more, the fundraisers' own choice of words concerning the goodness and usefulness of the charitable goal can affect potential donors (Smith and Schwarz 2012). Fundraisers also bring the recipients of the donation (seriously ill children) closer to the potential donors, making them more identifiable, which triggers warm empathy and results in giving donations (Small et al. 2007).
People also prefer to donate to charities that they think are trustworthy. Talking about the therapeutic summer camp and how it helps the children signals the reliability of the foundation in two ways: (1) the potential donors learn how their donation would be used; (2) by investing a large effort into supporting the Bátor Tábor, the fundraiser shows the trustworthiness of the foundation, regardless of its aims (donors might donate to any foundation if it is important to a friend).
We have found that the significant effect of communication of socially oriented goals on the number of donors only appears when the number of positive words is used but not when the ratio of the words is used as an explanatory variable. The reason for this difference could be either that potential donors perceive the communication of the socially oriented goal as an additive variable, reacting to the absolute number of words or it could be again a side effect caused by the length of the introductory text (that shows a positive relationship with the number of donors, Table 4). As it is mentioned above, we could not exclude that fundraisers who write a longer introduction text might invest more heavily in fundraising (for example, advertising their fundraising on social media, or talking more about their running to their friends, etc.), or might be simply more communicative, and therefore they reach more people and collecting more money. A further study, applying a well-designed questionnaire for the fundraisers can reveal the connections between these effects.
A charitable run is a complex costly signal that signals altruism, physical well-being, a caring attitude and social sensitivity, all of which can be attractive traits both for females and males. We studied here how the communicated subjective cost as an altruistic signal affects opposite-and same-sex donors. Earlier it was shown in an online donating situation that physically more attractive fundraisers receive more and larger donations compared to less attractive fundraisers. Further, male donors compete with each other by responding larger amount if a male donates a large amount to more attractive female fundraisers previously (Raihani and Smith 2015). We have shown here that fundraisers communicating a higher cost in their costly signal have a higher ratio of opposite-sex donors. This result is consistent with the general view that charity runners may be viewed as attractive mating partners, which leads to a higher ratio of opposite-sex donors if charity run is communicated to be more costly. Note that unlike the results of hypotheses 1 and 2, the ratio of the opposite-sex donors does not depend on the length of the introduction text; therefore, it could be considered a more solid result.
The positive relationship between the ratio of oppositesex donors and the subjective cost may emerge roughly in four different ways: as the fundraisers emphasize their difficulties more (1) the number of opposite-sex donors increases and the number of same-sex donors remains constant; (2) the number of opposite-sex donors remains constant and the number of same-sex donors decreases; (3) the number of opposite-sex donors increases while the number of same-sex donors decreases; (4) the absolute number of opposite-sex donors remains constant but more opposite-sex donors decide to publish their names instead of donating anonymously. Unfortunately, the dataset used in this study is not suitable to test which scenario is happening, but further, with a more elaborated data, we can explore it in the future.
A positive relationship between the subjective cost and the ratio of opposite-sex donors could be created by an increasing number of opposite-sex donors choosing to donate non-anonymously. As no correlation could be demonstrated between the ratio of non-anonymous donors and the subjective cost (measured by the number of sentences: p = 0.14, rho = 0.08, measured by the ratio of the sentences rho = 0.077, p = 0.119), this hypothesis should be at least partly rejected.
The reputation of the fundraisers depends not only on the fundraising activities but also on the success of the fundraising. It could be assumed that the more donations the fundraiser collects, the greater the reputation increase they receive. Potential same-sex donors might feel their reputation threatened by the fundraisers and compensate not by donating, and therefore increasing their own reputation, but by refusing to donate to decrease the reputation of the fundraiser. This latter strategy would increase the relative reputation of the (potential) donor as well. Unfortunately, the dataset used in this study is not suitable to test this hypothesis.
To conclude, based on real-life data, the present research shows that fundraising functions as a costly signal and how fundraisers communicate the cost and goal affects the potential donors. Communicating a higher subjective cost attracts more potential donors, and even though the individual donations remain invariable, a higher number of donors results in a greater amount donated overall. It has also been demonstrated that fundraisers who communicate a higher subjective cost attract a higher ratio of oppositesex donors.

Declarations
Ethics approval This research did not contain any studies involving animal or human participants, nor did it take place in any private or protected areas. The data used in this study is publicly available on the webpage of the studied charity running community (elmenykulonitmeny.hu).

Consent to participate
As the data used in this study is publicly available on the webpage of the studied charity running community (elmenykulonitmeny.hu), no patient consent statement was needed.

Conflict of interest
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.