Procedure

All participants provided informed consent before study inclusion, and they were instructed they could quit the procedure at any time. Most of the informants were illiterate, thus the consent was oral. The study complied with the Declaration of Helsinki on Biomedical Research Involving Human Subjects. Data were collected in the form of interviews in Ki-Swahili (Hadza) and the Yali language (Yali). The local assistants read all questions aloud in one-to-one dialogs, and further explanations were provided if necessary. More details are provided below. No other group members were allowed to stay nearby, in this way we tried to avoid at least some biases associated with social desirability.

The second and fourth digits of participants were measured with a digital Vernier caliper with.01 mm accuracy—we took measures of the second and fourth hand digit from a mid-point on the ventral crease proximal to the palm to the tip of the finger (Manning et al., 1998). Participants who reported injuries or deformities of at least one of these digits were excluded from the statistical analysis. Each measurement was collected twice from each participant. The means of the first and the second measurement of right and left 2D and 4D, as well as the right and left 2D:4D ratios were calculated following the procedure described by Manning et al. (1998) and Manning (2008). Intra-class correlation coefficients (ICCs) were calculated to assess the reliability of right and left hand 2D:4D between the two measurements. The ICCs were 0.93 and 0.96 for the right 2D:4D and left 2D:4D for the Hadza sample, and 0.86 and 0.97 for the Yali sample, respectively.

We measured maximum voluntary contraction (MVC) of the grip flexors, commonly known as hand grip strength using the Harpenden spring dynamometer (Balogun et al., 1991). Each participant was instructed to take a comfortable position and squeeze the dynamometer with their dominant hand as hard as they could, with a dominant hand lifted. We took measures three times. ICCs between measurements were high for both hands (0.98 and.99 for the Hadza and 0.82 and 0.87 for the Yali, for right and left hand, respectively). To analyze hand grip strength, we used the average of all three grips.

To assess helping behavior among participants, we asked them to estimate their involvement in helping others using four statements: (1) “I struggle to help my relatives if they are in need” (kin altruism); (2) “I can’t refuse if people ask for help” (helping in-group members); (3) “I help my friends if they have problems” (helping friends); (4) “I help those who helped me in the past” (direct reciprocity). Possible answers were rated on a Likert scale, from 1 (“never”) to 5 (“always, or regularly”). Initially the statements looked as follows: (1) “I help my relatives”; (2) “I help other people (in-group members)”; (3) “I help my friends”; (4) “I help those who helped me in the past.” Each of these statements addressed a certain type of altruistic behavior close to the specifications presented by Nowak (2006). We decided to present the questions in current format, because in a small pilot study obtained a lot of comments from our respondents, asking to specify in which case and if a potential benefiter has been asking for help or not. Before the main study, we tested the level of understanding of these statements among a group of Hadza and Yali people to exclude misunderstandings. In most cases, these questions were accompanied by further explanations and examples suitable to each society, e.g., helping with a child (both societies: caring for a child when mother is busy; bringing food or water for a mother with a new born baby); helping with carring meat after hunting (men in both societies); helping with cooking (both societies); helping with house construction and cleaning a new field (Yali); protecting in disputes (both men and women, in both societies); protecting against strangers (Yali men). Sharing on demand (under pressure) is something basic for Hadza society. Direct address with a request for help also seems common.

To verify whether participants comprehended the questions, we asked them to provide examples of particular activities and memories of situations when they helped. A short list of questions about helping behavior was used to obtain the most precise answers possible while maintaining respondent attention and interest in the interview. The Cronbach’s alpha values for total helping were reliable (in the Hadza: Cronbach’s α = 0.71, n = 223; in the Yali: Cronbach’s α = 0.79, n = 83).

Apart from society type, we also recorded information about participants’ sex and age. We decided to divide the sample into age groups, as many respondents did not know their exact age; thus, they were approximated in accordance with the informants’ personal information about their life events and their groupmates’ comments. Consequently, we divided the participants into six age groups, from 18–19 years (group 1) to 60 years and older (group 6).