Über Arne bin ich auch diese interessante Studie gestoßen:
We examine gender biases in the attribution of leaders’ outcomes to their choices versus luck. Leaders make unobservable investment choices that affect the payoffs of group members. High investment is costly to the leader but increases the probability of an outcome with a high payoff. We observe gender biases in the attribution of low outcomes. Low outcomes of male (female) leaders are attributed more to their selfish decisions (bad luck). These biases are driven by male evaluators. We find no gender differences in the attribution of high outcomes.
Quelle: Do women receive less blame than men? Attribution of outcomes in a prosocial setting
Also in der Übersetzung:
Wir untersuchen geschlechtsspezifische Verzerrungen bei der Zuschreibung der Ergebnisse von Führungskräften auf ihre Entscheidungen und nicht auf Glück. Anführer treffen unbeobachtbare Investitionsentscheidungen, die die Auszahlungen der Gruppenmitglieder beeinflussen. Hohe Investitionen sind für die Führungskraft kostspielig, erhöhen aber die Wahrscheinlichkeit eines Ergebnisses mit hoher Auszahlung. Wir beobachten geschlechtsspezifische Verzerrungen bei der Zurechnung von schlechten Ergebnissen. Niedrige Ergebnisse von männlichen (weiblichen) Anführern werden eher auf ihre egoistischen Entscheidungen (Pech) zurückgeführt. Diese Verzerrungen werden von männlichen Bewertern verursacht. Wir finden keine geschlechtsspezifischen Unterschiede bei der Zuschreibung von hohen Ergebnissen.
Also: wenn etwas schief geht, dann ist es bei Frauen eher Pech, bei Männern eher ihre Schuld, da sie egoistische Entscheidungen getroffen haben. Bei guten Ergebnissen hingegen gibt es keine unterschiedlichen Zuschreibungen.
Aus der Studie:

Fig. 1. Evaluators’ posterior beliefs that the leader has chosen high investment against Bayesian posteriors, by the leader’s outcomes. (a) Low outcomes, (b) High outcomes.
Aus der Studie:
Fig. 1 shows evaluators’ posterior beliefs given a low outcome (panel a) and a high outcome (panel b) against the Bayesian posteriors beliefs as derived using evaluators’ prior beliefs. In each panel, we present the graphs separately for male leaders (gray solid line) and female leaders (black dotted line). The dashed 45° line represents the situation where evaluators’ posteriors fully coincide with the Bayesian benchmark. The graphs are plotted based on the estimates reported in Table B1 of the Online Appendix (see, e.g., Eil and Rao 2011 for a similar estimation approach).
The figure reveals several key insights. First, evaluators generally do not conform to the Bayesian benchmark, as evidenced by departures from the 45-degree line in both panels. Second, panel (a) reveals that while evaluators’ posterior beliefs for low outcomes are lower than those predicted by Bayes’ rule for both male and female leaders, their posterior beliefs for male leaders are much lower than those for female leaders (p-value = 0.067 in column 4 of Table B1). Third, panel (b) reveals that evaluators’ belief-updating patterns in response to high outcomes are almost identical between male and female leaders. We investigate these findings further by estimating Eq. (3), which allows us to disentangle whether departures from the Bayesian benchmark are driven by evaluators’ treatment of their prior beliefs and/or their treatment of the leader’s outcomes.
Und aus den Ergebnissen:
In environments where the outcomes of leaders are determined by a combination of luck and unobservable actions, are outcomes of male and female leaders attributed differently by their evaluators? We answer this question in an environment where the actions taken by leaders affect both their own welfare and those of the group members (evaluators). Such environments are pervasive in many real-world settings both in the public and private domain. Gender biases in evaluations may emerge in these situations if, for example, women are expected to behave in a more prosocial manner.
We find an asymmetry between the evaluation of high outcomes and low outcomes. High outcomes of male and female leaders are not treated differently, suggesting that men and women are deemed to be equally altruistic (but less so than what a Bayesian would believe) after observing a high outcome. However, while the low outcomes of male leaders are attributed more to their selfish decisions, those of female leaders are attributed more to bad luck. Hence, in the case of failure, men are assigned more blame than women for being selfish. This is despite the fact that there are no gender differences in the evaluators’ prior beliefs about male and female leaders’ prosocial preferences. That is, evaluators start from a gender-neutral position, but they update their beliefs differently based on the gender of the leader when they observe a low outcome.
We find that this gender bias in the evaluation of low outcomes is driven by male evaluators and potentially, by evaluators who are prosocial. One interpretation of these results is that male evaluators may see the need to treat female leaders more favorably, thus giving them a greater benefit of the doubt in the face of failure. Interpreted in this way, one possible explanation for our findings is benevolent sexism (Glick and Fiske, 1996). Unlike hostile sexism, benevolent sexism tends to lead to behaviors toward women that are often characterized as prosocial.22 However, such biases in the evaluation process may still lead to adverse outcomes for women. For example, gender biases in evaluations that favor women may hinder the development of their careers and increase the possibility of backlash against female leaders in the long run.23 In general, gender biases are important to understand because they may lead to distortions in the incentives provided to all decision makers in positions of power (male and female) and may harm the future actions taken by them. Future research can shed light on the distortionary impact that such gendered evaluations can have on decision making by leaders and the consequent labor market outcomes.
Interessanterweise liegt es also an den Männern, dass die Bewertung abweicht. Frauen werden von Männern insofern großzügiger behandelt und es wird ihnen eher Pech zugestanden.
Das wird hier dem „benevolent Sexism“ zugeordnet, mir würde es passender erscheinen darauf abzustellen, dass Männer sich im Bezug auf Männer eher in intrasexueller Konkurrenz befinden und sich netter gegen Frauen verhalten wahrscheinlich eher evolutionäre Vorteile gebracht hat.
Passt insofern aber gut zu „Weibliche Unterverantwortlichkeit (Female Hypoagency) und männliche Hyperverantwortlichkeit (Male Hyperagency)“
Arne schreibt: „Männliche Chefs werden Opfer von „wohlwollendem“ Sexismus“. Aber tatsächlich sind natürlich die Frauen hier die Opfer des wohlwollenden Sexismus, denn die Ansicht, dass sie nur Pech hatten mag positiv klingen, letztendlich nimmt man sie aber damit nicht im gleicher Weise ernst wie die Männer und das ist natürlich schlecht.