Together the studies suggest that men have more perceived power in the public domain, however, this domain has a lower preference weighting than the private domain where women have more power than men

Weighting power by preference eliminates gender differences. Sverker Sikström ,Laura Mai Stoinski,Kristina Karlsson,Lotta Stille,Johan Willander. November 5, 2020. https://doi.org/10.1371/journal.pone.0234961

Abstract: Power can be applied in different domains (e.g., politics, work, romantic relationships, family etc.), however, we do not always reflect on which domains we have power in and how important power in these domains is. A dominant idea is that men have more power than women. This notion may be biased because the concept of power is associated with public life. We introduce the concept of preference-weighted power (PWP), a measure of power that includes different domains in life, weighted by the domains’ subjective importance. Two studies investigated power from this perspective. In Study 1, participants generated words related to power, which were quantified/categorized by latent semantic analysis to develop a semantic measure of the power construct. In Study 2, we computed a PWP index by weighting the participants' self-rated power in different power domains with the importance of having power in that domain. Together the studies suggest that men have more perceived power in the public domain, however, this domain has a lower preference weighting than the private domain where women have more power than men. Finally, when preferences for power in different domains were considered, no gender differences were observed. These results emphasize gender difference in different domains and may change how we perceive men’s and women’s power in our society.

General discussion

As discussed in the introduction, people’s concept of power is often biased towards power in public domains (e.g, [8]), where men earn higher salaries than women and fill the majority of powerful positions in work and politics [13–15, 36]. This suggests that men hold more public power compared to women and supports the idea that men have more power in general. The goal of the present study was to broaden the current view on power and gender, by introducing the concept of preference-weighted power. Looking at people's preferences suggests that private domains, for example family, friends and romantic relationships, are viewed as being more essential in life. However, these domains have often been neglected from a power perspective. Therefore, we applied the concept of preference-weighted power (PWP). The findings from these measures converge on the idea that women have more perceived power, in social relationships which as shown in Study 1, also were the domains that were most commonly generated as the most important are in life. However, when in Study 2, weighing the relative importance of powers across domains, there were no gender differences in domain-independent power, as measured by PWP.

General power, public and male power. In Study 1 a statistical semantic approach to measure and define people's concepts of power was applied. In particular we investigated how the participants associated men and women with the concepts of power, work power and social relationship power. We hypothesized that the participants’ view of power would be biased towards their concept of work power as well as their concept of male power. Consistent with our hypothesis, the words describing the concept of work power were significantly semantically closer to the words describing power in general, compared to the words describing social relationship power. Thus, supports the notion that people’s idea of power is actually biased toward the public domain (e.g., [7]). Furthermore, the results showed a stronger association between power in general and participants' idea of male power, compared to their view of female power. That people perceive power as stereotypically masculine could imply that the participants associate men with more power than women. The data from the present study thereby confirms a socially conditioned power concept. However, the current paper also provides an opportunity to recondition these social constructs by providing empirical data suggesting that women have power in domains perceived as important independently of gender, e.g., family, friends. In addition, we explored the connection of male and female power to work versus social relationship power. The results suggest that the words created by the participants to describe their view on work power were closer related to their concept of male power compared to female power, implying that people view work power as a stereotypically masculine domain. The opposite was true for social relationship power. Here, the words generated to describe social relationship power were closest related to the participants’ view of female power, suggesting that social relationship power is perceived as a predominantly female domain of power. The results are interesting as they reflected typical gender-role stereotypes, picturing women to anticipate communal goals such as family and other social relationships, while men are expected to possess agentic traits and to pursue goals related to career and work. People often associate power with success in work related domains as well as with people who possess agentic traits. These findings are also consistent with previous literature showing that men are stereotypically perceived as more agentic, as such for example competitive, achievement oriented and assertive (e.g., [37]) and associated to executive positions at work and in politics, where people might perceive men as more powerful than women (e.g., [16, 17, 37, 38]). However, these premises might require a revision of how we look at power, as the common held view on power neglects other possible domains where power can be exerted. Thus, the result emphasizes the need to revise the concept of power in the context of gender.

The most important area in life. In study 1, we also investigated the relationship between domains where men and women have power and the domains that people think are important in life. Consistent with our hypothesis, the words female participants used to describe their domains of power were semantically closer to words describing what both genders considered important in life, compared to the words generated by men to describe their domains of power. This suggests that women have more power in domains that are important to people. As the participants primarily described private domains, when asked what is important in life, the findings are well aligned with previous literature [20, 21]. Previous research highlights the subjective as well as objective importance of private domains in life [21]. Social relationships, such as family, partner and friends are more important for people as well as their well-being and health compared to material needs, and these findings seem to be even stronger for men [10, 12, 39]. Even though in the last decades men have become increasingly involved in housework and childcare, alongside to women's larger economic and political representation and contribution [36, 40, 41], women still show a larger involvement in private domains of life (in USA; [42]) where they have power over the majority of decisions associated with family, healthcare, food, daily purchases etc. [43–45].

Preference weighted female power and power in social relationships. As addressed in the introduction, power can be defined in many different ways, and different power concepts stress different decision-making domains, depending on the specific interest of the investigation. Further, as indicated in Study 1, people often relate power with work/public power, without reflecting on other domains of power and their personal preference for these domains (e.g., [8]). Moreover, although previous research has highlighted the subjective and objective significance of private domains in life, very little prior research has been conducted in which personal preferences for these domains have been investigated (e.g., [21]). Therefore, we introduced the concept of preference weighted power in Study 2, and measured it by weighting the participants’ self-rated power with their preferences for specific domains. Consistent with our hypothesis, women had relatively more perceived power in social relationships compared to the work domains. When testing gender differences, women showed significantly higher perceived private power than men. Again, the results reflect the notion of women's larger power in private domains. This is consistent with the concept of women’s larger dyadic power, defined as the power to influence others in close relationships (e.g., [8, 35]). This study did not demonstrate men’s larger PWP power compared to women. However, men had relatively more perceived power in public compared to private domains, which is consistent with the results of Study 1, suggesting that work power is perceived as a male domain of power.

An additional focus of the study was to investigate gender differences in the domain independent PWP scores, including both private and public domains. Previous literature has often neglected private domains of power and hence attributed more power to men [16, 17, 46–48]. In contrast, no significant gender difference became apparent when using PWP as a power measure in this study and the results were consistent, regardless of the applied approach to compute the preference weights. Thus, by extending the concept of power to private domains (e.g, social relationships) and by taking preferences for specific power domains into consideration, we were able to demonstrate the possibility that men and women have similar amounts of power. This shows that, despite the still prevalent inequalities in public power, women are not powerless.

Power was investigated using common self-report measures as well as semantics. Here, including statistical semantics improved the study’s sensitivity beyond mere keyword counting. Further, we used factor analyses to assess the underlying structure of the PWP scores in Study 2. Serving as a manipulation check, the analyses supported the notion that power can be divided into public versus private power, thus enabling us to study the powers separately. In addition, applying different approaches to compute the PWP scores allowed us to control possible artefacts (e.g. response biases).

Methodological considerations. The results of the present study suggest that, despite gender differences in domain specific kinds of power, men and women do not differ in preference weighted power. However, the current project of course also has a number of limitations and shortcomings. First, as the majority of participants resided in the US, generalising the results of the study to other cultural contexts is difficult. For example, traditional gender role stereotypes, reflected by the concepts of male and female power measured in Study 1, could be even stronger pronounced in less gender-equal societies, while being less prominent in high-parity nations that are closer to achieving gender equality than the US (e.g., in the Scandinavian countries, [49]). The traditional division of labor between the genders has undergone great changes over time. As fathers perceive a growing desire to play an active part in childcare [50], the number of stay-home-fathers has increased together with an overall greater involvement in domestic tasks [51, 52]. Simultaneously, a rise in women’s economic contributions has been observed, reducing the likelihood of fathers to be the family’s main or even sole source of income [40, 41]. As men and women’s roles become more similar, people’s gender role stereotypes might change concurrently (see Social Role Theory; [53, 54]). These changes could improve gender parity, as traditional gender role stereotypes play an important part at maintaining inequalities between men and women in society [55, 56]. Additionally, to the extent that observers perceive these social roles and gender stereotypes to develop, they might also perceive power disparity between men and women to diminish (see also; [8]). Thus, prevalent gender differences in public as well as private powers might be less pronounced in future societies. However, many men still fill in the role of the major financial provider, while women adapt the role as the main caretaker in the family [42], earn less than men for comparable work and are underrepresented in high status positions at work and politics [13–15, 36]. Therefore, diminishing inequities in public domains as well as raising awareness of these problems is still of major importance.

Second, internet users are not completely representative of the general population [57, 58]. Since the studies were conducted via Mechanical Turk, the generalizability of our finding to the whole US population might be limited. As we neither included a measure of the participant’s gender role attitudes, nor investigated other potential influences like age, education or social status, the results might look different when comparing people of different age, educational and financial backgrounds. For example, previous research suggests that younger age as well as higher education and income is associated with more egalitarian or critical attitudes towards traditional gender roles [55, 59]. However, despite the generalizability problems of internet-based studies, previous research has shown that internet users are still more representative than the more commonly used convenience samples [57, 58]. Finally, the large sample sizes ensured a sufficiently high statistical power.

Finally, as noticed by the various definitions of general power, the traditional use of the concept as delineated by Townsend et al. [3] refers to different forms of power, in particular ‘power over’ and be ‘powered by’. To elaborate the matter further, power is exercised on many levels, such as between and within areas of life (e.g., [2, 3, 60, 61]). For example, a person might have high power over decisions made in their private life (e.g., parenting), while simultaneously having to obey the orders from their supervisor in a work setting. Furthermore, decision-making can also differ within domains; for example, one member of a romantic dyad could have control over the economy, and the other make decisions on how to manage specific household tasks and children's education. Thus, power is a complex construct. In the present work, forms and levels of power were not explicitly addressed. Participants freely interpreted the concept of power in general and in relation to gender, social relationships, and work. Using the same method, it would be of interest to explicitly investigate different forms and levels of perceived power.

Furthermore, semantic analyses of power could be expanded. In the present study, LSA was used to analyze power associations. LSA also enables analysis of more extensive collections of text then we collected in the present study. In a study by Townsend and colleagues [3], Mexican women’s manifestly described power experiences were analyzed with a qualitative approach. In research by Karlsson et al. [28], two quantitative methods were used to analyze women's and men's memory reports. LSA to measure the latently described (i.e., the underlying meaning in the expressed words) and linguistic inquiry word count (LIWC; e.g. [62]) to measure the manifestly described (i.e., the actual words said). In line with the present work, it was found that the female participants latently were oriented towards social relationships in their memory descriptions than the male participants were. Thus, there are many analytic methodologies that could shed further light on the dimensions of power.

In sum, the results illustrate how people’s definition of power is biased towards public domains, which are further stronger associated with men than women. However, it was also found that women have more perceived power in the private domain. This highlights the need to broaden our perception of power, as power can be exerted in many important domains in life. Because when considering many different domains of power, weighted by their relative importance (PWP), we demonstrated a lack of domain-independent gender differences in preference weighted power. Taken together, the results of this project may significantly change how we perceive power.