Although businesses now employ more female managers than ever before, women’s advancement into senior leadership roles remains much slower than for men. While there are a variety of structural causes driving gender inequity in the workplace, one important factor is the disparity in how men and women are given developmental feedback. Identifying and reducing bias in feedback on past performance is somewhat more straightforward, since this sort of feedback tends to be more quantitative — but feedback focused on how employees should change and grow as leaders in the future is fundamentally qualitative, making it much harder to analyze.
However, with computerized text analysis, it’s possible to quantify differences in feedback between men and women, as well as how these differences can drive employees down different leadership paths. In our recent study, we used a form of machine learning known as “topic modeling” (which has recently become popular as a tool for analyzing political Tweets — see the Methodology Corner below for more details) as well as comprehensive qualitative analysis to investigate a large, complex dataset of developmental feedback.
Specifically, we explored gender differences in a dataset of open-ended written feedback for 146 mid-career leaders, provided anonymously by more than 1,000 of their peers and leaders while taking part in a leadership development program. We also asked participants to rate their leaders’ performance numerically, giving us a quantitative baseline for comparison that enabled us to control for objective differences in leaders’ performance.
Based on these analyses, we found four key differences in how advice was framed for female leaders and for male leaders:
It is important to note that all of these messages were generally framed as positive, and it is possible that the people providing this feedback genuinely believed in the potential of these women to reach senior leadership roles. However, providing men and women with equally positive feedback does not mean that the feedback is free of gender bias — nor do good intentions eliminate the very real harm that this bias can cause. Our research shows that even if it is ostensibly positive, feedback provided to women tends to be less actionable and less useful for leadership progression than feedback given to men, making it less likely that women will advance to more senior positions.
How can gender bias in developmental feedback be corrected?
The good news is, this subtle bias can be mitigated through deliberate action. To make their developmental feedback more gender-inclusive, managers must scrutinize the messages they communicate in that feedback.
Importantly, it is vital for managers to examine how they provide feedback not just to their female employees, but to their male employees as well. After all, the goal is not simply to treat women more like men, but rather, to encourage leadership practices in all employees that include the best of both traditionally feminine and traditionally masculine traits. For instance, both assertiveness and collaboration are essential for leadership. As such, to effectively combat gender bias, managers should encourage all employees to develop both qualities — which may (on average) mean more conversations with male employees about developing collaboration skills, and more conversations with female employees about developing assertiveness.
So what does this look like in practice? Through our research, we found a few simple ways that managers can overcome their biases and provide more equitable feedback in each of the four areas identified above:
Vision: Too often, women get pigeonholed into delivering, rather than developing vision. To help them move past their areas of technical expertise into broader leadership roles, managers should encourage female employees to think strategically about the wider context in which the organization operates. Invite them to develop and articulate a personal vision for their team, rather than overly focusing on operational details and execution — and find opportunities to publicly recognize these contributions. Some conversation starters include:
“What is your personal vision for the team/organization?”
“How does it fit in with the bigger picture?”
“How can you involve others in developing this vision?”
Conversely, encouraging men to focus both on visionary and operational skills means that beyond vision-setting, developmental conversations should consider tactical areas for improvement. Some questions to ask include:
“What are the operational or tactical aspects of the job you need to pay more attention to?”
“What areas of expertise do you need to develop?”
Political Skills: Workplace politics can seem undesirable, but research shows that political behaviors such as networking, negotiating, and influencing others are not only positive, but vital for progression to senior roles. Simply “coping” with politics (or worse yet, attempting to avoid it entirely) is a reactive mindset that tends to get in the way of effective leadership.
Instead, managers should encourage their female employees to embrace a proactive political mindset. Help them to appreciate the importance of political engagement, and encourage them to map out key players, identify hidden agendas, and deliberately build relationships — not just with their peers, but with those in power, who can help them get things done. Some conversation starters include:
“How do you feel about workplace politics? What might be constructive ways of engaging in politics, in your role?”
“Who are the key players in your work area/organization and what are their agendas?”
“Who do you need to form relationships with and whose support do you need to progress towards your leadership goals? How will you do that?”
Similarly, men might be prompted not to focus just on developing strategic relationships with those above them in seniority, but also to foster supportive alliances with their peers. Some conversation starters include:
“How might you build connections with colleagues outside your normal groups?”
“Which of your colleagues are you least likely to work with, and how might you — and they — benefit from developing a closer relationship?”
Asserting Leadership: Encouraging men to be assertive while asking women to focus on getting along with others implicitly gives your male employees a mandate to forge ahead and take on leadership roles, while women are directed towards more collaborative endeavors. Instead, managers should invite women to be explicit about their leadership aspirations and proactively pursue development opportunities. Some conversation starters include:
“What are your leadership aspirations?”
“How will you pursue them? What and who might enable you?”
“In a year’s time, what steps will you have taken to achieve that leadership role?”
Importantly, collaboration is also an important component of good leadership. As such, in addition to encouraging their male employees to pursue their leadership aspirations, managers should also invite men to develop collaboration skills and a team-oriented mindset. Some conversation starters include:
“How team-oriented and collegiate are you in various work contexts?”
“In what ways could you develop these skills?”
Confidence: In our research, we consistently found that men were told they needed to develop confidence for specific skills, such as managing meetings or communicating with different audiences, while women were given more generic advice to simply “become more self-confident” without concrete guidance around how to do that. Indeed, past research has shown that decision-makers often cite lack of confidence as a justification for women’s slower progression into senior roles — without offering specific, actionable feedback for how to develop that confidence.
To address this during developmental conversations, managers should discuss confidence with respect to specific domains or skill sets, rather than talking about self-confidence as a generic trait (and thus something that can be inherently lacking). Try starting the conversation by asking:
“What specific skills do you feel less confident about? How can you develop them?”
“What skills do you feel confident about? How can you better leverage them in your role?”
“What behaviors can you use to demonstrate your confidence to others?”
Developmental feedback (provided either informally or via official management processes) is a significant yet often-overlooked driver of professional growth. It is one of employees’ few explicit opportunities to learn about how they should change and develop as a leader, and as such it plays a major role in paving the way to leadership. Our research demonstrates how differences in developmental feedback can direct women along different — and less effective — leadership pathways than men, creating long-lasting gender inequities.
Luckily, understanding this subtle gender bias is the first step towards correcting it. By identifying and refocusing developmental conversations in the four key areas of bias we’ve outlined above, managers can begin to overcome their unconscious biases and more effectively support the development of all of their employees.
Key Messages in Developmental Feedback Provided to Male and Female Leaders
Source: Elena Doldor, Madeleine Wyatt, and Jo Silvester© HBR.org
Methodology Corner: Tackling Qualitative Big Data with Topic Modeling When people talk about “big data,” they’re typically referring to databases containing large volumes of numerical data, such as compensation rates, demographic breakdowns, or other quantitative metrics. But big data can also include qualitative data, such as the extensive spoken and written information collected by managers and HR departments in the form of staff surveys, performance evaluations, and developmental feedback. While this kind of data can be trickier to analyze, a tool called “topic modeling” enables the automated analysis of these large volumes of text. Topic modeling, a form of Natural Language Processing (NLP), uses machine learning to impose structure on textual data by identifying common themes, or “topics,” without the need for manual analysis. The method uses an algorithm that looks at how often individual words appear, where they appear in different types of documents, and how different words are associated with each other (for example, it might find that the word “time” often appears next to the word “management”). It then creates keyword groupings which can be used to infer common topics. For example, an analysis of news websites in 2020 might find that the words “testing,” “vaccine,” “lockdown,” and “virus” often appear together, which could form the topic “coronavirus.” The method also makes it possible to see which source documents contain greater concentrations of particular topics, and analyze the data with respect to factors such as gender, ethnicity, work group, and even work outcomes such as performance and engagement. This makes it possible for organizations to answer questions like, “Do the finance and marketing departments give different kinds of feedback in their staff surveys?” or, “On average, do employees from minority backgrounds receive emails that are different in content from those that white staff receive?” This can help uncover biases in communication that might otherwise go unnoticed. Topic modeling can be an efficient method for analyzing textual big data — but as with any tool, it’s important to consider its ethical implications. Topic modeling can be used to mine emails, Slack messages, or even voice-to-text phone conversations, creating an ethical minefield of privacy issues. It’s important to ensure that employees are aware of how their work communications may be used, and that personal or sensitive material is kept separate from content used for analysis (since it may be challenging to prevent an unsupervised machine learning model from extracting information from non-work-related data). Topic modeling also has several practical limitations. First, although the process is largely automated, it does require human input on how many topics to extract, and the parameters of those topics. This means that models can be hard to replicate, because analysts who make different choices could end up with different topics being extracted from the same data set. In addition, interpreting the output of a topic modeling algorithm and translating sets of representative words into coherent themes can be challenging: Although the raw output from topic modeling software can provide some insights, the results will generally still need to be explored in greater detail to understand the nuances of the identified topics. As such, topic modeling can be a useful tool for organizing text, but based on our experience, it should be considered a starting point — rather than a finished product — for qualitative big data analysis.
Elena Doldor is an Associate Professor in Organizational Behaviour and Co-Director of the Centre for Research in Equality and Diversity at the School of Business and Management, Queen Mary University of London, UK. Her research examines gender and ethnic diversity in leadership.
Madeleine Wyatt is an Associate Professor in Human Resource Management at the University of Kent, UK and a Chartered Occupational Psychologist. Madeleine’s research examines diversity at work and the role informal and political processes play in individuals’ leadership journeys.
Jo Silvester is a Professor in Work Psychology at Loughborough University, UK. Jo’s research examines leadership emergence and effectiveness in complex work environments, with a particular focus on politicians and political work.
Source : https://hbr-org.cdn.ampproject.org/c/s/hbr.org/amp/2021/02/research-men-get-more-actionable-feedback-than-women