Case Study 4 - BU618 Pt 2
Problem Statement
Ethnic diversity is a vital driver of innovation, better decision-making, and enhanced customer understanding, particularly in diverse markets (Kandola & Fullerton, 1994). However, achieving equitable representation of underrepresented groups (UGs) across all organizational functions, particularly in industries with historically low diversity, remains a challenge. In the financial services sector, understanding how UGs are distributed across functions is critical to identifying barriers and developing strategies to improve diversity.
This analysis focuses on UG representation in two key functions within a large financial organization: Sales and Professional Services. Specifically, we aim to test whether significant differences in UG representation exist between these two functions. Additionally, we explore how other factors—such as team location (London or not), group size, the number of female team leads, and the proportion of male employees—may influence UG representation.
By investigating these factors through an independent t-test and multiple linear regression, this study seeks to identify potential systemic barriers that may prevent UGs from being equitably represented across all functions. The results will inform targeted diversity and inclusion (D&I) initiatives designed to foster a more inclusive workforce in line with commercial goals and employee engagement objectives.
Data Description
The dataset includes 29,976 employees across 928 teams in the UK, aggregated from an employee engagement survey. The primary variables are:
• Function: Sales or Professional Service
• UG Representation: Proportion of underrepresented groups in each team
Ethnicity data was collected voluntarily, and only teams with 10+ employees are included for anonymity. An independent t-test will assess differences in UG representation between functions, and a multiple regression will explore additional influencing factors.
Methods of Analysis
To examine whether there is a significant difference in the proportion of UGs between the Sales and Profes- sional Service functions, two primary statistical tests were used: Levene’s test for homogeneity of variances and an independent t-test.
Levene’s Test for Homogeneity of Variance
Levene’s test was applied to assess whether the variances in UG representation across the two functions were equal. This step ensures the validity of the independent t-test’s assumptions regarding equal variances.
Independent Samples t-Test
An independent t-test was conducted to compare the proportion of male employees—used here as a proxy for UG absence—between the Sales and Professional Service functions. Despite Levene’s test indicating unequal variances between the groups, the t-test was performed assuming equal variances due to the robustness of the test with large sample sizes. The null hypothesis posited no significant difference in UG representation between the two functions.
Multiple Linear Regression
To further explore the factors influencing UG representation, a multiple linear regression model was fitted. This model included five predictors: team location (London or not), function (Sales or Professional Services), group size, the number of female team leads, and the percentage of male employees in the team. The aim was to assess how these factors interact to affect UG representation and to identify any additional disparities that might exist beyond functional differences.
Key Results
The results of Levene’s test indicated a significant difference in variances between the two functions (F(1, 925) = 30.99, p < .001), suggesting that the variances are not equal. However, the independent t-test was conducted assuming equal variances, given the large sample size. The t-test revealed a highly significant difference in UG representation between the Sales and Professional Service functions (t(925) = 22.89, p < .001). Specifically, the proportion of male employees was significantly higher in the Sales function (M = 71.26) compared to the Professional Service function (M = 44.40). The 95% confidence interval for the mean difference ranged from 24.56 to 29.16, confirming a substantial disparity in UG representation between the two groups.
The multiple linear regression analysis further highlighted that both team location and function significantly impacted UG representation, with the overall model explaining 17.4% of the variance in UG prevalence (Adjusted R² = 0.168, F(5, 724) = 30.49, p < .001). Key findings include:
• Team Location (London or Not): Teams based outside of London had significantly lower UG represen- tation compared to those based inside London (B = -0.085, = -0.36, p < .001).
• Function: Professional Service teams exhibited higher UG representation than Sales teams (B = 0.038, = 0.17, p < .001).
Other factors, such as group size, the number of female team leads, and the percentage of male employees, did not significantly contribute to UG representation (all p > .05).
Interpretation and Conclusion
The findings reveal a significant difference in UG representation between the Sales and Professional Service functions, with Sales teams showing a higher male proportion and lower UG representation. This suggests that UGs are underrepresented in customer-facing, higher-paid Sales roles compared to non-customer-facing Professional Service roles. The analysis also indicates that teams outside London have notably lower UG rep- resentation than London-based teams, which may highlight geographic disparities in recruitment or retention practices.
These findings point to the need for targeted D&I strategies that address both functional and geographic disparities in UG representation. Specifically, the organization may need to examine recruitment practices, particularly in areas outside London and in customer-facing roles, to ensure more equitable representation of UGs across all areas.