In this example, we have team-level data from a UK-based financial company. The company has 29,976 employees in 928 teams which are divided in two main functions – sales and professional service. In short, sales staff are customer-facing employees; whereas professional service staff are non customer-facing employees such as product development, finance, marketing.

**The Question**

The question we want to answer is whether there is a significance difference in the percentage of the employees with different ethnic backgrounds between sales and professional service. The rationale is that, potentially, the prevalence of minority groups across the teams may, to some extent, depend on the type of the team (or function in our example). For instance, the sales team tend to draw more employees with different ethnic backgrounds.

We can use the variable BAME in the dataset to test our hypotheses. The term “BAME” – Black, Asian or Minority ethnic – is primarily used in UK and is appropriate since the data is from a UK-based financial institution. The data type of the variable BAME is the percentage of the team made up of applicable employees. There is another variable in the dataset called “PercentageMale” (the percentage of the team made up males). Since we already discussed the topic of gender and job role in the last post, we can revisit the gender issue in this example as a bonus.

Independent Samples T-test

Independent Samples T-test

The method we use to analyze is **Independent Samples T-test**. It’s used for comparing whether the means of two samples are statistically different from each other. The independent variables are sales and professional service (two categorical variables). The dependent variable is the percentage of BAME (a continuous variable).

**The Results**

In regard to the percentage of the ethnic diverse employees in sales and professional service, the t-test is significant: t(640.61) = -5.66 and p-value is < 0.001. The bigger the absolute t value is, the more differences there are between the two groups.

Welch Two Sample t-test data: sales.bame and prof.bame t = -5.6553, df = 640.61, p-value = 2.345e-08 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: -0.06335727 -0.03069866 sample estimates: mean of x mean of y 0.09683168 0.14385965

In plain English, in comparing the sales and the professional service functions, the proportion of BAME staff is significantly lower in sales than in the professional service. The average percentage of BAME in sales is 9.68%; whereas that of the professional service is 14.39%. The analysis shows that the likelihood of this difference occurring by chance is less than 1 in 1,000. It suggests that we have an issue that needs attention within the sales functional group.

In regard to the percentage of male employees in the two groups, the t-test is also significant: t(828.77) = 22.56 and p-value is < 0.001.

Welch Two Sample t-test data: sales.male and prof.male t = 22.557, df = 828.77, p-value < 2.2e-16 alternative hypothesis: true difference in means is not equal to 0 95 percent confidence interval: 24.52151 29.19569 sample estimates: mean of x mean of y 71.26096 44.40235

We can conclude that the proportion of the male staff in sales is significantly higher than the professional service. The average percentage of male staff within sales is 71.26%, and by comparison the average proportion of male staff within the professional service function is 44.4%. The analysis indicates that we have a clear issue of male dominance within the sales functional group. It potentially needs attention if we want gender proportions to be similar across the organization.