Outline

2

Human Capital Analysis

Name

Institution

Course

Professor

Date

Human Capital Analysis

This report scrutinized a data set consisting of crucial metrics related to human capital to form a clear picture of components that influence employees’ experience, engagement, and diversity in organizations. Such key indicators as team diversity representations, team structures, the distribution of places, and surveys were scrutinized.  Descriptive statistics were utilized to get insight into the features, distributions, and main parameters of each variable. A visualization and a summary of these foundational characteristics, presented prior to inferential statistical tests, helped understand the results obtained. Relationships between factors were then explored using inferential methods, including analysis of variance (ANOVA), correlation and regression techniques, t-tests, etc. These techniques allowed for the discovery of any connections observed. Together, those techniques helped to establish more reliable results and illustrated the development of these indicators for organizational human capital. The strategy assists policymakers in targeting methods that will help maximize critical performance indicators by using the analytics tool.

Descriptive Statistics

The dataset involves information from departmental entities with variety measurements, employee socioeconomics, and overview reactions. There are 927 total observations across various categories. DepartmentGroupNumber codes designate individual units, with an average size of 470 members but a standard deviation of 270.05, showing a vast range from 1 to 937 employees. This wide variability implies the potential for differing resources and challenges depending on a group's scale (Curtis & Mont, 2020). Correlation tests later may uncover connections between size and other priorities, like diversity support.

Diversity is represented through BAME proportions, averaging 11.84% overall yet fluctuating substantially from 0-45% between groups, as the standard deviation of 11.33% indicates. This high variability in minority representation at the level of individual work units raises questions about equitable experiences and opportunities between different teams. Gender makeup trends are 58% male on average, according to PercentMale. However, the large standard deviation of 22.27% again confirms that percentages diverge markedly, from nearly all women or men in some areas to a balanced mix elsewhere (Ordonez-Ponce et al., 2020). Disparities in gender composition risk uneven support, particularly for numeric minorities in imbalanced contexts.

GroupSize finds teams encompass 32 people on average, with a minimum and a maximum of 10 and 71, respectively, with a standard deviation of 16 indicating wide variation. Comparisons may uncover connections between size and important dependent variables like productivity or satisfaction given different logistical demands and social dynamics in small vs. large teams. Measures of spread, like standard deviations, are effective in uncovering complexities within aggregate means (Antonioli et al., 2022). For example, averages paint diversity and gender balances positively across the whole sample, yet these same indicators show significant heterogeneity amongst observations. Such nuanced detail enriches interpretation and application.

While demographics shape the framework, perceptual factors also impact outcomes. Later tests explore connections between structural metrics and internal survey data reflecting experiences of engagement, integrity, and leadership. Multi-dimensional analysis enhances understanding of both direct and indirect drivers. A thorough assessment of variability provides a basis for hypothesizing differences based on subgroup characteristics like disproportionate BAME representation or extremely large or small departments. This supports a targeted examination of potential experience variances and inequities (Wulff & Villadsen, 2019). Accordingly, the descriptive analysis offers a basic setting and creates questions justifying further quantitative and qualitative requests to fully appreciate human capital indicators.

Correlation Analysis

The dataset's correlation matrix analyzes the connection between 21 distinct factors estimating employee engagement, authoritative integrity, management, and their subcomponents. Several noteworthy correlations emerge: The nine EMPsurvEngage variables zeroed in on unambiguous aspects of engagement are profoundly correlated with one another, ranging from .4365 to .8400. This suggests that employee ratings of distinct engagement facets tend to move together. EMPsurvEngagement, representing overall engagement, has strong correlations with its sub-variables, ranging from .7252 to .8818.

EMPorgIntegrity variables also exhibit high internal correlation, from .3405 to .6994, indicating different integrity factors are closely linked. EmpSurvOrgIntegrity closely tracks this set of variables, ranging from .5789 to .6883. EMPsurvSUP ratings mirror this trend, with correlations between .5585 and .8058. EmpSurvSupervisor parallels this subdomain’s specific measures from .6969 to .8917.

Moderate correlations are seen between engagement, integrity, and supervision domains. This implies some interaction but also independence between these experience categorizations. Correlations between observed variables and their aggregate scores present convergent validity for the composite measures (Antonioli et al., 2022). Strong relationships within subsections point to their internal consistency. Overall, the relatively high positive correlations signify that survey answers assessing similar constructs tend to co-occur, as would be expected. This lends credibility to utilizing the multi-item measures to represent broader attitudes.

Paired Sample T-tests

T-test 1

A paired samples t-test was conducted to compare the Location and LondonorNot variables. There were 928 paired observations. Location had ( M = 1.426, SD = 0.245) and LondonorNot ( M = 2.122, SD = 0.715). A Pearson correlation revealed a strong positive relationship between the variables, r = .894, p <.001. The t-test revealed significant differences in the means of the two variables, t(927) = 46.08, p < .001, with Location rated significantly lower than LondonorNot on average.

T-test 2

A second paired sample t-test was used to compare NumberTeamLeads and NumberFeMaleTeamLeads. There were 181 paired observations. NumberTeamLeads had ( M = 4.608, SD = 5.373) and NumberFeMaleTeamLeads ( M = 0.669, SD = 1.045). A Pearson correlation showed a weak positive relationship between the variables, r = .137, p < .001. The t-test revealed statistically significant differences between the means, t(180) = 22.07, p < .001, with NumberTeamLeads significantly higher than NumberFeMaleTeamLeads on average. These results indicate meaningful differences in the descriptive statistics of the paired variables.

Analysis of Variance (ANOVA)

ANOVA 1

A one-way ANOVA was conducted to determine if there were differences in group size based on the percentage of males. The independent variable was percent males, and the dependent variable was group size. There were 927 participants across two levels of male percentage. Levene's test indicated homogeneity of variances ( p >.05). The ANOVA was significant: F (1, 1852) = 866.39, p <.001, ηp2 =.32. Thus, group size scores differed significantly based on the percentage of male levels.

ANOVA 2

A second one-way ANOVA examined differences in group size based on the independent variable function. There were 927 participants across two levels of function. Levene's test was not significant ( p >.05), indicating homogeneity of variances. The ANOVA was significant: F (1, 1852) = 3341.11, p <.001, ηp2 =.64. Therefore, group-size scores differed significantly based on levels of function. Given there were two levels of the independent variable, post-hoc tests were less warranted.

Regression Analysis

Multiple regression

Multiple linear regression was performed to predict PercentMale from GroupSize, NumberTeamLeads, NumberFeMaleTeamLeads, Location, LondonorNot, and Function. The linear combination of predictors was significantly related to PercentMale, F(6, 173) = 18.13, p < .001. The sample multiple correlation coefficient was .62, indicating that approximately 38.6% of the variance of PercentMale in the sample can be accounted for by the linear combination of predictors. In terms of individual predictors, only NumberFeMaleTeamLeads was significantly related to PercentMale (b = -3.96, t(173) = -2.88, p < .01).

Simple linear regression

A separate simple linear regression was conducted to predict NumberFeMaleTeamLeads from PercentMale. The regression equation was significant, F(1, 178) = 7.98, p < .01, with PercentMale accounting for approximately 4.3% of the variation in NumberFeMaleTeamLeads (R2 = .043). PercentMale negatively predicted NumberFeMaleTeamLeads, such that higher percentages of males were associated with fewer female team leads (b = -.009, t(178) = -2.83, p < .01).

Conclusion

The descriptive statistics, correlations, t-tests, ANOVAs, and regressions provided valuable insights into the composition and relationships within the dataset. Demographic variables such as gender, ethnicity, and team size showed considerable variability that warrants further exploration. Significant differences and predictors emerged between variables. While direct and indirect links were identified, more complex multivariate modeling is needed to fully understand the nuanced interplay of factors. Overall, the analyses found meaningful patterns and came up with hypotheses that can guide further quantitative and qualitative research into understanding important organizational and experiential indicators.

References

Antonioli, D., Ghisetti, C., Mazzanti, M., & Nicolli, F. (2022). Sustainable production: The economic returns of circular economy practices. Business Strategy and the Environment. https://doi.org/10.1002/bse.3046

Curtis, S. K., & Mont, O. (2020). Sharing economy business models for sustainability. Journal of Cleaner Production, 266, 121519. https://doi.org/10.1016/j.jclepro.2020.121519

Ordonez-Ponce, E., Clarke, A. C., & Colbert, B. A. (2020). Collaborative Sustainable Business Models: Understanding Organizations Partnering for Community Sustainability. Business & Society, 60(5), 1174–1215. https://doi.org/10.1177/0007650320940241

Wulff, J. N., & Villadsen, A. R. (2019). Keeping it within bounds: Regression analysis of proportions in international business. Journal of International Business Studies, 51(2), 244–262. https://doi.org/10.1057/s41267-019-00278-w