The Top 50 competition and related specialty list rankings are based on empirical data obtained through organizations completing and submitting the online survey.
Participation is free and companies that do business with Fair360 receive no preferential treatment.
Since 2001, this assessment is the most extensive, data-driven analysis of workplace fairness initiatives and outcomes focused on leadership accountability, organizational programs and practices as well as human capital metrics at some of the largest U.S. employers.
The assessment provides detailed insights into representation metrics across several dimensions and intersectionality segments including race/ethnicity, gender identity, sexual orientation, disability status and veteran/active military status of employees, leadership and board of directors. In addition, the assessment captures information regarding supplier diversity.
The insights and benchmarks derived from the assessment enable employers with 750 or more full-time employees to advance their fairness strategies to improve hiring and retention, leadership accountability, talent programs, workplace practices and supplier diversity.
To better understand the language used in the assessment, view our Top 50 Glossary of Terms.
The following is an overview of the various lists derived from leveraging the data submitted by participating employers:
Top 50 Companies
Since 2001, this list is composed of employers who had the highest overall aggregated scores as calculated using Fair360’s ranking algorithm. The results from companies making it into the Top 50 serve as the basis for establishing Fair360’s benchmark indices.
Noteworthy
This list is composed of employers whose data indicates they have the potential to make it into the Top 50.
Specialty Lists
These lists are composed of employers whose data indicates they are performing at a high level in a specific area of workplace fairness or corporate social responsibility.
Specialty Lists:
Industry & Geography
The following are the associated lists:
- Top Hospitals and Health Systems
- Top Regional Companies
- Top Utilities
We apply the Top 50 ranking algorithm and methodology to identify the top employers within each of these specific cohort groups. The Regional list is composed of companies that compete on a regional basis for talent.
Specialty Lists:
Representation
These lists recognize companies for outstanding results regarding specific representation practices.
The following are the associated lists:
- Top Companies for Asian American and Pacific Islander Executives
- Top Companies for Black Executives
- Top Companies for Board of Directors
- Top Companies for Latino Executives
- Top Companies for Native American Executives
- Top Companies for Talent Acquisition for Women of Color
- Top Companies for LGBTQ+ Employees
- Top Companies for People with Disabilities (PwD)
- Top Companies for Veterans
Specialty Lists:
Talent Development & Leadership Accountability & Data Transparency
These lists recognize companies for outstanding results regarding specific workplace fairness programs and practices.
The following are the associated lists:
- Top Companies for Employee Resource Groups (ERGs)
- Top Companies for Mentoring
- Top Companies for Sponsorship
- Top Companies for Supplier Diversity
- Top Companies for Executive Diversity Councils
Development of the Model
The development of the Fair360 Top 50 assessment, ranking methodology and model began in 2001. Each year, we review and adjust the model to be relevant to current and emerging fairness practices, programs, policies and dimensions of difference. Beginning in 2019, Fair360 adjusted the model to account for extreme outliers at the factor level as to not skew the mean – a process known as “fissurization.” In addition, we applied dynamic weightings to human capital metrics to utilize available U.S. Equal Employment Opportunity (EEO) industry data tables to calculate a ratio to increase efficacy when comparing companies within the same industry and normalizing the data across all companies to power the ranking.
All factors were judged against the following criteria:
- They needed to be rigorously developed based on sound statistical methods.
- They needed to capture the essence of specific workplace fairness practices.
- They needed to have a sufficiently high correlation with the other factors of the same dimension—providing assurance that each one was measuring the same aspect of organizational effectiveness.
Data Collection
The Fair360 Top 50 ranking is based on empirical data obtained through organizations filling out the survey via the Momentive survey platform. Data is retained to allow for year-to-year comparisons.
How Top Employers Are Measured
In the case of Fair360’s model, there are over 1,400 unique factors that fall into the five categories of corporate performance relative to fairness.
The model captures empirical data to identify and assess an organization’s assets (leadership, programs, practices and policies) and the resulting outcomes (human capital and supplier diversity) to determine all the associated lists. The following are the five categories of data collected to measure, rank and benchmark participating employers.
- Human Capital Metrics: Four levels of management, total management, total workforce, promotions, retention and new hires
- Organizational Structure: Board of Directors, CEO and Senior Leadership, and Executive Diversity Councils
- Talent Programs: Employee Resource Groups (ERGs), high potential, mentoring and sponsorship programs
- Talent Acquisition, Development & Management: Talent acquisition, training and talent management
- Supplier Diversity: Spend with companies owned by people from underrepresented groups
In calculating Top 50 scores, a slightly different weight was computed for each category based on the degree to which it was found statistically to contribute to overall organizational effectiveness and performance.
Fundamentally, the model rests on the belief that all five categories are interrelated and influence each other over time—what social scientists call “reciprocal causation.”
Fair360 has conducted longitudinal research to better understand how different parts of the model relate to one another—that is not the immediate goal of this ranking. Rather, the main purpose is to provide a “snapshot” of overall organizational effectiveness and performance at a point in time.
In the case of Fair360’s specialty lists, the same methodology, model and categories used for determining the Top 50 list are applied. However, we include only the relevant data points for the specialty list to power the specialty list rankings. For example, the Top Companies for LGBTQ+ Employees references the same category framework. However, the weightings are unique to identify best-in-class companies relative to their practices and outcomes in support of LGBTQ+ employees and suppliers.
As a point of caution, care should also be taken when comparing results with those from previous years because the universe of companies analyzed can expand or shrink from year to year.
Scoring of Companies
Companies are scored by:
- Standardization of responses where it’s needed, i.e., if the responses are counts, converting them into respective percentages. If the responses are dichotomous, creating composite variables by adding all the dichotomous responses within a subcategory construct.
- Post step one, for most questions, we calculate the participant’s z-scores per standardized questions or the composite variables with exception of a few where we calculate the ratio between the standardized percentages to the percentages from the latest available EEO dataset. For these groups, predetermined weights are assigned based on the ratio. In addition, for all the questions with a z-score, a constant term is added to the z-score to move the score toward a positive scale.
- The z-score and weighted (in the case of EEO comparisons) data points are then further standardized using minmax scaler, which makes the data fall in the range of 0 to 1.
- In grouping the data, based on the sub-category of the survey and weights that particular subcategory carries, the items from step three are added and weights are applied. Weights for each subcategory are in the range of 0-100% (the weights add up to 100%), thus output from this stage is in the range of 0-100. These subcategories roll up to the five final categories, i.e., Human Capital Metrics and Leadership Accountability.
- These five final categories are again added with specific weights (in the range of 0-100%) to determine the company’s final score.
- Final score determines the rank position on the Top 50 and/or a specialty list(s).
Factor analysis was calculated to determine the validity of the survey categorization.
Population of Ranked Companies
The Fair360 Top 50 results are derived exclusively from survey submissions from employers with at least 750 employees in the U.S. Companies are evaluated within the context of their own industries. Subsets of the same data submission are used to determine our specialty lists. Participants must:
- Answer all applicable questions in this assessment.
- Complete and submit the NOD Disability Employment Tracker.
- Submit a signed letter by the CEO or U.S. Managing Partner, Chief Human Resources Officer (CHRO) or a corporate officer other than the Chief Equity Officer or person responsible for workplace fairness, affirming all data submitted is accurate. This can be submitted via DocuSign, which is directly accessible via the survey collection portal.
*Please note, participants still receive a free report card even if the NOD Employment Tracker is not completed.
Hospitals and Health Systems
Hospitals and Health Systems complete the same survey as other major U.S. employers competing for a spot on one of the lists. There are questions that only hospitals and health systems will complete.
Missing Data
To handle missing data, data is stored in two different formats: one for reporting and another for the algorithm. For the algorithm, most of the missing data is imputed by 0.
Thus, for part of the survey where we have percentages, imputing 0 would return 0%, composite variables are not impacted as they are summations and the addition of missing data or 0 has no impact.
For reporting, the data is shown as is, except for certain cases where it breaks general mathematical rules.
Testing and Validating the Model
To build the model, Fair360 utilizes multiple linear regression analysis to examine the relationship between measured variables (factors) and latent constructs. This approach allows for the analysis of the entire model simultaneously, including all factors captured in the assessment, as well as all five categories of corporate performance and overall organizational effectiveness. Each year, we run Cronbach alpha testing to assess the reliability, or internal consistency, of the category and items in the assessment. For the 2024 competition, the Cronbach Alpha Reliability score for the overall model was .93.
A series of tests are run to ensure:
- Construct validity (the degree to which the factors measured what they claimed to measure)
- Reliability (freedom from random error)
In the end, the model confirmed that:
Taken together, the five categories reflect a single higher-level construct (organizational “effectiveness”) and that each has a substantial factor loading on that construct, averaging 0.65 since 2017.
Future of the Model
Fair360 is committed to a process of continuous improvement regarding the practices, methodology and processes for capturing, analyzing and reporting data related to the Top Companies competitions and benchmarks. While the underlying model framework (five categories) has been consistent through the years, new factors may be introduced if they are determined to serve as better proxies for the associated fairness practices, policies, programs and outcome metrics (human capital and supplier diversity).
Fair360 will also continue to analyze its computational methods to ensure that we deliver fair and trustworthy rankings and benchmarks.
If you have any questions or comments, please email top50@fair360.com.
