Using Multiple Regression to Examine Compensation Practices (Overview)

6

Click here to load reader

description

This Chapter presents an overview of an exceptionally powerful statistical technique that can be applied to determine whether significant differences exist among applicant or employee groups. Specifically, Multiple Regression (MR) will be reviewed and then shown how it can be applied for evaluating employee compensation for the purpose of determining whether males or whites are compensated more highly than females or minorities due to reasons that are not legitimately job related. Clearly, the practice of making simple comparisons between the average salary of males and females or whites and minority groups is inadequate for developing a complete understanding of whether or not evidence exists for discrimination. What is needed is a statistical tool that allows a researcher to compare the average salaries of various groups of employees while controlling for the legitimate job-related factors that are also influencing employee pay. MR is a statistical tool that allows practitioners to do exactly that. Because MR is such a powerful tool for finding differences in compensation that cannot be explained by legitimate job-related factors, it is the method of choice that has been used by the courts for decades, and has more recently been endorsed by the OFCCP. Learn more about the BCG Institute for Workforce Development by going to www. BCGInstitute.org Visit http://bcginstitute.org/?AIBookSeries to learn about the Adverse Impact and Test Validation webinar series based on Dr. Biddle’s book.

Transcript of Using Multiple Regression to Examine Compensation Practices (Overview)

Page 1: Using Multiple Regression to Examine Compensation Practices (Overview)

Using Multiple Regression Analysis to

Examine Compensation Practices

BCGi: Adverse Impact & Test Validation Book Series

Page 2: Using Multiple Regression to Examine Compensation Practices (Overview)

Multiple Regression and Pay Analysis

• Multiple Regression (MR) is a robust statistical method that can

effectively determine whether potential pay discrimination exists

in a given job or group of jobs

• MR is able to make this determination by statistically controlling

for differences in job qualification factors that may exist between

groups, then evaluating the “leftover” pay difference for

statistical significance

• MR has long been the tool of choice in audit and litigation

settings

• Most employers can readily conduct a MR analysis within just

hours after compiling data

www.BCGinstitute.org

Page 3: Using Multiple Regression to Examine Compensation Practices (Overview)

Why is this Topic Important to HR/EEO Professionals?

• Why do I need to know about this topic? – If you work in a mid- to large-sized employer, it is not uncommon to find

statistical evidence of unfair pay practices

– Federal enforcement agencies have created a lot of attention over fair pay issues, and the stakes are high (for federal contractors especially)

– Compensation issues have a long history in the courts and the tools for evaluating the issues are well established

• What are the key essentials I need to know about this topic? – Conducting a pay analysis does not take long with the correct tools

– Data selection and setup are key

– There are a variety of methods for making corrective pay actions; selecting the best one based on your situation takes serious evaluation

• What are the consequences surrounding these issues? – Being reactive v. proactive can be 10X more expensive

– MR analyses are easy to do incorrectly

– Taking actions to make “corrective pay” based on weak regression (or other analyses) can lead to just as much liability as finding clear pay discrimination evidence

www.BCGinstitute.org

Page 4: Using Multiple Regression to Examine Compensation Practices (Overview)

Presentation Overview

• Overview & Background

• How Does Multiple Regression (MR) Work?

• Steps for Conducting Multiple Regression (MR)

Analysis

• Final Recommendations

www.BCGinstitute.org

Page 5: Using Multiple Regression to Examine Compensation Practices (Overview)

Resources

• Adverse Impact and Test Validation: A Practitioner's Handbook by Daniel A. Biddle, Ph.D.

– Purchase online at www.BCGinstitute.org

• Adverse Impact and Test Validation Book Series Webinars

– Recordings available online for all BCGi Platinum Members

– Webinar slides available online to all BCGi Standard & Platinum

Members

• BCGi Membership

– Free Standard Membership

– Premium Platinum Membership

www.BCGinstitute.org

Page 6: Using Multiple Regression to Examine Compensation Practices (Overview)

About Our Sponsor: Biddle Consulting Group (BCG)

BCG is an HR firm dedicated to providing the highest products and services related to

Equal Employment Opportunity (EEO), Affirmative Action and Employee Selection.

• BCG’s Consulting Services

― Affirmative Action Plan Outsourcing

― Compensation Analysis

― EEO/AA Litigation Support (Plaintiff and Defendant)

― Job Analysis

― Test Development and Test Validation

• BCG’s Software Products

― Adverse Impact Toolkit™

― AutoAAP® affirmative action plan development software

― AutoGOJA® job analysis software

― C4 call center testing software

― COMPARE™ compensation analysis software

― CritiCall® dispatcher/call-taker personnel selection software

― ENCOUNTER soft skills video situational judgment testing software

― OPAC® office skills testing software

― TVAP™ Test Validation & Analysis Program software

www.Biddle.com