The OFCCP wants more data. The agency’s current proposal for all of this additional data has been described by some as both “burdensome” and “stealthy”. McGuire Woods, a 900-person law firm based in Richmond, Virginia, went so far as to say that the OFCCP “does not understand the private sector or have any apparent concern about the burdens and confidentiality issues these proposals place on contractors.” In fact, the OFCCP itself estimated it would take 103.2 hours and cost $135,000 to collect and provide all the data that could be requested in the revised Scheduling Letter.
The proposed changes to the scheduling letter would increase the number of data submission items from 11 to 13. The first major change to the itemized listing is with respect to support data. Item 8 is an entirely new requirement, asking for the establishment’s leave policies regarding the Family Medical Leave Act, pregnancy leave, and religious accommodations. If these policies are included in the company handbook, the OFCCP requests the handbook be submitted. Because you’ll be required to submit handbooks and all associated policies, it’s critical that these items are kept up-to-date.
The next major change is that contractors will be required to provide data on employment activity for applicants, hires, promotions, and terminations by job group, job title, and by gender and specific racial and ethnic groups. There are also some key changes to the analysis pool for each kind of employment activity.
With respect to hiring, contractors would be required to submit the count of those applicants who did not disclose their race or gender. With respect to promotions, contractors would be required to submit the actual pool of candidates considered or who applied for the promotion by job group and by job title. As for terminations, contractors would be required to submit the actual pool of candidates considered for termination by job group and by job title and must specify if the termination was voluntary or involuntary.
These proposed changes have some pretty big implications in terms of data collection. Since you’ll be required to submit counts of applicants who don’t self-identify their race or gender, you should create new race and gender categories for these applicants. I suggest “non-ID” gender and “non-ID” race. I recommend using these categories, rather than leaving gender or race blank. Using the “non-ID” categories will allow you to tabulate your counts by job group and job title easily, and can help you keep your candidate demographics data clean. One of the key things to look for in cleaning your applicant data is unidentified gender, race, or other protected status. Blank values should be an indicator that something is missing. If the candidate did not self-identify, race and gender is NOT missing – it’s not identified. There’s a difference. If you record this difference, you’ll save yourself some time and effort in terms of data cleaning. You won’t have to go back and verify the demographics for the non-identified candidates, and any blank values you do find will tell you that you need to do follow-up on those records.
One other thing that can streamline your data collection and cleaning is to use a consistent set of demographic values. Code your gender as either “F”, “M”, or “non-ID”. Code your race and ethnicity as “W”, “B”, “H”, “A” and so forth. Ideally, you should avoid any text entries. I can’t tell you how many times I’ve seen client data that has a mix of race and ethnicity entries, such as African Americans being coded as Black, AA, AfrAm, African American, and so forth. Each variant of the same category in your data – even in terms of capitalization or spacing – makes an analysis difficult, because all of the variants have to be standardized to a common value before the analysis can proceed. It’s a simple thing to set these categories up with consistent values, and it can really save you a lot of time in terms of data cleaning.
This is true not only for demographic characteristics, but for all of your data. Any variants on job titles, such as coding managers as managers, mgr, mngr, and so forth, can increase the time required to clean your data exponentially. Develop a fixed set of values for each data element, and stick to that fixed set. It seems like a pretty simple thing, but you’d be surprised how many organizations don’t do this. And it can really save you a lot of time and effort in terms of data clean up.
Moving on to promotions, you’re going to need a way to group promotion applicants and candidates by vacancy. The easiest way to do this is to use a system similar to your applicant tracking system. Your ATS should be grouping candidates by requisition number or some other unique identifier for each vacancy. Your promotion tracking system should be doing the same thing.
If you’re using a paper-based system for candidate tracking, now might be a good time to look into an electronic system. And I’m talking about more than just an Excel spreadsheet – I mean a real applicant tracking system. There are several cost-effective options that are scalable to your needs, are easy to use, and easy to implement.
But getting back to promotions, you’re going to need some kind of CTS – a “candidate tracking system” – that can track promotion candidates by vacancy. I would recommend setting up the system to differentiate between “job progression” events and promotion events. A job progression event is where an employee is promoted from, say, Analyst I to Analyst II, as a result of achieving a certain time in job, passing a qualifications test, etc. If these job progression movements aren’t competitive, you’re essentially going to have a candidate pool of one. To easily distinguish these movements from other competitive promotions within your organization, you should have an indicator in your CTS that tells you this is a job progression movement and not a competitive promotion.
One other complication with promotions is that some organizations use a formal bidding procedure, some use a tap-on-the-shoulder process, and some use a combination of both. You’re going to have to provide data on all of your promotion events, regardless of the selection procedure. Consult with legal counsel to discuss your promotion process, and if you don’t have a formal bidding system, talk with legal counsel about whether you should have a formal process in place.
One final thing to keep in mind with respect to your promotions data is that any information used in making promotion decisions should be kept within the Candidate Tracking System. For example, if performance ratings are used to evaluate promotion candidates, that information needs to be maintained as part of the candidate’s record in the CTS. Similarly, if educational attainment, seniority, time in job, skills and qualifications, or other factors are considered in the promotion decision, that information should be maintained as part of the candidate’s record. Ideally, you should be able to re-create the promotion selection process from your CTS. This means not only keeping information on the candidates, but promotion advertisements, how candidates will be evaluated, how much weight will be given to each of the criteria, and so forth.
Your data collection process for terminations should be very similar to that for promotions. You’re going to need to identify all employees who were eligible for termination, as well as the metrics used to select an employee for termination. Just like the promotion data, you should be able to recreate your termination decisions from your termination pools data. And don’t forget to have a flag in your data set that indicates whether the termination was voluntary or involuntary. Everything I’ve said about hiring and promotion in terms of data collection and maintenance holds true for termination selections as well.
Let’s now move from selection decisions to compensation decisions. Under the proposed scheduling letter, contractors would be required to submit compensation data for all employees “including but not limited to” full time, part time, contract, per diem, day laborers, and temporary workers.
Employee compensation is defined as the base salary, wage rate, and hours worked. Additional compensation or adjustments should be identified separately and includes bonuses, incentives, commissions, merit increases, geographic differentials, and overtime. The OFCCP is encouraging contractors to submit any other factors used to determine pay including education, past experience, duty location, department, and function. Contractors will also be required to submit organizational policies to reasonably explain compensation practices.
Not only will contractors will be required to produce this additional data, the OFCCP is mandating that the data be provided electronically for each individual employee. The agency is going to be reviewing each employee’s compensation at the individual employee level. Currently, compensation is reviewed in the aggregate, and contractors are required to submit annualized total compensation data by salary range, rate, grade or level, grouped by race and ethnicity, and by gender.
Most contractors won’t be accustomed to producing compensation data electronically at the employee level, so you may need to take some extra steps here to make sure that your compensation data is ready for production. It’s going to be very important that your compensation data is as comprehensive as possible. Each employee’s pay record should include all the information necessary to determine that employee’s status – full time, part time, temporary, contract, per deim, or day laborer. It should also include all information relating to how that employee is paid – whether he receives commissions, is eligible for bonuses, receives a shift differential or locality adjustment, and so forth. Any adjustments – such as red circling – should be clearly indicated on the employee’s compensation record, and supporting documentation regarding those adjustments should be provided to the OFCCP.
Before producing your compensation data, make sure it’s clean. Check to make sure that you’re reporting compensation consistently within salary range, grade or level. You want to avoid a mix of hourly rates and salary amounts within the same grouping. Review your data to make sure it passes the “smell test” – look to see if you have unusual hourly rates – rates under minimum wage, rates that are too high to possibly be correct hourly rates, salary amounts that are too low to be correct salaries, and other things that are clearly out of line. Take a look at the hours worked data. Red flags indicating potential errors include excessive amounts of overtime for an employee within a given grouping, no overtime hours for a given employee when others in the grouping have overtime hours, or unusually high or low regular hours worked.
Cleaning your data can seem like an overwhelming task. And it can be, if you don’t have good data collection protocols in place. Even if you do have good collection protocols, errors can still happen. Take a look at your data and follow up on anything that seems out of line. And remember, there are resources available to assist you with your data clean up and production. A fresh set of eyes can often identify errors that you yourself may not have caught.
When it comes to data production, the most important thing is to get it right the first time. Make the extra effort to ensure that what you produce is clean and correct. You don’t want to be in a situation where you need to do one supplemental production after another because of errors identified after the original production. Take the time to look at your data, note anything that looks out of line, follow up and find the explanation – whether it’s an error or has a legitimate explanation – before you send that information to the OFCCP.