The problem with the way compensation professionals look at salary survey data is that they focus solely on the averages or middle of the range, according to James R. Beatty, professor at San Diego State University, and Fred Whittlesey, chief compensation officer, Payscale, Inc., at their presentation during World at Work's 2006 Total Rewards Conference. However, said Beatty, if we put our feet in the fire and our heads in the freezer, we might on average have a comfortable temperature but that doesn't mean we are comfortable.

Both men agreed that there is a new survey data paradigm that requires compensation professionals to understand the variations in pay and not simply to look at the averages or the middle. According to Whittlesey, salary survey data has been collected and reported the same way for 20 or 30 years. However, this no longer meets the needs of business or reflects the reality of the marketplace.

Whittlesey noted that compensation professionals have been unable to explain significant variations in pay because they don't collect data on all of the variables that affect pay. The job itself (the variable typically used to collect and report salary data) is only one variable. Other variables that must be considered are specific to individuals including qualifications, experience, specialized knowledge, education including where the individual received an education, etc. According to Whittlesey, if we identify the variables that impact pay, we can be far more precise in determining the appropriate pay range for an individual.

The new paradigm requires the collection of data that encompasses all of the variables, asserted Whittlesey. Employers need multi-source databases that are updated in real time. He also states that the data will have to be submitted by employees as well as employers. Moreover, the larger the data base the more valid the data.

As a statistician, Beard noted that it is important to look at the actual data points reported in survey data and not just the averages. If there is significant variation in pay between individuals, compensation professionals need to understand why? How much do individual pay rates vary? How can the variations be explained?

Whittlesey summarized by saying that the future of "competitive intelligence for pay" incorporates the following:

• Multi-source databases updated in real time
• Employer-sponsored employee submission of data
• Continual validation by requerying of submitters
• Integrated job description, tasks, scope, and individual databases
• Powerful statistical models with user-friendly interfaces
• Participative decision-making in tactical compensation decisions