What is the difference between predictive and criterion data and how is each used?

What will be an ideal response?

Predictive data is information about measures used to make projections about outcomes. For example, data collected to predict turnover or job performance is predictive data. Similarly, one could measure the intelligence of job candidates to see if it predicts some component of job success. This is also predictive data. In terms of the general staffing system, predictive data can come from any part of the hiring process and can include information on sourcing quality, the basic qualifications of applicants, and their traits, competencies, and values.

Criterion data is information about important outcomes of the staffing process. Traditionally, this data includes measures of the job success of employees. Criterion data can include job performance, training outcomes, promotability ratings, days missed, safe workdays, and any other job outcome of interest. At a general level, criterion data can also include all outcome information relevant to the evaluation of the effectiveness of the staffing system against its goals. This can include measuring a company's return on investment related to its staffing measures, employee job success, time-to-hire, time-to-productivity, promotion rates, turnover rates, and new hire fit with company values.

Criterion data is used to measure and assess desired outcomes of interest. Predictive data is used to make predictions about those outcomes. If statistically and practically significant relationships are found between predictive and criterion data, then the predictors can and should be used to make staffing decisions. This type of data-based decision making will provide a competitive advantage to organizations over those that do not use data to make staffing decisions.