Magazine article Personnel

Beyond "Data Validity": Improving the Quality of HRIS Data

Magazine article Personnel

Beyond "Data Validity": Improving the Quality of HRIS Data

Article excerpt

Beyond "Data Validity": Improving the Quality of HRIS Data

The quality of data in human resources information system (HRIS) databases is in some respects almost a philosophical issue. That is, the "quality" of the information about people and jobs that resides in the HRIS frequently lies "in the eye of the beholder" and can be called "good" or "bad" only in the context of current business requirements, a given application of HRIS data or processes, or an individual HR manager's perception of "what's right" in terms of data accuracy, timeliness, or significance of the information conveyed.

Consider, for example, some of the situations that have been known to inspire a major new HRIS data-quality improvement program: . A person's name is carried as J. T. Smith in one place, John T. Smith in another, and Jack Smith in yet another. All are technically "true," but which is correct? . Functional applications run on microcomputer (such as a salary-planning system) are downloaded from a centrally maintained database at the end of each month. The central or master files are updated with changes and new information on the first of each month, so the PC application is never currently "accurate." . Different systems or functions define the same data element differently. To HR employees, for example, "base pay" usually means a person's basic hourly, monthly, or annual pay; to payroll employees, base pay includes premium pay for overtime; and to benefits employees, base pay may include bonuses used to figure life insurance or long-term disability insurance. . The level of accuracy of certain data elements--for example, whether specific data are 92% or 99.9% accurate--may never have been worth the time and cost of data-quality efforts. New business requirements or government regulations can suddenly make this "bad data," whereas last year nobody may have noticed or cared how accurate it was. . Programs and procedures used to extract data that are correct in the database distort the data or render them useless to the user. An example of this is seen when age calculations based on date of birth fail to account for leap years or 31-day months.

In these and similar situations, the quality of the existing data in the HRIS is not so much a matter of validity (that is, of whether the data can pass edits and be entered into the database). Validity, in any case, is an essentially artificial systems concept: Data are valid if they pass certain systematic edits or if they match up with other listings that a user has created that he or she has established as "valid."

Assuming the Users' Perspective

Systems edits and validity checks (see box on p. 55) are essential to ensuring data quality in an HRIS, but they are often just the beginning of a continuing data-integrity program that will ultimately address the real needs of users for meaningful, up-to-date, accurate information. This occurs because HR information deals with human behavior--a subject that is almost infinitely variable--and is used by a range of different organizational functions with different perceptions of "what's right" in data definitions, timing requirements, and other characteristics of information.

In order to establish a continuing database-integrity program that goes beyond "data validity" in providing corporate management and functional users with quality data, HRIS managers must stand in the shoes of those who use the data in the HRIS. This approach is an extension of the original needs analysis that preceded systems development or the acquisition of a packaged system. The focus is on user requirements for data to meet organizational and functional business requirements instead of on some theoretical "ivory-tower" concept of the absolute truth or falsity of data.

Consider, for example, the previously mentioned controversy over Jack Smith's name. An "absolute-truth" approach might be to use the name as it appears on his birth certificate. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.