HUMAN FACTORS AUDIT:DESIGN REQUIREMENTS FOR AUDIT SYSTEMS

DESIGN REQUIREMENTS FOR AUDIT SYSTEMS

Returning to the definition of an audit, the emphasis is on checking, acceptable policies, and consis- tency. The aim is to provide a fair representation of the business for use by third parties. A typical audit by a certified public accountant would comprise the following steps (adapted from Koli 1994):

1. Diagnostic investigation: description of the business and highlighting of areas requiring in- creased care and high risk

2. Test for transaction: trace samples of transactions grouped by major area and evaluate

3. Test of balances: analyze content

4. Formation of opinion: communicate judgment in an audit report

Such a procedure can also form a logical basis for human factors audits. The first step chooses the areas of study, the second samples the system, the third analyzes these samples, and the final step produces an audit report. These define the broad issues in human factors audit design:

1. How to sample the system: how many samples and how these are distributed across the system

2. What to sample: specific factors to be measured, from biomechanical to organizational

3. How to evaluate the sample: what standards, good practices, or ergonomic principles to use for comparison

4. How to communicate the results: techniques for summarizing the findings, how far separate findings can be combined

A suitable audit system needs to address all of these issues (see Section 3), but some overriding design requirements must first be specified.

2.1. Breadth, Depth, and Application Time

Ideally, an audit system would be broad enough to cover any task in any industry, would provide highly detailed analysis and recommendations, and would be applied rapidly. Unfortunately, the three variables of breadth, depth, and application time are likely to trade off in a practical system. Thus a thermal audit (Parsons 1992) sacrifices breadth to provide considerable depth based on the heat balance equation but requires measurement of seven variables. Some can be obtained rapidly (air temperature, relative humidity), but some take longer (clothing insulation value, metabolic rate). Conversely, structured interviews with participants in an ergonomics program (Drury 1990a) can be broad and rapid but quite deficient in depth.

At the level of audit instruments such as questionnaires or checklists, there are comprehensive surveys such as the position analysis questionnaire (McCormick 1979), the Arbeitswissenschaftliche Erhebungsverfahren zur Ta¨tikgkeitsanalyse (AET) (Rohmert and Landau 1989), which takes two to three hours to complete, or the simpler work analysis checklist (Pulat 1992). Alternatively, there are

simple single-page checklists such as the ergonomics—working position—sitting Checklist (SHARE 1990), which can be completed in a few minutes.

Analysis and reporting can range in depth from merely tabulating the number of ergonomic standards violated to expert systems that provide prescriptive interventions (Ayoub and Mital 1989).

Most methodologies fall between the various extremes given above, but the goal of an audit system with an optimum trade-off between breadth, depth and time is probably not realizable. A better practical course would be to select several instruments and use them together to provide the specific breadth and depth required for a particular application.