Without first deciding what data are needed and why, embarking on a job analysis process can be a highly inefficient and largely wasted effort. Some novice analysts wind up asking numerous questions during interviews that generate answers that are not needed. Management and analysts must have their direction--their reason--for the data-gathering and job description preparation effort clearly in mind and must plan and structure this effort for high yield. Considerable preliminary work must be done before data generating starts. Otherwise it is like scatter bombing. You do not know where the target is so you spray the area, hoping you will hit it. You may hit it but is it worth the cost?
Subjects cannot remember everything. Techniques, such as questionnaires, interviews, and even work logs, rely on employee memory and ability to mentally capture and communicate the full scope and depth of the job. But employees, even though they do the job--and do it well--often cannot think through and recall all they do when sitting down in an interview or when filling out a questionnaire, for example.
Frequently, tasks are performed but once a year. One is likely to forget to record or report these. Other, perhaps low-priority but still time-consuming, tasks are often disregarded in reporting by incumbents. It is necessary to allow plenty of time for incumbents to think about their jobs. Analysts must be skilled in stimulating recall with appropriate questions on questionnaires and in inter views, and must have workers report at the proper intervals with work logs.
Person specifications and job characteristic data confusingly mixed with data on time usage. Job analyses too frequently do not distinguish job or task dimensions from person specifications and job characteristics. Task dimensions take up time; person specs and job characteristics do not. Job analysis is often negligent in not focusing on how workers spend their time. This is the key type of information desired for preparing job descriptions, but frequently analyses slip into excessive emphasis on specs and characteristics. It is valuable to know, for example, that the worker needs such and such skills to do the job and that the job is, perhaps rather routine or dangerous at times, but these are specs and characteristics.
Formal job analyses are incomplete. Too many companies try to write job descriptions directly, from scratch--without really gathering any data-or from data acquired informally. In a recent study the author found that 85 percent of the organizations surveyed had job descriptions prepared for at least some of their employees, but only 10 percent had conducted any rigorous data gathering for job description preparation. The common approach was either to have employees directly write their own job descriptions or to have managers do it for their employees. Such an approach usually produces job descriptions in a rather short time with little pain, but the job descriptions produced are generally far from valid. A well-done formal data-gathering effort tends to succeed at a much higher rate because it is planned and thought through.
Possible flaws are anticipated in advance and avoided. Redundancy procedures can be built in to assure acquisition of accurate data. Higher authority not involved Job analyses perhaps more often than not lack the blessings, guidance, and overt support of higher management echelons.
Often employees are not made aware of the high priority of the job analysis as an organizational engagement. Management fails to communicate this to them. Often, also, employees see higher management avoiding subjecting themselves to job analysis efforts. Under such conditions it is difficult for employees to take job analysis seriously-as anything but a bureaucratic tack-on simply invented to maintain bureaucratic muscle and control over employee behavior.
Crisis data acquisition. Data gathering is not infrequently commonly done in a crisis. Another common scenario is for the organization to be moving along without much thought given to the desirability of job descriptions. All of a sudden someone suggests the need for some job descriptions. Management then rushes around trying quickly to get some together. The process is not properly coordinated. The types of data really needed are not identified. In short, this crisis approach produces a third-rate job description. Management then attempts to use the developed job descriptions, but finds after a while that the job descriptions are not good enough to use. They may finally abandon them and file them away for the long term. At some later date someone suggests the need for job descriptions again and the process starts all over. Does this sound familiar? This is the crisis approach to job description preparation--no planned, regular, continuous effort.
Too much pirating of data. When people are asked to prepare job descriptions apart from any formal attempt to gather data, they often look for the path of least resistance (PLR). The PLR is often to steal from other sources. Employees (supervisors and subordinates) often dig out old job descriptions and essentially copy them. Sometimes they get job descriptions that have been prepared for other workers doing similar jobs in the company or outside or take data from standardized job descriptions prepared by the industry or an outside professional group. Sometimes they try to piece together job descriptions from various other company records and documents. All these sources can help, but help must be distinguished from substitution. When this type of simple approach is used, what an organization is left with are empty, fictitious job descriptions.
Lack of standardized data gathering procedures. When the interview is used for one subject, a questionnaire is used with another subject, and, say, direct observation is used with a third, you have a problem with data gathering. The data derived will defy attempts at comparison. It will be impossible to put together three job descriptions that provide accurate comparison of the jobs. Similarly, when one subject is allowed thirty minutes for filling out a questionnaire and another has two weeks, you have a situation that prevents valid comparison of jobs on the basis of the data collected.
Some companies are notorious for uncoordinated data gathering across departments. Production, for example, uses one type of questionnaire; the marketing department uses another type. It even gets worse, for example, when a work log sheet with predetermined categories of activities is used in one department, and a work log that leaves activity definition completely up to the incumbent is used in another department. The only way data will truly be useful company-wide is for data to be generated systematically with standardized means across employees and across departments.
Lack of help during the process. Seldom do data providers get enough advice and guidance during the process. Quality orientation for data providers is necessary but not sufficient. Incumbents are going to need help answering questions, doing logs, and performing all the other steps in the process.
Those conducting the job analysis need to be readily available to provide help upon request. Indeed, they need to intervene actively--to solicit questions and comments from respondents--and to search continually for ways to ease the process along. Proper intervention during the process can motivate constructive incumbent contributions too. Workers can come to sense the great importance of the analysis and how it can be of real benefit to them and the company.
Data gathered by novices. Job analyses are too commonly conducted by novices. Companies have a habit of delegating job analysis work to newcomers who lack expertise and experience in the process. Perhaps this is because it is a good way to give the new person a wealth of valuable experience fast. Perhaps it is because many people consider job analysis to be a rather routine endeavor, which involves little complexity or heavy decision making. Whatever the reason, it is usually not enough to support amateurism. Job description data are so vital that only those who have developed a high skill at gathering it should gather it. It may be a fundamentally basic task, but it has to be done right or the organization will eventually suffer internally.
Performance data gathered during the job analysis. One does not wish to answer how well workers are performing. This determination is left for another time--during performance evaluations and reviews. Performance data are vital but have nothing to do with data about the design of the job.
Job analyses that go astray and deal with performance issues are missing the point. They will never generate the data needed for a job description. You can ask what performance criteria and what performance standards exist. This can help clarify job design. But going beyond this and attempting to evaluate incumbent performance is outside the scope of the job analysis. It should not be mixed with it.
Employees find it difficult to describe what they do. Job analysis efforts are frequently hampered by employees' inability to state in words what they really do. Interviews, questionnaires, and work logs all face this problem. Many fine employees struggle when asked to state what they do and to categorize into major areas of responsibility what they do. They can do the job but cannot tell you about it. To overcome this problem takes skillful interviewing, carefully worded and structured questionnaires, and work logs that minimize the need for written phrases by the respondent by requiring check marks instead of written phrases.
See the following articles for more information: