Professional Documents
Culture Documents
Job Pricing
Chapter 1: Internal vs. External Equity
Chapter 2: Analyzing Survey Data
The completion of job evaluation does not signal the end of the wage and salary determination
process. The next step is “job pricing”—setting a dollar amount for the job—and it’s not as simple as
it may sound.
Chapter 1: Internal vs. External Equity
What you have as a result of your job evaluation program is a hierarchy of jobs based on (for exam-
ple) point values. In other words, you can be assured of internal equity when it comes to matching up
the various evaluation “scores” with actual wage levels because all of the jobs in the organization have
been compared with each other and have been evaluated accordingly. But internal equity alone can’t
guarantee employee satisfaction or protect your firm from a legal challenge. You must be aware of
what other firms in your area or industry are paying for similar jobs. Once this information has been
obtained and you have determined that your wage and salary structure compares favorably, you have
achieved external equity as well.
The most common means of obtaining this information is through wage and salary surveys.
Although conducting such a survey or even trying to make sense of the hundreds that are available
might seem like an overwhelming and probably unnecessary task to the small or medium-sized
employer, it should be remembered that almost every company does take surveys, even though they
may only do so informally. For example, most employers pay attention to the salaries that job appli-
cants are demanding, to the rates that departing employees say they will be receiving from their new
employers, to the information about wages that is passed around at seminars and business meetings,
and to what is said about wages in the news media. This informal survey activity is going on all the
time, and from the fragments of information that are collected emerge decisions concerning pay
increases, fringe benefits, and overall pay policies.
If your company isn’t in the habit of consulting or conducting wage and salary surveys and yet has
always managed to attract and maintain a sufficient number of qualified employees, you may not see
the need for any kind of formal survey activity. But consider the following: A $1.00 per hour over-
payment in a labor-intensive firm can have a drastic effect on its ability to compete and even on its eco-
nomic survival; while a $1.00 per hour underpayment can lead to higher recruitment costs, the hiring
of less skilled employees, increased training costs, and more turnover—not to mention lower morale
and motivation throughout the organization. In short, you may think you’re surviving just fine with-
out all this concern about what other companies are paying their employees, but you have no way of
knowing just how much your lack of concern is costing your firm in other, less obvious ways.
How Compensation Surveys Are Used
Most employers that do utilize wage and salary surveys on a regular basis have found them to be an
invaluable planning tool. Among other things, wage and salary surveys can be used to:
• Determine where your company’s pay rates for certain jobs or groups of jobs stand in relation
to the labor market.
• Are the companies that have participated in the survey about the same size as our company?
(You don’t want to end up comparing yourself to a few corporate giants.)
• Does the survey cover jobs that are similar to those in our firm?
• Are the survey findings affected by a few particularly high- or low-paying firms, or has the
average company been given sufficient weight?
• Do the participants have formal job evaluation and wage and salary administration programs,
or do they set their rates arbitrarily?
• Have any of our employees left the firm to go to work for other companies that have partici-
pated in the survey?
• Have we hired many employees from these participating companies?
• Does the survey cover at least some firms that we regard as competitors?
Your aim should be to select surveys of competitive companies within a relevant geographic area.
The participants should all have sound wage and salary programs, and survey data should not be undu-
ly influenced by a few large, progressive companies that can afford to pay more.
Professional surveys are conducted regularly by large employers, professional and consulting organi-
zations, trade associations, and the government.
BLR, the publisher of this reference, conducts annual surveys of exempt and nonexempt compensa-
tion, consisting of rates for over 100 benchmark positions in several hundred geographic and demo-
graphic categories. The Survey of Exempt Compensation is published in five regional editions. The Survey
of Nonexempt Compensation is published in separate state editions. The surveys are sent free of charge to
participants and subscribers to BLR’s loose-leaf services, What to Do About Personnel Problems in [Your
State] and Employee Compensation in [Your State]. BLR also conducts an annual survey of employee bene-
fits. For additional information, contact BLR’s Response Center at (800) 727-5257 or go to
www.blr.com.
Conducting Your Own Survey
The other approach, of course, is to design and conduct your own compensation survey. The big
advantage here is that you can pick the companies who will participate in the survey, and thus you
can be more certain of getting the type of data you want. But the collection and analysis of survey
data is an expensive undertaking, and you should be sure that the benefits will outweigh the costs
involved. If a “canned” survey developed by another organization will supply the data you need for
decision-making purposes, there is little to be gained by going to all the trouble yourself—and possibly
much to be lost through your lack of experience.
Most companies who perform their own surveys do so in order to remain competitive within the
local area where they recruit their employees. A compensation survey of the area enables the company
to know what rates of pay the local market demands, and to direct its efforts toward recruiting and
retaining the best possible workers.
While this isn’t the place to get involved in a detailed discussion of survey design and implementa-
tion, mention should be made of the fact that a good compensation survey requires both planning and
expertise. A human resources manager with no experience in conducting such surveys cannot expect to
elicit the needed information without doing a great deal of research and preferably obtaining some out-
side help. It is difficult enough to be sure you’re getting accurate, unbiased information without hav-
ing to worry that you’ve inadvertently undermined your own objectives. Your goal should be to obtain
a true picture of the rewards being offered by all organizations competing in the labor market. A sur-
vey that limits itself to a particular industry, to companies of a certain size, to a sector of the economy
(public or private), or to only a portion of the actual labor market will result in biased data that may
cause compensation problems instead of clarifying or solving them.
There are legal pitfalls involved in collecting survey data yourself. Contacting local firms in your
industry for salary information on specific employee categories may be seen as being anticompetitive.
Ask yourself whether your inquiries are likely to have that effect as they could amount to illegal wage
fixing in violation of antitrust laws. If you do contact other employers, you probably ought to consult
competent legal counsel before making the first call.
Middle 50 percent: Also termed the interquartile range, this measure is the range of data resulting
from discarding the highest 25 percent and lowest 25 percent of the reported data.
For several reasons, the ideal measure for survey comparison is the mean of the middle 50 percent of
reported rates. The exclusion of the high and low quartiles eliminates from the data the trainees and
persons who are grossly overpaid and with whom comparison is not desired anyway; this reduces the
main drawback of the mean—its sensitivity to extremes—and yields the most accurate possible meas-
ure of central tendency for survey analysis.
By way of illustration, consider the following hypothetical array of reported survey data: $340,
$340, $340, $350, $350, $350, $360, $360, $360, $360, $370, $380, $380, $380, $390, $420. The
mean reported rate in this array is $364.38, which is somewhat higher than either the median of $360
or the mode, also $360. This discrepancy between the mean and the median results from the inclusion
of a few very high rates, especially the one at $420. By excluding the highest and lowest 25 percent of
the data, we reduce the range of the array, which was $340 to $420, down to $350 to $380; the mean
of this middle 50 percent range then becomes $361.25, which is considerably closer to the median,
because of the elimination of the extreme rates.
If the mean of the middle 50 percent is not provided, then the mean of the total range of data
should be used. The exception to this would be when the reported mean and median for the same job
are quite different, indicating that a number of extreme rates are skewing the mean; in such case it is
reasonable to use the median.
An additional point: It should be recognized that survey data are often published months after
being collected. In some surveys, for instance, the survey may not be made available until as long as
nine months after the data were collected, and rates of pay in the market will of course have changed in
that time. One way to deal with this lag for the short period is to update the survey data in proportion
to the change in the cost of living, as measured by the Consumer Price Index (CPI). In this way, sur-
vey data several months old may be made usable, as increases in rates of pay tend to be fairly closely
correlated to increases in the CPI.
A final consideration is the possible impact of the firm’s participation in the survey on the reported
statistical measure. The firm obviously desires not to compare itself to itself, but rather to other firms;
therefore, if the firm participated in the survey, its rates should be deleted from the reported survey
data before a comparison is made. This can be done in several ways; a simplified approach is demon-
strated in the following example.
Assume the survey reports a mean weekly salary for 121 clerks of $314; also assume that included in
those 121 clerks are 13 from your firm, for whom a mean rate of $306 was reported at the time the sur-
vey was taken. The mean salary for the companies excluding your firm may be calculated as follows:
121($314) – 13($306) = $315 The total number of employees times their mean, divided by the
108 number of employees reported by the other companies.
In the example given, the mean of the rates reported by companies not including your firm would
thus be $315. Where the number of employees reported by a particular firm is a small proportion of
the total, the impact of the firm’s rates will be minimal. Nevertheless, this potential impact should be
considered and dealt with if necessary.
From this summary you can prepare a “scatter chart” or “scattergram” that shows job evaluation
points and labor grades along the bottom axis and survey dollar values on the vertical axis. Dots repre-
senting various benchmark jobs are plotted on the chart by finding the proper number of evaluation
points along the bottom or x-axis and then moving vertically up the y-axis to the appropriate survey
wage. Each dot, as can be seen in the scattergram below, can be said to represent the relationship
between what the company feels the job is worth internally (job evaluation points) and what is being
paid for similar work in the labor market (survey dollar values).
Scattergram 1
Survey
Dollar
Values
575
575
550
550
525
525
500
500
475
475
450
450
425
425
400
400
375
375
350
350
Points 200
200 400
400 600
600 800
800 1000
1000 1200
1200 1400
1400
Grade 21 22 23 24 25 26 27 28 29
It can be observed from the scattergram above that the higher the internal job evaluation points are,
the higher the surveyed dollar rate tends to be. Such a relationship is to be expected and is a means of
verifying the accuracy of the evaluation process. If the points got higher as the survey dollar values got
lower, this would certainly be an indication that something was wrong. Even more clearly than the
summary table, however, the scattergram demonstrates that as the evaluation points rise, so do the dol-
lar values in the labor market, but with some variance.
The next objective is to find the one, single straight line that when drawn through the dots will be
the average of all the data. This form of analysis can best be done using a statistical procedure known
as a “least-squares conversion.” The exact steps to be followed are described in most books on statis-
tics. However, a high degree of accuracy can be obtained simply by looking at the data and fitting a
line through the dots carefully with a ruler, as shown in the following scattergram.
Scattergram 2
Survey
Dollar
Values
575
575
550
550
525
525
500
500
475
475
450
450
425
425
400
400
375
375
350
350
Points 200
200 400
400 600
600 800
800 1000
1000 1200
1200 1400
1400
Grade 21 22 23 24 25 26 27 28 29
This line represents the average relationship between internal job worth as measured in job evalua-
tion points, and external job worth as measured by labor market surveys. In the jargon of wage and
salary administration, it is called the “community wage curve.” When an employer states that it pays
wages “equal to or better than the going rate in the community,” this line represents that promise. To
be more precise, if the organization wants to pay its employees the going rate, it will use this line,
which represents the average pay in other firms for similar work, to determine how much its “average”
employees will be paid.
Scattergram 3 shows how this is done. A line is drawn horizontally at the point where each labor
grade intersects the community wage curve. This line will become the midpoint of the rate range for
each grade. Now for some definitions:
Midpoint is the going rate in the salary survey sample and the rate that will be established for the
employee performing 100 percent of the job duties at 100 percent efficiency under normal supervision.
It should be halfway between the minimum and the maximum.
Minimum rate is the level of pay to which an employee who meets the minimum qualifications for
the job is entitled (depending on company policy, this may mean the ability to perform 75 percent of
the job duties at 75 percent efficiency, under normal supervision).
Maximum rate is the highest rate to be paid for work in the labor grade. Normally this would be
paid only to a person who performs duties well beyond those called for in the job description, at the
highest possible efficiency under little supervision.
Scattergram 3
Survey
Dollar
Values
575
550
525
500
475
450
425
400
400375
400 600
600 800
800 1000
1000 1200
1200 1400
1400
350
Grade 20 21 22 23 24 25 26 27 28 29
Spread is the distance between the minimum rate and the maximum rate, expressed as a percentage
of the midpoint. In other words,
If you want a 30 percent spread for a grade with a midpoint of $400, then .30 x $400 = $120, thus
the minimum rate will be $340 and the maximum rate $460. There is no rule as to how great the
spread must be; however, these guidelines may be of help: First, there is probably no job that requires a
spread of more than 50 percent. Second, as a starting point, use this formula as a guide to setting the
spreads: Convert the midpoint to a weekly rate, add 100 to it, divide by 1,000, and multiply that
amount times the midpoint to get the spread. For example, with a grade that has a midpoint of $360,
add 100 to it ($460), divide by 1,000 (.46), and multiply that times the midpoint (.46 x $360 = $166).
This amount is the distance between the minimum and the maximum, resulting in a rate range that
looks like this:
Minimum Midpoint Maximum
$277 $360 $443
At first this may appear to be a complicated procedure, but it gets easier with experience and it
works well.
Taking the midpoints from each of our 10 labor grades in Scattergram 3, we can apply this approach
to create a rate structure:
Grade Minimum Midpoint Maximum
20 $303 $357 $411
21 $323 $380 $437
22 $343 $403 $463
23 $362 $426 $490
24 $382 $449 $516
25 $401 $472 $543
26 $421 $495 $569
27 $440 $518 $596
28 $460 $541 $622
29 $479 $564 $649
This rate structure can also be expressed in chart form as shown in the following Rate Structure
Chart, with the bottom of each column representing the minimum rate for the grade, the top of each
representing the maximum, and midpoint represented by the line dividing the shaded sections:
650
600
550
500
450
400
350
300
Grade Grade Grade Grade Grade Grade Grade Grade Grade Grade
20 21 22 23 24 25 26 27 28 29
As you can see, the spread increases for each labor grade, making a longer “box” on the chart. This
greater distance between the minimum and the maximum is necessary because promotions come less
often to persons in higher grades, and a wider spread is needed to offer these individuals incentives over
a longer period of time. Also, there is a wider range of performance exhibited in the upper grades; to
illustrate, the difference between the worst performance of a clerk and the best performance is not real-
ly all that great but the difference between the best and worst executive secretary is significant, and the
rewards for varying levels of performance must be built into the rate structure.