Copyright
Leo Charles Pigage.

Job evaluation (Volume BEBR Faculty Working Paper v.5, no.3) online

. (page 3 of 4)
Online LibraryLeo Charles PigageJob evaluation (Volume BEBR Faculty Working Paper v.5, no.3) → online text (page 3 of 4)
Font size
QR-code for this ebook


■H Ti U ^


Cd 4< tJ cu
0) -rl d d P








3 rH


:3 d








C5< 0)


o< -H Tb -d


a in cd -H -H








<U ^^


oj ^H d d


Cd rH !-, 3








K


K bD cd cd


CO > CP bD en



















c




^.^


^






._


,— X


CO


H






^~"


H UA


J- j-


cv^ Ccj






o


■ — "












q:










<D


s


■£








a> a>
£5


C


O)






o


0}

(U
Q.


E






o
o




g










ro


\




Lt-















O w




+J 0)








60 a)




c a


0)


-H


u


-d -P


o


U 3


H


O JQ




u


c


O tfl


•H


cd -d




O


Ul


^a


60


(0 +J


C


hO <U


•H


c a


4->




m


+J M


crt


CO C








o -d


Ti


c


c




•H


l+H t,


U


O 60


60



<M -d 60 t-c

o a; c 0)



Vh -H (d -H



ocO g

c c



1 n



■H <U a 0) rH

=H Sh -H t< 0)

(fl 0) (U

iiH d 60 x: ^

O <U oJ 3 3

(1)

o; ? -d -P 4-1

w +j a c o



,c: to 0) u 'H



a -H cs +j c



TJ cd -P o tJ
oj o c j<; a



■H -H pq oj
•d "d =M ci(

C H Cm tn O



r-l Cm m c cd

•H O O -H



^ O -H ^ ^

3 P P Td Cm
G cd li C O



,Q 0) -H ti o

td > P 3

!h -H cd 60 O

U +J ,-1 C 4J

■d 'H 3 'H

•H +j p< rd -d

CO (U -H C -H

C Pi C -H P)

O (D cd ti Cd



in



>%

Q



•H +> -P

3 Cd a



■M


(1>


11)


Ti


■P
crt


^


0)




m


u




rt


•H




•H


3


to




n<


u


CI


t1


^


■P




P




>-,




(1)


tl


-^


,!<i


n1




crt


U


a


•p



p


M


3


cd


cd


H


t.


>






n


(11


-rt


H


U


o


(U


cd



o
o

o
o

(P <





+j


Ch


CH


cd


o


C






0)




(II


n


C


s




cd


p



(1) OJ
Oh ? C
(U P> (D
^ <D J3

-d -" '




(1) CO ^


60



P +J ? c



O *-



00



c



O

Q.



o
Q



sure that comparable data are being received. For example, wage rates
can be influenced by:

( 1 ) Average hourly earnings.

(2) Incentive bonuses or premiums.

(3) Nominal operating hours.

(4) Inclusion of overtime pay.

(5) Relationship of guaranteed day-w^ork rate and incentive plan
used.

(6) Number of persons in each job in each plant surveyed.

(7) Special benefits and privileges.

The above list, though not all-inclusive, does suggest the various diflfer-
ences that should be looked for.

All data should then be compiled and any data not comparable should
be eliminated from further use. Copies should then be supplied to plants
cooperating in the w^age survey.

Wage Curve

Even after eliminating irregular wage data collected during the wage
survey, one will soon discover that plotting the wage curve presents
some problems.

The ultimate purpose of the wage curve is to arrive at the relative
money rates which are to be paid for all the jobs which have been
evaluated. These rates should reflect, in general, the relative status of the
jobs evaluated.

The first step is to plot, on graph paper of convenient size, the relative
evaluations of the jobs along one ordinate, and the money rates along
another ordinate. See Figure VI. All usable wage data from the wage
survey, and rates within the plant, should be shown. (It is desirable to
key the data so that the sources of information from any particular plant
can be readily identified.) The resulting plot will show that no trend line
will be possible which will include all the points plotted. Rather, there
will be range trends, as shown by the dotted lines in Figure VI.

There are two ways in which trend lines can be established: mathe-
matically, by what is known as the least-squares method, or by estimating.

The method of least squares is defined and described in any text on
statistics. This method is sound mathematically. But it is not practical
here because the original data are based on judgment and not on abso-
lute facts. It is foolish to apply a detailed mathematical procedure to
data which are not absolute in the engineering sense.

A second reason why this method is undesirable is that the trend line



30



FIGURE Yl



o
q:



3
O

X

a>
v>
o

CO



.80]

.60

.40

.20

.00

.80

.60



.40



-r



c
o

-3



0>

a.
o



Note:



Total
Data from Chart I.



/Trend Lines



^



o Own company
• Company A
X Company B
a Company C



Points
Illustrative only.



is not necessarily straight, and since the method of applying the least
squares is easiest for a straight line, people tend to force a straight line
on data which should yield a curved line. Then, to establish labor grades,
the straight line is broken up — "roughed up" — into steps. What these
objections all add up to is that the method of least squares gives a false
sense of exactness. It looks scientific, but isn't.

Another matter to settle is the general shape of the trend lines.
Should they be straight lines or curved lines? There is no answer to this
question except as indicated by the data. It is reasonable to assume that
if a wide range of skills is being covered by the one evaluation program
and if the original measuring scales are arithmetic scales (for example,
5, 10, 15, etc.), the trend lines will be curved, concave upward, when
the data are plotted on ordinary arithmetic graph paper. If, however, the
original measuring scales used in the evaluation are geometric (for ex-
ample, 10, 20, 40, 80, etc.) the trend line will be a straight line. The
narrower the range of skills to be covered by the evaluation program, the
more nearly straight the trend lines.

Once the trend lines indicate the range of rates which prevail for the
various jobs, the next consideration is to settle where within the range
the particular plant can operate. There are two extremes: (1) at the
bottom of the range throughout, and (2) at the top of the range
throughout. Other alternatives are any number of places between the
range lines. Each company must set its own policy in this regard, on the



31



basis of the product and labor competitive markets and with due regard
to collective bargaining obligations if such exist.

If a single rate structure is the final aim for each labor grade, the
establishment of the average trend line will permit going directly to
establishing labor grades. However, if rate ranges are to be used for each
labor grade the trend lines will show one of the following tendencies of
the range between the lines: constant money range, constant percentage
range, or variable money and percentage range. Here, again, a policy
needs to be determined. Usually, the wider the range of skills covered
by any one job evaluation program, the more the trend will be toward
a variable percentage range. This is necessary to permit more variability
in rates for higher-skilled jobs. Furthermore, the use of rate ranges for
each job requires a policy on how the progression is to be made from
the lowest to the top rate for the job. Is this progression to be according
to seniority, merit, or a combination of seniority and merit?

Labor Grades

In the previous section. Wage Curve, a trend line was determined,
which relates the relative evaluation of the jobs with a money rate to be
paid. However, one will quickly discover that a great number of wage
rates will be present if the study stops with the wage trend line. Further-
more, since the whole system up to this point is based on judgment, even
though it is perhaps sound, many money rates will be a fraction of a cent
difTerent for very similar jobs. Such a fractional cent variance is nearly
impossible to explain and justify to those expected to accept and work
with the results of a job evaluation system. Also, to keep a record of so
many difTerent rates becomes an increasingly diflficult accounting problem
for management as well as a union.

To overcome the psychological aspect of small fractional cent variance
for similar jobs, jobs are grouped into what is known as labor grades.
This grouping of jobs into labor grades reduces the friction which would
result with small fractional cent variances because all jobs in any one
group, or grade, will carry the same rate or rate range.

It is desirable that a few basic rules be observed in establishing labor
grades. They are: (1) the number of grades should be established to
provide sufficient incentive for movement from grade to grade; (2) the
overlap between labor grades, if rate ranges are used, should not be
excessive; and (3) a rational pattern of difTerentials should exist through-
out the labor grades.

Considering the first basic thought, the number of labor grades estab-
lished, one finds that there is a workable minimum as well as a maximum
established through practice. The minimum condition is to have labor

32









FIGUR


E ML








1.39-
















a>








11 iyiilliyilll::!:;:













I: ::::::::::;::::::;::




o








il l|llillli:l:l::l:::j




% 1.23-








1: i:!;:;::;;!;ll;;i:i!














^








ii;:i




3












^ 1.12
















« 1.05-
o 1.00-




















.'" '"












m





















1


iiiiiiii


lllliil


4


iiiiSi





Labor Grades

grades with at least five or six percent of the base rate differentials
between successive grades. Values less than that will lead back to the
problem of fractional values, and it will be nearly impossible to secure
acceptance, as previously stated. The maximum limit of differentials
between successive labor grades will depend upon the range of skill classi-
fications to be covered by the job evaluation plan (i.e., unskilled through
semi-skilled to and including high-skilled jobs is the top consideration).
For the ordinary jobs usually found, and included in the job evaluation
plan, in a mass production industry the differential amounts to about
ten to fifteen percent of the specific base rate. Higher differentials make
the jump from labor grade to labor grade too abrupt and create distrust.

This whole situation can be illustrated graphically as shown in Figure
VII. In this case the differential between labor grade one and two is five
percent; between grades two and three, about seven percent; about nine
percent between labor grades three and four; and around eleven percent
between the fourth and fifth labor grades.

In the case of the overlaps between labor grades one should consider
the psychological impact of the status of workers in various grades.
Whether it is justified or not, each worker builds up a feeling of status
based somewhat on his particular labor grade assignment and the cor-
responding money which is attached to the labor grade. It is this status,
as well as the yet unsolved problems of impartial and inadequate systems
of merit rating, which has caused many plans to resort to a single rate
for each labor grade. This is to avoid the problem that arises as the
result of both overlap and merit rating.

When labor grades overlap there is a possibility of a person in one



33



FIGURE 2III



1.53



c 1-39
^ 1.35
q:

-z 1.23

=}
o

X



o
CD



1.16
1.12-
1. 10
1.05
1.00



If |-^::/; 2 : 3 4 5



Labor Grades

labor grade receiving the same hourly rate as a person in the next or
subsequent labor grade. Usually this overlap is spoken of in terms of
money. An illustration of no overlap is shown in Figure VII. Here each
successive labor grade — a single wage structure • — • pays a distinctly
diflferent rate. There can be no case of people assigned to different labor
grades receiving the same amount of money per hour, or what ever other
pay basis is used.

Should rate ranges be used in Figure VII one may get a picture as
shown in Figure VIII.

In Figure VIII an arbitrary rate of ten percent of the starting rate
was used in each case. But, note that between labor grades one and two
there is an overlap of five cents; betwen labor grades two and three, a
four cent overlap; and no overlap between grades three and four and
even a gap of four cents between grades four and five.

For example, if an arbitrary rate of twenty-five percent of the starting
rate is used for the rate range in each grade. Figure IX results. In this
case there is an overlap between each two successive labor grades. How-
ever, there is a serious overlap picture developing. Considering grades
one and four, there is an overlap of three cents. The type of work —
even the minimum work — which will be found in grade four should be
far superior in requirements to that in grade one; yet someone in grade
one could receive equal to or better than a person in grade four. Such a
situation hits strongly at the psychological status of the worker in grade
four. This successive overlap should be avoided. This is even more serious



34



1.74-



FIGURE IX



o



3
O



o
CD



.54-



.40-
.37-

.31-
1.25-
1.23-



.05-
.OOH



Labor Grades

if a job in labor grade one is on incentive and a job in labor grade four
is on day work.

A certain amount of overlap is essential for management to provide
flexibility of transfer and upgrading with no additional expense during
the employee trial period on the new job.

A third thought on establishing labor grades is to have a rational
pattern of differentials throughout the labor grades. In Figure VII this
pattern is 5 (difference between $1.05, grade two, and $1.00, grade one,
etc.), 7, 11, 16. Another pattern could be 5, 5, 5, 5; or 5, 6, 7, 8. But a
pattern of 5, 7, 8, 12 is hard to justify because a suppression is shown
in the middle, as contrasted with the implication of increasing importance
of successively higher labor grades.

There are many other important considerations in establishing labor
grades. It must be kept in mind that:

(1) Tradition is hard to overcome; so grouping of jobs into labor
grades may have to be a little different than strict evaluations order.

(2) Occupational grouping of jobs should be considered.

(3) General union feeling and prevailing practice in the area on the
number of labor grades should not be ignored.

(4) The greater the difference of the rate for the lowest paid job and
that for the highest paid job, the more labor grades one will have.



35



CROSS-CHECKING THE EVALUATION PROGRAM

The foregoing sections have outHned the general and procedural as-
pects of job or position evaluation. As is evident, much of the mechanics
is based on judgment. But, there is need for better guidance in the
judgment process than has been evident in practice to date. In this section
some procedures are given to assist in this judgment process. Under no
conditions are these procedures the absolute measure and neither are they
to be used as the sole criterion. There are too many problems within the
company and too many problems arising from forces outside the company
— problems and forces that cannot as yet be fully evaluated — to allow
a dogmatic solution.

Factors to be considered in cross-checking the evaluation program
include :

( 1 ) The effectiveness of each factor.

(2) The validity of each factor.

(3) The validity of factors in respect to each other.

(4) The continuance of validity of the plan.

Effectiveness of Each Factor

This means: is the factor necessary and is it fully used? In many cases
a factor (see Charts IV, V, VI, and VII) is listed with several possible
degrees of its use. Such listing assumes that the factor is necessary and
that it is to be fully used. The latter is important when considering item



CHART VUr


FACTOR: Experience


Degrees


Jobs assigned


1
2


1
6


3


3


4


8


5


II


6


2


7


3


8


1



36



CHART IX



FACTOR: Work Hazards


Degrees


Jobs assigned


1
2


5


3


5


4


21


5


4


6






three, the validity of factors in respect to each other. To check the first
point, one should tabulate the degrees for the factors and the correspond-
ing number of jobs assigned to each degree. Such a tabulation could be
as shown in Chart VIII. This shows that the factor is at least used
throughout its range. Whether it contributes to the plan properly will be
shown subsequently. In the following illustration. Chart IX, some ques-
tion is to be raised regarding the effectiveness of the factor, especially
since the trend is to simplify jobs in industry, rather than to make them
more complex. Therefore, if the higher degrees, especially those of a
factor, are not used at the time of installation of the job evaluation pro-
gram, it is even more likely that they will not be used in the future. This
has a special impact on the validity of factors in respect to each other.

Validity of Each Factor

The second consideration, the validity of the factor, is a measure of
how distinctly the factor is separated from other factors in the plan. The
first step is to tabulate the jobs and to set opposite each job the rating
for the factors to be checked against each other. A suggested tabulation
is shown in Chart X. The same data then can be plotted on a graph as
in Figure X.

If the trend line can be established as shown, if the points cluster close
to the trend line, and if the slope is at about 45 degrees, rising or falling
(assuming uniform scales on the graph), then there is a high correlation
between the two factors. In such a case, one factor or the other is
sufficient in the plan; if both are used, usually they are measuring the



37







CHART X






Job


Factor One


Factor Two




Job


Factor One


Factor Two




Rating


Rating






Rating


Rating




in degrees


in degrees






in degrees


in degrees


1


2


1




19


3


3


2


1


1




20


3


3


3


1


2




21


2


3


4


2


1




22


2


3


5


1


I




23


3


3


6


1


I




24


3


3


7


2


2




25


3


4


8


2


2




26


3


2


9


1


1




27


3


2


10


2


2




28


3


2


II


2


2




29


3


2


12


2


2




30


3


2


13


3


4




31


5


4


14


5


5




32


5


4


15


2


2




33


3


2


16


2


2




34


3


2


17


1


3




35


3


2


18


2


3

1











same thing. If no trend line can be quickly seen, or if a trend line run-
ning nearly horizontal results, each factor is measuring a relatively
different feature of the job.

Validity of Factors in Respect to Each Other

The third item, the validity of factors in respect to each other, con-
siders the importance attached to the individual factors. In Charts IV
and VI each factor is considered to be of equal importance and each
factor has the same number of points assigned to it as do each of the
remaining factors. In Chart V, however, some factors are considered
more important than other factors. This is shown by a different number
of degrees, so that even though all factors start out with the same number
of points, some factors can have a higher average rating than others.
Chart VII definitely shows the variable importance attached to respective
factors.

There is nothing wrong with the importance attached to each factor



38



FIGURE X



M 4

0)

0>

w.



0)

■° 3

I

0)

c

? z

o

u

o




Trend line



o2 ^7 o3



^ ol oi



12 3 4 5

Factor Two- degrees

except that the entire weighting is based on judgment. Furthermore, the
plan is explained as such to those who will operate under it. The purpose
of this check is to make sure that: ( 1 ) the judgment of factor importance
is correct, and then that (2) the average collective results of the job
evaluations are the same as the plan shows. For example, is it a correct
judgment of the factor "experience" to make it twice as important as the
factor "working conditions"? Do the average collective results of the
evaluations of all the jobs show this to be so?

To make this check on the relative importance of factors, consider the
following example. From Chart X the average rating for Factor One is
2.48; and for Factor Two, 2.37. (Secured by adding all the degree ratings
for Factor One and dividing by the number of jobs considered. Likewise,
for Factor Two.) Then add the difference, regardless of sign, between
2.48 and each job rating for the factor. For Job One, for example, the
difference is 0.48. Divide this total by the number of jobs considered.
For Factor One this arithmetic deviation is 0.84; and for Factor Two,
0.81. In terms of points. Factor One is 4.20 (secured by multiplying 0.84
by the 5-point intervals between degrees) and Factor Two is 8.10.

The specific plan from which the data were taken is shown in Chart
XL This shows that Factor Two is considered to be double the impor-
tance of Factor One. If this is the case, the arithmetic deviations for these
factors should have the same or nearly the same relationship. The actual
figures in this case show that the ratio is nearly two (8.10 dixided by



39



CHART zr



FACTOR


Degrees


1


1


3


4


5


One
Two
Three ^^^^^


5


10
x20


15
30


20
40


25
50









4.20). So the two factors in use are contributing, on the average, to the
results in accordance with the plan as designed. This is a reasonably close
approximation method to use to guide judgment.

Continued Validity of the Plan

A fourth point to check is the continuance of the validity of the plan.
It is important to recognize that from time to time new jobs will be
evaluated, and changed jobs will be re-evaluated. The main question
that arises is : How do these new evaluations change the actual collective
results from what they were at the time of installing the plan? The
method used to answer this question is equivalent to setting up a quality
control chart of much the same sort as that used by production
departments.

The procedure will be to determine the average rating for each fac-
tor at the time of installing the plan. Then periodically — every six
months or a year — thq, average factor ratings should be determined for
all evaluations which have been made within the six months or year
period. These average ratings are then compared with those at time of
installation. Differences will indicate whether or not the new evaluations
are more lenient or more severe than those that were adapted at the
time of installing the plan. This sort of check helps keep inequities from
developing.

INSTALLATION AND MAINTENANCE OF PLAN

Before a job evaluation plan is actually installed a considerable
amount of explanation and selling has to be done to those who have to
operate with it. This selling is a continuous process. It must go on
throughout the period of establishing the plan and ever afterwards.

Above all, remember a job evaluation plan is used in a dynamic situ-
ation, and it needs continual guidance by someone who has been assigned



40



this responsibility and who makes this responsibility an important part
of his work.

A procedure must be set up which, among other things, will :

( 1 ) Handle all grievances on present evaluations or re-evaluations.

(2) Periodically review' the jobs to ascertain whether or not the jobs
are the same as originally described.

(3) Periodically conduct sound wage surveys.

(4) Continue to receive support from all those closest to the jobs —
as by calling attention to job content change.

(5) Provide close check upon adherence to job description in work
assignments — except in emergencies.

(6) Provide a constant source of assistance to show and solve im-
pacts of the job evaluation plans on other personnel policies and labor
contract provisions.

CONCLUSION

This Bulletin is intended mainly to present the techniques of job
evaluation (or position evaluation) and to indicate many "fringe"
aspects which must be considered. It does not attempt to discuss when
and by whom job evaluation should be adopted. That is a decision to be
made by management and labor representatives in each situation. As is
evident, judgment forms an important part of any job evaluation. And,
since judgment is not positive in measurement, psychological and socio-
logical aspects take on added meaning in any evaluation. No attempt has
been made here to weigh the pro and con of any of the human relation-
ships. Instead, the reader is encouraged to investigate other current
publications on the subject, some of which are listed in the following
bibliography.



41



SELECTED BIBLIOGRAPHY
Books and Pamphlets

Baker, Helen, and True, John M. The Operation of Job Evaluation Plans, A


1 3

Online LibraryLeo Charles PigageJob evaluation (Volume BEBR Faculty Working Paper v.5, no.3) → online text (page 3 of 4)