Journal of Vocational Behavior 36, 176-180 (1990)

Some Extensions of the lachan Congruence Index

RONALDO IACHAN

Research Triangle Institute

This article extends an index introduced in Iachan (1984) to measure agreement

or congruence between a pair of three-letter codes. Three-letter codes are used
in Holland’s vocational theory to characterize occupations, work environments,
job preferences, college majors, personality types, etc. The congruence index is
extended to cover two situations: (a) when ties occur in the three-letter codes
and (b) when only two top letters are recorded and/or used. This article suggests
methodologies for breaking ties and for measuring agreement between two-letter
codes. In addition, a computer algorithm for computing the index is included in
an appendix. 0 I!990 Academic Press, Inc.

Iachan (1984) introduced a general measure of agreement between any

pair of three-letter codes that has found wide applicability in the career
counseling setting. The index has been used to assess the congruence
between a person’s Holland’s code-obtained, for example, from the
Self-Directed Search (SDS)-and one or more vocational aspirations or
vocational choices (Holland, 1985). A comprehensive review of earlier
research on person-environment congruence is provided in Spokane
(1985).
In the Holland typology (Holland, 1973) a three-letter code is formed
by selecting in sequential order three (the top three by some assessment)
of the following personality types or dominant work environment
characteristics:
R, Realistic
I, Investigative
A, Artistic
S, Social
E, Enterprising
C, Conventional.
Three-letter codes are used in Holland’s theory to characterize oc-
cupations (or clusters of related occupations), work environments, job
preferences, college majors, etc. Congruence is then measured by the
176
OOOl-8791/90$3.00
Copyright 0 1990 by Academic Press, Inc.
All rights of reproduction in any form reserved.
 CONGRUENCE INDEX 177

similarity (or agreement) between the three-letter codes. Different con-

gruence measures have been developed by Zener and Schnuelle (1976),
Kwak and Pulvino (1982), and Iachan (1984). The index constructed by
Iachan (1984) has been advocated by Holland (1985, p. 15, p. 85) as the
“most accurate technique” for determining the degree of agreement
between three-letter codes. Holland (1985, p. 85) further states that this
index “is expected to become the most popular way to assess the con-
gruence between a person’s SDS code and one or more vocational as-
pirations.” A summary description of the index computation procedure
is provided in Table 1.
Recent applications of the Iachan index have involved, for instance,
(a) agreement between current and past college majors, (b) agreement
between intended college majors and SDS summary codes, and (c) agree-
ment between current occupation code and SDS code or a code obtained
from the Career Decision-Making System (Harrington & O’Shea, 1982).
In several of these applications, two recurrent problems were appar-
ently encountered: (a) the occurrence of ties between two codes and (b)
the use of two-letter codes. This note suggests solutions to these two
problems, formalizing suggestions that I have given in various personal
communications. An appendix presents a computer program for the cal-
culation of the index. The program may be particularly useful if con-
gruence is to be assessed for a large number of subjects.
BREAKING TIES
The linear additive Iachan index makes the treatment of ties extremely
simple. Whenever two (or more) codes are tied, the index is computed
as the average of the individual index values. In the simplest case when
only two codes are tied, only two index values need to be averaged.
This approach, illustrated here, easily generalizes to multiple ties.
The occurrence and resolution of ties are best explained with the help
of a concrete example (other examples may be found in Holland, 1985,
pp. 24, 26). Suppose one wishes to compute the congruence between a
code of S E I and one of E S A/I, where the latter code shows a tie in
the third letter of the code (between Artistic and Investigative).

TABLE 1
Calculating Congruence

SDS code
First second Third Other
Occupation code letter letter letter letters

First letter 22 10 4 0
Second letter 10 5 2 0
Third letter 4 2 1 0
178 RONALD0 IACHAN

To compute the index for the two codes in the example, one first
computes the two individual congruence index values,
m(l) = congruence value between SE1 and ESA,
and
m(2) = congruence value between SE1 and ESI.
Then, the congruence measure is computed as the average,
A4 = (m(l) + m(2))/2.
More specifically, using the methodology described in Iachan (1984),
the two individual values in the example are
m(l) = 2*W(l, 2) + W(3,3) = 21,
and
m(2) = 2*w(l, 2) = 20,
so that their average is M = 20.5, the index value in this case.
TWO-LETTER CODES
This section extends the agreement measures between three-letter
codes constructed in Iachan (1984) to the two-letter code situation. The
most obvious class of applications arises when only the two top letter-
types are recorded in Holland typology (e.g., SDS) uses.
The basis for the extended index is the following set of weights, de-
veloped according to the same rationale as in Iachan (1984):
W(1, 1) = 5,
W(1, 2) = W(2, 1) = 2,
W(2, 2) = 1.
Table 2 gives a representation of these basic weight values. Using this
set of weights, a complete matrix of index values was generated and is
presented in Table 3.
SUMMARY
This article presents some extensions to the congruence index pro-
posed by Iachan (1984) that take advantage of the simple-linear addi-
TABLE 2
Congruence for Two-Letter Codes
Code-I First letter Second letter

First letter 5 2
Second letter 2 1
 CONGRUENCE INDEX 179

TABLE 3
Modified Index for Two-Letter Codes

Configuration Example Agreement index

(1, 2) RE, RE W(1, 1) + W(2, 2) = 6

(1. 0) RE, RI W(1, 1) = 5
(2, 1) RE, ER 2W(l, 2) = 4
(0, 1) RE, IR W(1, 2) = 2
(0, 2) RE, IE W(2, 2) = 1

tive-structure of the index. Other, more theoretical extensions given in

Iachan (1983) include the embedding of the index in a general family of
indices for ranked ordinal data, and the general case when the “top k”
categories (or “letters”) are ranked.
The extensions provided in this article were developed in response to
questions raised by various practitioners when using the congruence
index. These questions-hopefully answered here-fall into the three
following categories:
(a) How do I break ties in the code?
(b) How do I measure congruence between two-letter codes?
(c) Is there a computer algorithm for computing the index?
APPENDIX
A Computer Program for the Calculation of the lachan Congruence
Index (M)
The computer program described in this appendix reads the pairs of three-letter codes
for N subjects and outputs the Iachan agreement index, M. The three-letter codes are
represented as (1 x 3) vectors
x = (X(l), X(2), X(3)),
y = (Y(l), Y(2), Y(3)).
In addition, the (3 x 3) matrix (W) of basic weights, W(Z,J), needs to be read as well
(see Table 1 for values of W).
The basic program logic is to compare each entry of Y with each entry of X, and add
the corresponding value of W to the cumulative M-measure. The program is written in a
generic, FORTRAN-like language so that the code may be adapted to any high-level
language (e.g., PL-1, APL, or PASCAL). The program is particularly suitable for SAS.
Step 1: Read W-matrix and initialize cumulative index M:
M = 0;
DO Z = 1, 3;
DO .Z = 1, 3;
READ W(Z, J);
END;
END;
180 RONALD0 IACHAN

Step 2: Read vectors X and Y for each case:

DO I = I. 3;
READ X(r). Y(I):
END;
Step 3: Cumulate M for matched values:
DO I = 1: 3;
DO J = 1, 3;
IF X(l) EQ Y(J) THEN M = M + W(f, J);
END;
END;
Step 4: Output M if all file (N cases) has been read--otherwise, return to Step 2 to read
on;
IF END-OF-FILE THEN PRINT M;
ELSE RETURN TO STEP 2.

REFERENCES
Harrington, T. F., & O’Shea, A. J. (1982). Cureer decision-making system: Manual. Circle
Pines, MN: American Guidance Services.
Holland, J. L. (1973). Making vocational choices: A theory of careers. Englewood Cliffs,
NJ: Prentice-Hall.
Holland, J. L. (1985). The self-directed search: Professional manual. Odessa, FL: Psy-
chological Assessment Resources, Inc.
Iachan, R. (1983). Agreement and differentiation for ordinal data. Paper presented in the
Social Statistics Section of the American Statistical Association.
Iachan, R. (1984). A measure of agreement for use with the Holland classification system.
Journal of Vocational Behavior, 24, 133-141.
Kwak, J. C., & Pulvino, C. J. (1982). A mathematical model for comparing Holland’s
personality and environmental codes. Journal of Vocafional Behavior, 21, 231-241.
Spokane, A. R. (1985). A review of research on person-environment congruence in Hol-
land’s theory of careers. Journal of Vocational Behavior, 26, 306-343.
Zener, T. B., & Schnuelle, L. (1976). Effects of the self-directed search on high school
students. Journal of Counseling Psychology, 23, 353-359.

Received: June 23, 1989.