You are on page 1of 4

PsychologicalReports, 2005,96,408-410.

O Psychological Reports 2005

REVISITING T H E FACTOR STRUCTURE OF T H E KIRTON


ADAPTION-INNOVATION INVENTORY '
SUBIN IM MICHAEL Y. HU
San Francisco State University Kent State University

Summay.-The original l r t o n Adaption-Innovation Inventory, used to measure


innovative (as opposed to adaptive) individual cognitive styles, has been reported to
have three factors: Sufficiency of Originality, Efficiency, and Rule/Group Conformity.
In exploring the construct validity of the Kirton Adaption-Innovation Inventory, find-
ings from a 2003 study by Im, Hu, and Toh showed the existence of two subdimen-
sions of the Sufficiency of Originality factor-Idea Generation and Preference for
Change. In this study, using a sample of 356 household participants, with an average
age of 56.0 yr. (SD= 14.0), average income of $39,700 (SD= $19,200), and average of
15.0 yr. of education (SD=2), from the Arkansas Household Research Panel, we con-
ducted factor analyses. The results specific to our selected sample indicate that a four-
factor model recognizing the two subdimensions of Sufficiency of Originality has a
better fit than the original three-factor model.

The 32-item Kirton Adaption-Innovation Inventory (Kirton, 1976, 1978)


has been widely used to measure the innovative (those who do things differ-
ently) and adaptive (those who do things better) styles of individuals because
it has good reliability and content validity. In testing for the construct valid-
ity of the inventory, earlier studies focused on exploring the factor structure
(Kirton, 1976, 1978; Mulligan & Martin, 1980; Goldsmith, 1984; Beene &
Zelhart, 1988) by using exploratory factor analysis. From these studies, it ap-
pears that the inventory has three factors. The first factor is called the Suffi-
ciency of Originality. The second factor is called Efficiency which describes
people who are precise, reliable, and disciplined. The third factor, called
Rule-Group Conformity, describes people who fit well into a bureaucracy be-
cause they have proper respect for authority and rules. Bagozzi and Foxall's
confirmatory factor analyses (1995, 1996) support the construct validity of
the original Kirton Adaption-Innovation Inventory by empirically validating
the three-factor structure. However, the construct validity with regard to the
three-factor structure has been questioned because significant amounts (more
than 60%) of the test variance remain unexplained in factor analyses (Ham-
mond, 1986; Taylor, 1989a; Foxall & Hackett, 1992). In addition, the Suffi-
ciency of Originality dimension has more than ten items. Any factor that has
more than 10 items tends to split into different factors. Taylor (1989a, 1989b)

'The authors thank Barry Bayus, Kenneth A. Ballea, and Rex Toh for their help. Send corre-
spondence to Subin Im, Department of Marketing, College of Business, San Francisco State
University, San Francisco, CA 94132 or e-mail (subinim@sfsu.edu).
KIRTON ADAPTION-INNOVATION INVENTORY 409

and Foxall and Hackett (1992) claimed that the Sufficiency of Originality
subscale is not homogeneous; instead, it has two subdimensions, one related
to Idea Generation and the other related to Preference for Change. Most re-
cently, using confirmatory factor analysis on only the Sufficiency of Origi-
nality factor items, Im, Hu, and Toh (2003) showed that this factor is com-
posed of the two subdimensions previously mentioned. This raises two inter-
esting questions: would a four-factor model (Idea Generation, Preference for
Change, Efficiency, and Rule/Group Conformity) have more construct valid-
ity than the traditional three-factor model (Sufficiency of Originality, Effi-
ciency, and Rule/Group Conformity)? When confirmatory factor analysis is
used with the entire inventory of items, would the procedure reveal a three-
or four-factor model?
Using a sample of 356 household participants from the Arkansas House-
hold Research Panel, with an average age of 56.0 yr. (SD = 14.0), average in-
come of $39,700 (SD= $19,200), and average of 15.0 yr. of education (SD =
2.0), in the first stage we performed exploratory factor analysis (using the
principal axis factor method with varimax rotation) on the 32-item Kirton
Adaption-Innovation Inventory. The analysis gave three factors (Sufficiency
of Originality, Efficiency, and Rule-Group Conformity) which explained
37.4% of the total variance. But in contrast to previous studies which report-
ed that 13 items loaded onto Sufficiency of Originality, our results showed
that only 11 items loaded onto this factor. In our particular sample, two
items ("prefers to work on one problem at a time" and "prefers changes to
occur gradually") instead loaded onto the Efficiency factor, conceivably be-
cause they are considered to be characteristics of those who subscribe to the
"precise, reliable, and disciplined" nature of the Efficiency factor. Neverthe-
less, the three dimensions exhibited good internal consistency as reflected by
high Cronbach alphas (.87 for Sufficiency of Originality, .87 for Efficiency,
and .78 for Conformity).
In the second stage, we performed confirmatory factor analysis on the
32 items but with two different specifications. Model 1 specified the three
factors identified by exploratory factor analysis while Model 2 specified the
four factors as the Sufficiency of Originality factor divides into two separate
subdimensions, Idea Generation and Preference for Change, based on a re-
cent finding (Im, et al., 2003). Using the Maximum Likelihood criterion, we
compared the two models in terms of component fit, fit indices, and chi-
square statistics. First, with respect to component fit, all coefficients for fac-
tor loadings were significant (p< .O1) for both models, suggesting good com-
ponent fit for all 32 items. Standardized coefficients for the loadings ranged
from .36 to .76 for Model 1 and from .37 to .79 for Model 2. Second, the
baseline fit indices-Normed Fit Index (NFI), Incremental Fit Index (IFI),
Relative Fit Index (RFI), and Tucker-Lewis Index (TL1)-were all over 0.95
for both models, indicating that both models fit the data well. The Root
Mean Square Error of Approximation (RMSEA) values for both models were
.O7, showing an equally good fit for both models based on Browne and
Cudeck's cutoff criterion (1993). But, third, when we used chi-square esti-
mation to examine the overall fit for Models 1 and 2, we noted that the chi-
square difference was A ~ ~ = -xiMode1
x 2 = ~1208.02
~ - ~1148.16=59.86,
~ ~ ~
Adf = 461 - 458 = 3, p < .01. Thus, Model 2, with a lower chi-square statistic,
has a better overall fit than Model 1, favoring the four- over the three-factor
specification.
Our results from confirmatory factor analyses indicate the possibility of a four-factor
structure as suggested by Taylor (1989a, 1989b), Foxall and Hackett (1992), and Im, et al.
(2003). It should also be noted that our findings are not definitive as those of the previous
study (Im, et al., 2003). As is the case of most if not all social science studies, our findings are
specific to the sample we used for this study. We also acknowledge that our four-factor model
still has high unexplained test variance, so our results are susceptible to sampling variation.
Since we used the derived version of the Kirton Adaption-Innovation Inventory (although all
measurement items were collected from the Kirton original 1976 article), one should not direct-
ly compare our findings with the three-factor model reported in previous literature (e.g., Kir-
ton, 1976, 1978; Mulligan & Martin, 1980; Goldsmith, 1983; Bagozzi & Foxall, 1995, 1996).

REFERENCES
BAGOZZI,R. I?, & FOXALL,G. R. (1995) Construct validity and enerali~abilit~ of the Kirton
Adaption-Innovation Inventory. European Journal of ~ersonal'ty , 9, 227.1-227.23.
BAGOZZI, R. P., & FOXALL, G. R. (1996) Construct validation of a measure of adaptive-innova-
tive cognitive styles in consumption. International Journal of Research in Marketing, 13,
201-213.
BEENE,J. M., &ZELHART, I? F. (1988) Factor analysis of the Kirton Adaption-Innovation Inven-
tory. Perceptual and Motor Skills, 66, 667-671.
BROWNE,M. W., &CUDECK,R. (1993) Alternative ways of assessing model fit. In K. A. Bollen
& J. S. Long (Eds.), Testing structural equation models. Newbury Park, CA: Sage. Pp.
136-162.
FOXALL, G. R., &HACKETT,I? M. W. (1992) The factor structure and construct validity of the
l r t o n Adaption-Innovation Inventory. Personality and Individual Differences, 13, 967-
975.
GOLDSMITH, R. E. (1984) Personality characteristics associated with adaption-innovation. Jouv-
nal of Psychology, 117, 159-165.
HAMMOND, S. M. (1986) Some pitfalls in the use of factor scores: the case of the Kirton Adap-
tion-Innovation Inventory. Personality and Individual Differences, 7, 401-407.
IM, S., Hu, M. Y.,&ToH, R. S. (2003) Exploring the dimensionality of the originality subscale
of the Kirton Adaption-Innovation Inventory. Psychological Reports, 93, 883-894.
KIRTON,M. J. (1976) Adaptors and innovators: a description and measure. Journal of Applied
Ps~chology,61, 622-629.
KIRTON,M. J. (1978) Have adaptors and innovators equal levels of creativity? Psychological Re-
ports, 42, 695-698.
MULLIGAN, G., &MARTIN,W. (1980) Adaptors, innovators and the Kirton Adaption-Innovation
Inventory. Psychological Reports, 46, 883-892.
TAYLOR, W. G. K. (1989a) The Kirton Adaption-Innovation Inventory: a re-examination of the
factor structure. Journal of Organizational Behavior, 10, 297-307.
TAYLOR, W. G. K.(1989b) The Kirton Adaption-Innovation Inventory: should the sub-scales be
orthogonal? Personality and Individual Di'erences, 10, 921-929.

Accepted Februapy 11, 2005.


410E

ERRATUM

An error has been called to our attention:


IM, S. & HU, M. Y. (2005) Revisiting the factor structure of the Kirton Adaption-
Innovation Inventory. Psychological Reports, 96, 408-410.

The acknowledgement at the bottom of page 408 should read: The authors thank
Barry Bayus, Kenneth A. Bollen, and Rex Toh for their help.

Erratum published in:


Psychological Reports, 2005, 96, 243E. © Psychological Reports 2005

You might also like