Measuring a University’s Image: Is Reputation an Influential Factor?

Abstract

Today, the higher education sector can be considered a market and, within it, private university education is a common marketable service in the literature on higher education management. Research on the analysis of the variables that generate the university image has been the subject of numerous investigations. Although there is no generally accepted definition, most authors approach the measurement of image through multi-factor scales, with variables relating to functional and psychological elements. This research aims to contribute to study of the most determinant variables in measuring a product’s image, assessing especially the effect of the reputation construct. This was done through measuring the image of the “private university” product as perceived by citizens of Andalusia, based on a standardized model with three dimensions—functional and affective aspects and reputation. After adapting and validating the questionnaire, a two-phase procedure is performed with double validation through exploratory and confirmatory factor analysis. The results show an adapted scale valid for measuring the image of a generic product; with presentation and discussion of a series of advantages of incorporating reputation and measuring image through models with three dimensions. This article goes deeper into the possible influence of reputation as a determinant factor in measuring image, an assumption arising from some models for measuring image, something that so far has not been sufficiently contrasted.

Full text



Citation: Gutiérrez-Villar, B.;
Alcaide-Pulido, P.; Carbonero-Ruz, M.
Measuring a University’s Image: Is
Reputation an Influential Factor?
Educ. Sci. 2022,12, 19. https://
doi.org/10.3390/educsci12010019
Academic Editor: Han Reichgelt
Received: 11 November 2021
Accepted: 21 December 2021
Published: 30 December 2021
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2021 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
education
sciences
Article
Measuring a University’s Image: Is Reputation an
Influential Factor?
Belén Gutiérrez-Villar 1,* , Purificación Alcaide-Pulido 2and Mariano Carbonero-Ruz 3
1Departamento de Gestión Empresarial, Universidad Loyola Andalucía, Escritor Castilla Aguayo 4,
14014 Córdoba, Spain
2Departamento de Comunicación y Educación, Universidad Loyola Andalucía, Escritor Castilla Aguayo 4,
14014 Córdoba, Spain; palcaide@uloyola.es
3Departamento de Métodos Cuantitativos, Universidad Loyola Andalucía, Escritor Castilla Aguayo 4,
14014 Córdoba, Spain; mcarbonero@uloyola.es
*Correspondence: belengut@uloyola.es
Abstract:
Today, the higher education sector can be considered a market and, within it, private uni-
versity education is a common marketable service in the literature on higher education management.
Research on the analysis of the variables that generate the university image has been the subject of
numerous investigations. Although there is no generally accepted definition, most authors approach
the measurement of image through multi-factor scales, with variables relating to functional and
psychological elements. This research aims to contribute to study of the most determinant variables
in measuring a product’s image, assessing especially the effect of the reputation construct. This was
done through measuring the image of the “private university” product as perceived by citizens of
Andalusia, based on a standardized model with three dimensions—functional and affective aspects
and reputation. After adapting and validating the questionnaire, a two-phase procedure is performed
with double validation through exploratory and confirmatory factor analysis. The results show an
adapted scale valid for measuring the image of a generic product; with presentation and discussion of
a series of advantages of incorporating reputation and measuring image through models with three
dimensions. This article goes deeper into the possible influence of reputation as a determinant factor
in measuring image, an assumption arising from some models for measuring image, something that
so far has not been sufficiently contrasted.
Keywords: image measurement; reputation; private university; confirmatory factor analysis
1. Introduction
The implementation of marketing to university is an area that is growing in impor-
tance as competition between universities increases. Thus, the irruption of marketing
philosophy in higher education institution (HEI) is understood as a competitive strength
where globalization of HEIs has played a part [
1
,
2
]; and has pushed the universities to
develop new marketing strategies to increase brand engagement [
3
]. According to this
marketing philosophy, the study of two intangible assets in higher education, image and
reputation, have received increasing academic research in recent years [4–8].
There is considerable consensus on the benefits of a good university image, such as a
source of differentiation from other competitive institutions [
9
], or to improve the relations
with stakeholders [
10
]; Mainly, a good image can contribute to the achievement of new
students [11–13].
However, the analysis of image in perceptual meaning becomes an ambiguous con-
cept [
14
,
15
]. This may be due to related but different concepts revolving around the
idea of image, these being: corporate image, brand image, product image, brand per-
sonality, positioning, identity, and reputation, which are sometimes used erroneously as
synonyms [
16
–
18
]. Certainly, reputation, and its differentiation from image, is a subject
Educ. Sci. 2022,12, 19. https://doi.org/10.3390/educsci12010019 https://www.mdpi.com/journal/education

Educ. Sci. 2022,12, 19 2 of 13
that has given rise to various stances, without there being consensus as to the relation of
dependence between these two concepts [19].
Nevertheless, there is considerable agreement on the premise that image can be
measured, but it is hard to measure [
20
]. Various authors support the idea that the image
of a product, brand or company are subjective phenomena that can be measured by
compiling the opinions of individuals or groups of individuals, with no need for them
to be consumers [
17
,
21
–
23
]. However, there is no generally accepted model to measure
image, with the methodologies proposed, and the variables used in published models
being disparate [15].
Following on from the above, this study aims to contribute more evidence to the
discussion on perceptual measurement of university image with a twin objective: from
the theoretical bases and results of empirical work published, it was decided to submit
to examination the brand image measurement scale proposed by Martínez et al. [
24
],
formed of three dimensions—functional image, affective image and reputation, to confirm
its validity when the object measured is a product unconnected to a specific brand; and
secondly, the study seeks to confirm the significance of reputation as an independent and
necessary construct to measure image.
The brand measurement scale was chosen due to its versatility since it considers a set
of items adaptable to each specific context –as the authors themselves point out; because it
is conceived as universalizable and has been used successfully to measure brand image in
different sectors of activity with the model being validated in subsequent studies [25–27].
In this way, the main research question is to know what are the factors that shape the
image of a private university in the minds of people in general, regardless of their level
of studies or other sociodemographic characteristics. In addition, given the importance of
the reputation construct in the university context, this study aims to determine whether
reputation is one of the influential factors in the measurement of the image of a private
university, when no specific institution or brand is mentioned.
The Spanish university system has 83 universities, of which 50 are public and 33 are
private. Although private universities account for 40% of Spanish universities, according
to official statistics from 2018 onwards, they only account for around 14% of university
students. In Andalusia, there is only one private university to date. The decision to focus
the study on private education is considered pertinent, as this topic is widely debated in
society, with all individuals finding it easy to give their appreciation and opinion.
2. Literature Review and Hypothesis
2.1. Concept of Image and Levels of Measurement
Academic interest in the concept of image dates to the 1950s, and authors at that
time were suggesting that human behavior depends more on image than on objective
reality itself [
28
,
29
]. Transposed to the world of products and brands, since the mid-20th
century, many authors have proposed definitions of image which highlight the perceptual
component leading individuals to form a mental image of the object (product, brand,
company, sector or even person) through a set of sentiments, ideas and attitudes, rather
than a sum of physical attributes and characteristics.
The review of the literature on the concept of image reveals many commonalities in
the definitions that emphasize the cognitive or mental process through which individuals
create an image. Many authors describe aspects such as ideas, sentiments, attitudes,
mental constructs, perceptions, beliefs, expectations, impressions, or stereotypes as essential
determinants [4,15,19,30].
This research is based on the definition of image proposed by Currás [
17
] (p. 29),
who sees it as: “the sum, result, or accumulation of beliefs, attitudes, experiences, senti-
ments, impressions or information that a subject—consumers, stakeholders, individuals in
general—has of an object”.
In addition, a more detailed study of the degree of generality or specificity of the
object of the image allows what is usually known as levels of image measurement. In

Educ. Sci. 2022,12, 19 3 of 13
this case, the few authors who have addressed this topic open the field of study, with it
being possible—and interesting—to analyze individuals’ mental image of objects other
than brands. Although the number of levels forming these studies does not coincide, there
is unanimity regarding the existence of a product image, understood as what individuals
perceive about a product or service in general, beyond brands or companies [
28
]. In
consequence, there are different objects/levels that can be of interest for the measurement
of image—a sector, a generic product, a company, or brands [
21
]. This paper focuses on
the measurement of a generic service, the private university, without alluding to a specific
university brand.
2.2. Reputation and Image
As happens with image, reputation has been studied in different areas related to
business [
31
], such as strategy, communication, marketing, sociology and even finance. But
unlike image, many of the studies published on reputation do not give a definition, taking
this for granted [32].
However, this study takes as a reference the definition proposed by Fombrun et al. [
33
]
(p. 87), where reputation is: “collective representation of a company’s past behavior and its
consequences, which shows the company’s capacity to deliver valuable results to multiple
stakeholders”. Also, as with image, reputation can be applied to different objects, such as
products or services [19,34].
Another important consideration arrived at after reviewing studies is the existence of
difficulties in establishing differences between the concepts of image and reputation [
35
,
36
].
So, a minority of work considers them similar [
31
,
32
]; but the majority argue that they are
interrelated concepts:
1.
some studies consider reputation as the result of combining the different images of a
company [37,38];
2.
and others consider it as the global assessment and influential element in measuring
the image of products or brands [
12
,
20
,
23
,
24
,
39
–
42
]. This study is based on the
second group.
Following the review above, the first hypothesis is formulated:
Hypothesis 1 (H1). Reputation is a significant factor in measuring a product’s image.
2.3. Measurement Scales for University Image
There is great consensus in research based on university image as a multivariate
concept generated in different subjects [
10
,
12
,
20
,
30
,
39
,
43
,
44
]. Also generalized is the thesis
covering inter-related cognitive and emotional aspects [
39
]. The cognitive (knowledge
about the object) and affective (motivations, sentiments and benefits) dimensions are
complementary, allowing better measurement and understanding of image [23].
There is no consensus on how to measure image and how many factors are necessary,
which has led to multiple measurement scales. In this context, there are two methodolo-
gies [
24
]: (1) ad-hoc scales that can capture image but cannot be generalized; and (2) stan-
dardized scales which, although allowing generalization, are not globally recognized.
Therefore, three multi-dimensional scales including reputation as a factor have been
reviewed. Two ad-hoc scales designed to measure university image—those of Beerly
et al. [
39
] and Sung & Yang [
4
]—and a standardized scale proposed by Martínez et al. [
24
],
which was used for vehicles and would be applicable to all types of products.
The first of the above-mentioned models establishes six dimensions to measure the uni-
versity’s brand image: the university’s orientation and training, reputation, massification,
accessibility, age and affective dimension. This scale was reduced to four dimensions in
the works published later [
23
,
45
], where the dimensions associated with cognitive aspects
emerge as significant: reputation and emotional aspects.

Educ. Sci. 2022,12, 19 4 of 13
The second model [
4
] finds three factors significant: personality of the university,
external prestige, and reputation; with the university’s external prestige being the factor
with the most weight in the image.
To explore other possibilities for study, this research opted to use the standardized
scale [
24
], which has not previously been applied to measure university image. According to
the authors, brand image would be formed of three constructs: functional image, covering
aspects related to attributes or benefits linked to the brand; affective image, including the
personality of the brand; and reputation, formed of the global perception of the brand over
time. In the original proposal, a scale of 9 items was used, three for each construct (see
Table 1).
Table 1. Studies carried out with the image measurement standardised scale of Martinez et al.
Items Proposed in Each of the Three Dimensions
Significance in Studies
Martínez et al.
(2004) [24]
Martínez et al.
(2005) [25]
Martínez et al.
(2007) [26]
Buil et al.
(2008) [27]
FUNCTIONAL IMAGE
F1 The products are high quality. Signif. Signif. Signif. Signif.
F2 The products present characteristics that other brands do not have. ——– Signif. Signif. ——–
F3 The products have better characteristics than the competition. ——– ——– ——– Signif.
F4 Consuming this brand is very unlikely to cause problems or
unforeseen events. ——– Not Signif. ——– ——–
F5 The competition’s products are usually cheaper. ——– Signif. ——– Not Signif.
F6 The models are cheap in relation to other brands. Signif. ——– ——– ——–
F7 I like the design of the models. Signif. ——– ——– ——–
AFFECTIVE IMAGE
A1 It is a brand that arouses positive feelings. ——– Signif. Signif. Signif.
A2 This brand conveys a personality that differentiates it from
rival brands. ——– Signif. Signif. Signif.
A3 Buying this brand says something about the class of person you are.
——– Not Signif. ——– ——–
A4 I have an image of the type of people that buy this brand. Signif. Not Signif. ——– ——–
A5 It is a brand that does not disappoint its customers. ——– Signif. Signif. Signif.
A6 The brand conveys values that differentiate it from other brands in
the sector. Signif. ——– ——– ——–
A7 The brand strives to innovate in new models, services
and/or technology. Signif. ——– ——– ——–
REPUTATION
R1 It is one of the best brands in the sector. ——– Signif. Signif. Signif.
R2 It is a brand that is committed to society. ——–
R3 It is a much consolidated brand in the market. ——– Signif. ——– ——–
R4 The brand is highly regarded. Signif.
R5 The brand is a professional in its category. Signif. Signif. Signif. Signif.
R6 You can trust this brand. Signif.
It should be noted that the standardized scale has been applied to different brands
in various sectors, such as: cars [
24
], dairy products, sports footwear, and electronics [
25
];
generic brandless products [
26
] –a tube of toothpaste and a box of chocolates–; and goods
of frequent consumption, lasting consumer goods and services [27].
In all these studies the scale was validated, resulting in explaining the three constructs
(functional, affective, and reputation), although as the authors themselves state, the 9 items
originally proposed to assess brands in the car sector should be adapted if applied to brands
in other sectors, and can be extended or substituted with others, as long as respecting the
three-dimensional structure.
Following the review above, it was decided to propose a three-factor model, leading
to the formulation of the last hypothesis:
Hypothesis 2 (H2).
The most correct formulation to measure the image of the generic product
“private university” is obtained by applying a standardized multi-factor scale with three dimensions:
functional image, affective image, and reputation.

Educ. Sci. 2022,12, 19 5 of 13
3. Materials and Methods
3.1. Sample
The type of sampling used was intentional. However, since the questionnaire has been
implemented in a previously existing consumer panel, we believe that the stability and
maturity implied by its past will have transferred to the sample, strengthening its repre-
sentativeness. So, the questionnaire launching method was online through a consolidated
consumer panel, managed by a marketing research company with a consistence control
in registered persons. The initial sample consisted of 778 questionnaires. This number
was the maximum we could afford with the limited budget we had available for primary
data collection. Before proceeding with the estimation and validation of the model, the
sample was filtered, eliminating the atypical observations, which resulted in a total of
27 questionnaires being eliminated.
It is not the objective of this work to study differences based on sex or age. Never-
theless, the panel leaders were asked to include men and women of different ages in the
sample, so that the sample would not be composed only of young individuals, a priori
more familiar with the subject matter of this study.
The questionnaire was validly completed by 751 Andalusians. The sample was
461 men (61.4%) and 290 women (38.6%) between 18 and 65 years old. The distribution
by age bracket was equal in the 4 age brackets chosen: 150 persons aged 18 to 25 (20%);
150 persons aged 26 to 35 (20%); 150 persons aged 46 to 55 (20%) and 151 persons aged 56
to 65 (20.1%).
3.2. Instrument
As stated above, there is no agreement in the significance of all the items used in
the research published [24–27]. This implied elaboration of a questionnaire including and
adapting all the significant items in the studies reviewed previously, as compiled in Table 2,
only adding one item in this study for reputation.
Table 2. Brand image measurement scale adapted to the “private university”.
Functional Image (F) Affective Image (A) Reputation (R)
f1. Education and training in
the private university are of
high quality.
a1. The private university
arouses positive sentiments.
r1. The private university is
the best option.
f2. The private university’s
premises are of high quality.
a2. The private university
conveys a personality that
differentiates it from the
state university.
r2. The private university is
committed to society.
f3. The research produced in a
private universities is of
high quality.
a3. Enrolling in a private
university says something about
the class of person you are.
r3. The private university is
much consolidated in the
market.
f4. Private universities present
characteristics that state ones
do not have.
a4. The private university does
not disappoint its customers.
r4. The private university
occupies very high positions
in university rankings.
f5. The private university
gives high value in relation to
the price that must be paid
for it.
a5. The private university has a
differentiated ideological
component.
The addition of item r4, relating reputation to the position in rankings was not tested
previously in this scale of image but was considered appropriate in the light of recent
research [
46
,
47
], which establishes its relevance and links these concepts in the univer-
sity domain.

Educ. Sci. 2022,12, 19 6 of 13
As for the formulation of the questions, each one is presented as a statement, and
respondents are asked to show their level of agreement or disagreement on a 7-point Likert
scale (1 = complete disagreement, 7 = complete agreement).
3.3. Data Collection and Analysis Procedure
To test the hypotheses proposed, the factor analysis method was applied with data
arising from a survey made of Andalusian adults. Not having a generally accepted model
to measure the image of the private university, but a generic brand this could be likened
to, led us to divide analysis of the issue into four phases, as presented in Figure 1. Phase 1
involves adapting the items of the standardized image measurement model and carrying
out the corresponding pre-test. In Phase 2, the validity and significance of the items initially
proposed are explored. In Phase 3, the standardized model is confirmed in its version with
the three original dimensions. In Phase 4, the validity of the result obtained is contrasted.
The difference between the third and fourth phases is that while in the third we admit
the possibility of improving the initial estimation by modifying the model, in the fourth,
there is only validation of the result of the first. This strategy of validation meant dividing
the sample.
Educ. Sci. 2022, 12, x FOR PEER REVIEW 6 of 13
3.3. Data Collection and Analysis Procedure
To test the hypotheses proposed, the factor analysis method was applied with data
arising from a survey made of Andalusian adults. Not having a generally accepted model
to measure the image of the private university, but a generic brand this could be likened
to, led us to divide analysis of the issue into four phases, as presented in Figure 1. Phase
1 involves adapting the items of the standardized image measurement model and carry-
ing out the corresponding pre-test. In Phase 2, the validity and significance of the items
initially proposed are explored. In Phase 3, the standardized model is confirmed in its
version with the three original dimensions. In Phase 4, the validity of the result obtained
is contrasted. The difference between the third and fourth phases is that while in the third
we admit the possibility of improving the initial estimation by modifying the model, in
the fourth, there is only validation of the result of the first. This strategy of validation
meant dividing the sample.
Figure 1. Phases, procedures applied, and sample used.
4. Results
4.1. Phase 1. Pretest of the Questionnaire
As already mentioned, 50 valid questionnaires were used in a pre-test. Validity and
reliability were checked through analysis of the Cronbach alpha coefficient (α). A reliabil-
ity value of around 0.70 is considered acceptable [48]. The figure obtained here was α =
0.819, implying validation of the initial questionnaire.
4.2. Phase 2. Initial Validation through EFA
In exploratory factor analysis, examination of the data is a necessary preliminary
step, to allow better prediction and more accurate evaluation of the dimensional nature.
Otherwise, atypical cases, extreme answers, can unduly influence the result of a multivar-
iate analysis [49]. Measuring the Mahalanobis distance to the centroid of the distribution
indicated the presence of 22 outliers, which we decided to eliminate [50], leaving a final
sample formed of 728 observations.
Although various procedures are available in statistical programs (such as SPSS), in
practice two methods are most used an exploratory factor analysis (EFA): Principal Axes
and Principal Components. The Principal Axes method only contemplates the variance
Phases, techniques
and sample
1. Adaptation of the
items and preparation
of the initial
questionnaire
Pretest sample:
50 individuals
2. Testing and results by
means of exploratory
factor analysis
Total sample:
728 individuals
3. Contrast of the image
measurement model in its
formulation with three
dimensions. Analysis by
confirmatory factor analysis
(CFA)
One half of the sample:
363 individuals
4. Validation of the
image measurement
model obtained in the
3rd phase by CFA
Other half of the
sample:
365 individuals
Figure 1. Phases, procedures applied, and sample used.
4. Results
4.1. Phase 1. Pretest of the Questionnaire
As already mentioned, 50 valid questionnaires were used in a pre-test. Validity and
reliability were checked through analysis of the Cronbach alpha coefficient (
α
). A reliability
value of around 0.70 is considered acceptable [
48
]. The figure obtained here was
α
= 0.819,
implying validation of the initial questionnaire.
4.2. Phase 2. Initial Validation through EFA
In exploratory factor analysis, examination of the data is a necessary preliminary
step, to allow better prediction and more accurate evaluation of the dimensional nature.
Otherwise, atypical cases, extreme answers, can unduly influence the result of a multivariate
analysis [
49
]. Measuring the Mahalanobis distance to the centroid of the distribution indicated
the presence of 22 outliers, which we decided to eliminate [
50
], leaving a final sample
formed of 728 observations.

Educ. Sci. 2022,12, 19 7 of 13
Although various procedures are available in statistical programs (such as SPSS), in
practice two methods are most used an exploratory factor analysis (EFA): Principal Axes
and Principal Components. The Principal Axes method only contemplates the variance the
variables have in common or co-variance, excluding specific variance and that which can be
attributed to measurement. In this case, Principal Components was chosen, since the aim of
this analysis, more than a prediction, is the detection of latent dimensions (here, the three
blocks forming the questionnaire, functional image, affective image, and reputation) [49].
The results suggest that three factors should be chosen, as from the fourth factor
onwards the associated eigenvalues are far below 1. Besides, the percentage of explained
variance with the three factors chosen reaches 69.935%, a high percentage, which indicates
acceptance of the three factors.
The next step is to contrast the coincidence of their significance. One way to approach
factor content is rotation, which consists of their re-organization, reducing as much as
possible the variance of each item shared by different factors, at the expense of a simultane-
ous reduction in the explained variance. After rotating the model, the explained variance
is 61.431%.
The factor loadings after rotation show that the 5 items relating to functional image
form one factor and that the 4 relating to reputation form another. However, in the case of
affective image, two of its items are not significant: variable a3. “Enrolling in the private
university says something about the class of person you are”, and variable a5. “The private
university has a differentiated ideological component”. With these results, affective image
will be studied from now on with 3 items, due to eliminating the variables about the
association between the type of person and the private nature of the university, and about
the ideological component.
4.3. Division of the Sample
This study proposes a previous procedure to CFA of dividing the sample, which
is common in some areas of quantitative research. This assumes validation of a scale
principally through a procedure of dividing the total sample (728 questionnaires) in two
homogeneous sub-samples: one for training (confirmation of the scale, corresponding to
Phase 3) and the other for validation (ratification of the results, corresponding to Phase 4).
The reason for this strategy is that it would not be advisable to use the same observations
for construction and validation, as if to evaluate the model we use the same observations
as to construct it, we would probably obtain a good result, due to the fact of the parameters
of the model being estimated with the same sample as for subsequent validation.
Among the few methods available for division of the sample, the one developed
by [
51
] was chosen. This is an extension of the Wilcoxon-Mann-Whitney test of equality of
means without requirements of normality. The contrast is applied to successive divisions
of the sample in two sub-sets, obtained randomly, with the similarity being evaluated
through a statistic of known approximate distribution (specifically chi-squared) until the
sub-samples obtained are statistically homogeneous for the chosen level of significance
(5% in this case). Applying this procedure to the initial sample results in two sub-samples
of 363 and 365 respectively, whose associated p-value calculated is 0.948, very close to
one, which leads us to accept the division with strong guarantees of their similarity. As
explained previously, the first sub-sample will be used to confirm the model (Phase 3) and
the second to validate it (Phase 4).
4.4. Phases 3 and 4. Estimation and Validation of the Model through CFA
Both the existing literature on the factors measuring image and the previous results
indicate it may not be completely clear that there is an affective factor independent of
reputation and functional elements. Setting out from the result obtained from EFA, the
third phase deals with confirming this, even improving it, through slight modifications to its
structure. In the fourth phase, the resulting model is put to test to check its statistic validity.

Educ. Sci. 2022,12, 19 8 of 13
4.5. Phase 3. Estimation of the Model
To estimate and validate the CFA models we used EQS 6.1 software, choosing the
maximum likelihood (ML) method to estimate the parameters, aiming to find the esti-
mates that make the occurrence of the sample obtained most probable, maximizing its
likelihood [
52
]. In addition, this is a very robust estimation method that can be applied
when there is no condition of univariate and multivariate normality of the data (as in this
case) without serious losses of the theoretical properties of the estimates obtained [
49
]. The
most important aspect of the results obtained is evaluation of the model’s goodness of fit.
In each case, we obtained some of the usual indices in this type of analysis, namely the
comparative fit index (CFI), the normed fit index (NFI), the non-normed fit index (NNFI),
and the root mean square error of approximation (RMSEA).
The CFI is a measure that compares the fit obtained for the estimated model with what
would be obtained on the assumption of no relation between the variables used. Its values
range between 0 (no adjustment of the model to the data) and 1 (perfect adjustment of
the model), with 0.9 being the appropriate threshold value [
53
]. The NFI and NNFI are
similar measures, which situate the estimated model on a scale between the extremes of the
nil model and perfect fit, differing in the fact of considering or not the models’ degrees of
freedom. Both have values between 0 and 1, considering the models with the highest values
in these indices as the most adjusted. Finally, RMSEA, as a measure of the magnitude of the
errors committed, should be interpreted in the opposite direction, with the most suitable
models being those where this indicator has a lower value. A common reference is that its
value should not be above 0.06 [53].
Carrying out the confirmatory procedure with the first sub-sample (363) and calculat-
ing the adjustment indicators of the model to measure private university image with the
three factors (functional, affective and reputational) and the items determined in the ex-
ploratory Phase 2, confirmation of the model did not turn out, as expected, to be completely
satisfactory (CFI = 0.919, NNFI = 0.895, NFI = 0.905 and RMSEA = 0.119). However, analysis
of the Lagrange multipliers indicated it was possible to improve the model by modifying its
structure slightly.
Those modifications consisted of allowing the functional factor (F) to share one in-
dicator (r3, opinion about consolidation in the market) with reputation (R) and another
with the affective factor (f5, quality-price relation). Finally, everything seemed to indicate
that variable f4 (private universities present characteristics that state ones do not have)
could be linked to the three factors. We consider this possibility is equivalent to stating that
the variable is not appropriate to identify the elements defining university image, since it
should be associated with all three and is not really a characteristic of any. Therefore, we
decided to eliminate it from the model.
The indicators for the model recalculated with these corrections (CFI = 0.965, NNFI = 0.950,
NFI = 0.952 and RMSEA = 0.086), much more accurate than the initial ones, together with
the results’ coherence with theory, led us to consider it definitive, as shown in Figure 2.
Educ. Sci. 2022, 12, x FOR PEER REVIEW 8 of 13
4.5. Phase 3. Estimation of the Model
To estimate and validate the CFA models we used EQS 6.1 software, choosing the
maximum likelihood (ML) method to estimate the parameters, aiming to find the esti-
mates that make the occurrence of the sample obtained most probable, maximizing its
likelihood [52]. In addition, this is a very robust estimation method that can be applied
when there is no condition of univariate and multivariate normality of the data (as in this
case) without serious losses of the theoretical properties of the estimates obtained [49].
The most important aspect of the results obtained is evaluation of the model’s goodness
of fit. In each case, we obtained some of the usual indices in this type of analysis, namely
the comparative fit index (CFI), the normed fit index (NFI), the non-normed fit index
(NNFI), and the root mean square error of approximation (RMSEA).
The CFI is a measure that compares the fit obtained for the estimated model with
what would be obtained on the assumption of no relation between the variables used. Its
values range between 0 (no adjustment of the model to the data) and 1 (perfect adjustment
of the model), with 0.9 being the appropriate threshold value [53]. The NFI and NNFI are
similar measures, which situate the estimated model on a scale between the extremes of
the nil model and perfect fit, differing in the fact of considering or not the models’ degrees
of freedom. Both have values between 0 and 1, considering the models with the highest
values in these indices as the most adjusted. Finally, RMSEA, as a measure of the magni-
tude of the errors committed, should be interpreted in the opposite direction, with the
most suitable models being those where this indicator has a lower value. A common ref-
erence is that its value should not be above 0.06 [53].
Carrying out the confirmatory procedure with the first sub-sample (363) and calcu-
lating the adjustment indicators of the model to measure private university image with
the three factors (functional, affective and reputational) and the items determined in the
exploratory Phase 2, confirmation of the model did not turn out, as expected, to be com-
pletely satisfactory (CFI=0.919, NNFI=0.895, NFI=0.905 and RMSEA=0.119). However,
analysis of the Lagrange multipliers indicated it was possible to improve the model by mod-
ifying its structure slightly.
Those modifications consisted of allowing the functional factor (F) to share one indi-
cator (r3, opinion about consolidation in the market) with reputation (R) and another with
the affective factor (f5, quality-price relation). Finally, everything seemed to indicate that
variable f4 (private universities present characteristics that state ones do not have) could
be linked to the three factors. We consider this possibility is equivalent to stating that the
variable is not appropriate to identify the elements defining university image, since it
should be associated with all three and is not really a characteristic of any. Therefore, we
decided to eliminate it from the model.
The indicators for the model recalculated with these corrections (CFI = 0.965, NNFI =
0.950, NFI = 0.952 and RMSEA = 0.086), much more accurate than the initial ones, together
with the results’ coherence with theory, led us to consider it definitive, as shown in Figure
2.
Figure 2. CFA results of the first sub-sample.
Figure 2. CFA results of the first sub-sample.

Educ. Sci. 2022,12, 19 9 of 13
4.6. Phase 4. Validation of the Model
The model estimated and modified in Phase 3 was subject to evaluation using the
second sub-sample (365 questionnaires) with no option for change. It is a question of
determining whether the goodness-of-fit achieved in the previous phase was inherent to
the model, or if this had been adapted exclusively to the sample with which it had been
obtained (overtraining). In both cases, as detailed below, the result was satisfactory.
The expectations for the 4th phase, on the assumption of the model’s validity, are of
stability, in that the model proposed, without new modifications, should be acceptable.
Moreover, if the structure detected during the first phase is correct, not only the global fit
measurements but also the factor loadings should be similar in both estimates.
Tables 3and 4present, respectively, the goodness-of-fit indicators and the values
estimated for the model’s parameters in both phases.
Table 3. Global measurements.
Phase 3 Phase 4
CFI 0.965 0.963
NNFI 0.950 0.947
NFI 0.952 0.951
RMSEA 0.086 0.088
Table 4. Estimated parameters.
Factor Variable Phase 3 Phase 4
F
f1 0.865 ** 0.883 **
f2 0.776 ** 0.777 **
f3 0.816 ** 0.838 **
f5 0.286 ** 0.479 **
r3 0.176 ** 0.358 **
R
r1 0.852 ** 0.854 **
r2 0.892 ** 0.903 **
r3 0.835 ** 0.827 **
r4 0.754 ** 0.537 **
A
f5 0.53 ** 0.341 **
a1 0.82 ** 0.785 **
a2 0.755 ** 0.769 **
a4 0.772 ** 0.732 **
Cov(F,R) 0.631 0.708
Cov(F,A) 0.766 0.822
Cov(R,A) 0.901 0.926
** Parameter statistically significant at 1%.
Both Tables 3and 4reveal that, although the structure was estimated from a set of
values and partly adapted to them, the degree of reliability obtained is not the result of
that adaptation, as the indicators are confirmed to be practically identical with the different
samples. Therefore, the results obtained in Phase 3 with the functional, reputational and
affective factors, partially modified with regard to the model by Martinez et al. (2004) [
24
],
are not the result of the sample used for the estimate.

Educ. Sci. 2022,12, 19 10 of 13
5. Discussion and Conclusions
5.1. Discussion
Image and reputation are key assets with which institutions can differentiate their
products from those of their rivals, and so it would be useful for managers to have appro-
priate measuring techniques for use in distinct contexts.
When reviewing the specifics about image and reputation, it is clear that, being
concepts generally used in empirical work in management areas, a consensus has not been
reached regarding their definitions. Although one stream argues that the concepts are
similar, most of the studies reviewed focus on their divergence, defending the need to
distinguish between image and reputation. Even within this stance of divergence, it is
worth going deeper into the relation between both constructs.
Following the theoretical review, despite nuances, the definition of the image pro-
posed [
17
] seems best to us, as it can be generalized to any product and any stakeholder.
Therefore, image would be defined as the sum of an individual’s beliefs, attitudes, experi-
ences, sentiments, or information regarding an object. In relation to reputation, it is more
difficult to find work going deeper into its definition, and the definitions reviewed are less
homogeneous than in the case of image. Even so, there is a predominance of authors who
consider reputation as the assessment or global attitude of a company or brand’s behavior;
and some consider that other objects, such as goods or services, have a reputation [19].
We also found that when designing instruments to measure image, there is a certain
gap between what is established theoretically as a possible target of image measurement
(product, brand, company, country, etc.) and the empirical contributions published, which
are usually greatly focused on similar subjects of study, with brand image being measured
most frequently. It was very difficult to find work analyzing a generic product, with no
brand, as is proposed here. In our view, this level of analysis would be extremely valuable
for subsequent decision-making affecting brand image, or even guiding the first stages in
launching a new product.
This study used an adaptation of the brand image measurement scale proposed by
Martinez et al. (2004) [
24
], applied to citizens’ perception of the private university. In
this model, perceptions of a brand are grouped around three basic dimensions. Firstly,
individual’s value functional attributes, such as quality and perception of the price. In
addition, affective attributes, which are linked to elements of a symbolic and emotional
character. Finally, reputation is included. This model was chosen for three reasons:
1. the universality of the measurement scale proposed;
2. its validation through other work in different sectors;
3. considering reputation as a necessary factor in measuring image.
Exploratory factor analysis led to the conclusion that it is possible to measure the
image of a product in individuals’ minds, without referring to any brand, and that this is
a multi-dimensional construct, with cognitive and emotional components. Therefore, the
first hypothesis of this study is accepted.
The first hypothesis, seeking to contribute more empirical evidence about the rela-
tion between image and reputation, is also accepted, since through confirmatory factor
analysis, reputation is found to be a significant factor at the time of measuring a brandless
product’s image.
The second and final hypothesis of this study—the best formulation to measure the
image of the “private university” product is obtained by applying a standardized multi-
factor scale with three dimensions: functional image, affective image, and reputation—is
also accepted.
Concerning the overall results, there are similarities in relation to the results of the
study by Martínez et al. [
25
], since both models have similar goodness-of-fit indicators,
which corroborates the general validity of their theoretical proposal.
Nevertheless, it is worth making some changes about the initial model. In the first
place, it is proposed to eliminate from the scale one of the variables included in the block of
functional aspects—private universities present characteristics that state ones do not have -,

Educ. Sci. 2022,12, 19 11 of 13
a result that coincides with the last study published by its original authors [
27
]. Secondly, a
change is proposed so that the price variable loads not only on the functional factor but also
on the affective factor, which is logical since perception of the price involves both affective
and functional elements.
Furthermore, it is proposed to incorporate an additional variable to measure the
reputation of a university, i.e., the position in university rankings, something of great
relevance in the sphere of higher education. The analyses confirm that the position in
rankings is a significant variable in measuring university image, also being connected to
the functional factor as well as the reputational one, although with much greater weight in
the latter.
5.2. Conclusions
The relevance of the product category concept is explicit in the positioning concepts
of image measurement, but very little research has been done on it. In the analysis of an
organization’s image, it is assumed that potential customer’s group products hierarchically
into different levels of specificity. For example, if an adult individual were to consider
which path to follow in his or her studies, the possible products would initially be grouped
into product categories (i.e., university versus vocational training), into product types (i.e.,
private university versus public university) and finally, by brand. However, most published
empirical studies are concerned with measuring the brand image of a particular university.
A standardized multi-factor scale with three dimensions- functional image, affective
image, and reputation- has proved valid for measuring the image of the private university,
in general, without referring to any brand.
The results obtained suggest the existence of a differentiated image of private univer-
sities compared to public universities, which, to a certain extent, justifies their coexistence
with state universities. This work offers some keys to private universities in terms of
management and differentiation: if they work on functional aspects, it is convenient to
allocate resources to improve the quality of education and research, without neglecting the
price charged to students; if they focus attention on the management of emotional aspects,
they must find a clearly differentiated brand personality compared to public universities
and, finally, if they seek to activate their reputation, they must improve their positions in
international rankings, among other aspects.
However, the results of this research, given the type of sampling employed, should be
confirmed by other studies that extend the sample to a broader scope. It would also be of
great interest to compare the results between two or more different countries.
In this line, and for future lines of research, it would be of interest to contrast whether
there are significant differences in the measurement of the image of private universities
according to the profile of the respondent.
In the same line of research, it would be very interesting to incorporate image as a
second-order construct, which would allow evaluating the weight of each of the three factors
(functional, affective, and reputational) in its construction, in particular, that of reputation.
Finally, mention must be made of the methodological contribution as regards the
management of the sample, both in its sub-division in two parts used to estimate and then
validate the model (a not common practice in this type of study, although theoretically very
recommendable) and the supervised way in which this sub-division was carried out to
ensure the statistical homogeneity of both fractions.
Author Contributions:
Conceptualization, B.G.-V. and P.A.-P.; Formal analysis, M.C.-R.; Method-
ology, M.C.-R.; Writing—original draft, B.G.-V., P.A.-P. All authors have read and agreed to the
published version of the manuscript.
Funding: This research received no external funding.

Educ. Sci. 2022,12, 19 12 of 13
Institutional Review Board Statement:
Ethical review and approval were waived for this study, due
to a market research company external to Loyola Andalucía University being used. This company
“IDDEALIA Consulting” has the ISO 26362:2009 certification that guarantees the ethical aspects in
this type of opinion panel services.
Informed Consent Statement:
Informed consent was obtained from all subjects involved in the study.
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
Liu, Y.; Green, A.; Pensiero, N. Expansion of Higher Education and Inequality of Opportunities: A Cross-National Analysis.
J. High. Educ. Policy Manag. 2016,38, 242–263. [CrossRef]
2.
Khoshtaria, T.; Datuashvili, D.; Matin, A. The Impact of Brand Equity Dimensions on University Reputation: An Empirical Study
of Georgian Higher Education. J. Mark. High. Educ. 2020,30, 239–255. [CrossRef]
3.
Farhat, K.; Mokhtar, S.S.M.; Salleh, S.B.M. Role of Brand Experience and Brand Affect in Creating Brand Engagement: A Case of
Higher Education Institutions (HEIs). J. Mark. High. Educ. 2021,31, 107–135. [CrossRef]
4.
Sung, M.; Yang, S.-U. Toward the Model of University Image: The Influence of Brand Personality, External Prestige, and
Reputation. J. Public Relat. Res. 2008,20, 357–376. [CrossRef]
5.
Çetin, R. Planning and Implementing Institutional Image and Promoting Academic Programs in Higher Education. J. Mark. High.
Educ. 2004,13, 57–75. [CrossRef]
6.
Lafuente Ruiz de Sabando, A.; Forcada Sainz, F.J.; Zorrilla Calvo, M.P. The University Image: A Model of Overall Image and
Stakeholder Perspectives. Imagen Univ. Modelo Imagen Glob. Perspect. Grupos Interés2018,19, 63–86. [CrossRef]
7.
Del-Castillo-Feito, C.; Blanco-González, A.; González-Vázquez, E. The Relationship between Image and Reputation in the Spanish
Public University. Eur. Res. Manag. Bus. Econ. 2019,25, 87–92. [CrossRef]
8.
Irfan, A.; Sulaiman, Z.; Qureshi, M.I. Student’s Perceived University Image Is an Antecedent of University Reputation. Int. J.
Psychosoc. Rehabil. 2020,24, 650–663. [CrossRef]
9.
Iqbal, M.J.; Rasli, A.B.M.; Ibn-e-Hassan Rasli, A.B.M. University Branding: A Myth or a Reality. Pak. J. Commer. Soc. Sci. PJCSS
2012,6, 168–184.
10.
Nguyen, N.; Leblanc, G. Corporate Image and Corporate Reputation in Customers’ Retention Decisions in Services. J. Retail.
Consum. Serv. 2001,8, 227–236. [CrossRef]
11.
Walsh, C.; Moorhouse, J.; Dunnett, A.; Barry, C. University Choice: Which Attributes Matter When You Are Paying the Full Price?
Int. J. Consum. Stud. 2015,39, 670–681. [CrossRef]
12.
Aghaz, A.; Hashemi, A.; Sharifi Atashgah, M.S. Factors Contributing to University Image: The Postgraduate Students’ Points of
View. J. Mark. High. Educ. 2015,25, 104–126. [CrossRef]
13.
Massoud, H.K.; Ayoubi, R.M. Do Flexible Admission Systems Affect Student Enrollment? Evidence from UK Universities. J. Mark.
High. Educ. 2019,29, 84–101. [CrossRef]
14.
Pérez, A.; Bosque, I.R. del. Identidad, imagen y reputación de la empresa: Integración de propuestas teóricas para una gestión
exitosa. Cuad. Gest. 2014,14, 97–126. [CrossRef]
15.
Stern, B.; Zinkhan, G.M.; Jaju, A. Marketing Images: Construct Definition, Measurement Issues, and Theory Development. Mark.
Theory 2001,1, 201–224. [CrossRef]
16.
Balmer, J.M.T. Corporate Identity, Corporate Branding and Corporate Marketing—Seeing through the Fog. Eur. J. Mark.
2001
,35,
248–291. [CrossRef]
17. Curras Perez, R. Identidad e imagen corporativas: Revisión conceptual e interrelación. Teoría Praxis 2010,7, 9–34. [CrossRef]
18. Lee, J.L.; James, J.D.; Kim, Y.K. A Reconceptualization of Brand Image. Int. J. Bus. Adm. 2014,5, 1–11. [CrossRef]
19.
Martínez, I.; Olmedo, I. Revisión teórica de la reputación en el entorno empresarial. Cuad. Econ. Dir. Empresa
2010
,13, 59–77.
[CrossRef]
20. Kazoleas, D.; Kim, Y.; Anne Moffitt, M. Institutional Image: A Case Study. Corp. Commun. Int. J. 2001,6, 205–216. [CrossRef]
21. Barich, H.; Kotler, P. A Framework for Marketing Image Management. MITSloan 1991,32, 94–104.
22. Keller, K.L. Conceptualizing, Measuring, and Managing Customer-Based Brand Equity. J. Mark. 1993,57, 1–22. [CrossRef]
23.
Cervera, A.; Schlesinger, W.; Mesta, M.Á.; Sánchez, R. Medición de la imagen de la universidad y sus efectos sobre la identificación
y lealtad del egresado: Una aproximación desde el modelo de beerli y díaz (2003). Rev. Esp. Investig. Mark. ESIC
2012
,16, 7–29.
[CrossRef]
24.
Martínez, E.M.; Montaner, T.M.; Pina, J.M.P. Propuesta de una metodología. Medición de la imagen de marca. Un estudio
explorativo. ESIC Mark. 2004,35, 199–216.
25.
Martínez, E.; Gutiérrez, T.M.; Pérez, J.M.P. Propuesta de medición de la imagen de marca: Un análisis aplicado a las extensiones
de marca. RAE Rev. Astur. Econ. 2005,33, 89–112.
26.
Matínez, E.M.; Montaner, T.M.; Pina, J.M.P. Estrategia de promoción e imagen de marca: Influencia del tipo de promoción, de la
notoriedad de la marca y de la congruencia de beneficios. Rev. Esp. Investig. Mark. 2007,11, 27–52.
27.
Buil, I.B.; Martínez, E.M.; Pina, J.M.P. Un modelo de evaluación de las extensiones de marca de productos y de servicios. Rev. Eur.
Dir. Econ. Empresa 2008,17, 167–186.

Educ. Sci. 2022,12, 19 13 of 13
28. Gardner, B.B.; Levy, S.J. The Product and the Brand. Harv. Bus. Rev. 1955,33, 33–39.
29. Martineau, P. Sharper Focus for the Corporate Image. Harv. Bus. Rev. 1958,36, 49–58.
30.
Arpan, L.M.; Raney, A.A.; Zivnuska, S. A Cognitive Approach to Understanding University Image. Corp. Commun. Int. J.
2003
,8,
97–113. [CrossRef]
31. Kennedy, S.H. Nurturing Corporate Images. Eur. J. Mark. 1977,11, 119–164. [CrossRef]
32.
Williams, R.J.; Barrett, J.D. Corporate Philanthropy, Criminal Activity, and Firm Reputation: Is There a Link? J. Bus. Ethics
2000
,
26, 341–350. [CrossRef]
33.
Fombrun, C.J.; Gardberg, N.A.; Barnett, M.L. Opportunity Platforms and Safety Nets: Corporate Citizenship and Reputational Risk;
SSRN Scholarly Paper ID 1088404; Social Science Research Network: Rochester, NY, USA, 2008.
34.
Iglesias, S.; Calvo, A.; López, V.A. Una aproximación empírica al concepto de reputación. In Gestión Científica Empresarial: Temas
de Investigación Actuales; Netbiblo: La Coruña, Spain, 2003; pp. 87–104, ISBN 84-9745-051-5.
35.
Ruiz, B.; Gutierrez-Broncano, S.; Esteban, Á. Desarrollo de Un Concepto de Reputación Corporativa Adaptado a Las Necesidades
de La Gestión Empresarial. Strategy Manag. Rev. 2012,3, 9–31.
36.
Orozco-Toro, J.A.; Ferré-Pavia, C.C. Los índices de reputación corporativa y su aplicación en las empresas de comunicación.
In Comunicaciói Risc: III Congrés Internacional AssociacióEspanyola d’Investigacióde la Comunicació; Universitat Rovira i Virgili:
Tarragona, Spain, 2012; p. 42, ISBN 978-84-615-5678-6.
37. Chun, R. Corporate Reputation: Meaning and Measurement. Int. J. Manag. Rev. 2005,7, 91–109. [CrossRef]
38.
Ingenhoff, D.; Fuhrer, T. Positioning and Differentiation by Using Brand Personality Attributes: Do Mission and Vision Statements
Contribute to Building a Unique Corporate Identity? Corp. Commun. Int. J. 2010,15, 83–101. [CrossRef]
39.
Beerli Palacio, A.; Díaz Meneses, G.; Pérez Pérez, P.J. The Configuration of the University Image and Its Relationship with the
Satisfaction of Students. J. Educ. Adm. 2002,40, 486–505. [CrossRef]
40.
Flavián, C.; Torres, E.; Guinalíu, M. Corporate Image Measurement: A Further Problem for the Tangibilization of Internet Banking
Services. Int. J. Bank Mark. 2004,22, 366–384. [CrossRef]
41. Gómez, M. El Turismo Enológico Desde la Perspectiva de la Oferta; Editorial Universitaria Ramón Areces: Madrid, Spain, 2011.
42.
Almeida, M.G.C.; Coelho, A.M. El impacto de la reputación en el desempeño de la organización en la perspectiva de los miembros
de las cooperativas. ESIC Mark. 2015,46. [CrossRef]
43.
Luque-Martínez, T.; Barrio-García, S. Modelling University Image: The Teaching Staff Viewpoint. Public Relat. Rev.
2009
,35,
325–327. [CrossRef]
44.
Azoury, N.; Daou, L.; Khoury, C.E. University Image and Its Relationship to Student Satisfaction- Case of the Middle Eastern
Private Business Schools. Int. Strateg. Manag. Rev. 2014,2, 1–8. [CrossRef]
45.
Patlán, J.; Martínez, E. Evaluación de La Imagen Organizacional Universitaria En Una Institución de Educación Superior. Contad.
Adm. 2017,62, 105–122. [CrossRef]
46.
Safón, V. Inter-Ranking Reputational Effects: An Analysis of the Academic Ranking of World Universities (ARWU) and the Times
Higher Education World University Rankings (THE) Reputational Relationship. Scientometrics 2019,121, 897–915. [CrossRef]
47.
Luque-Martínez, T.; Faraoni, N.; Doña-Toledo, L. Meta-ranking de universidades. Posicionamiento de las universidades españolas.
Rev. Esp. Doc. Científica 2018,41, e198. [CrossRef]
48.
Nunnally, J.C. An Overview of Psychological Measurement. In Clinical Diagnosis of Mental Disorders: A Handbook; Wolman, B.B.,
Ed.; Springer: Boston, MA, USA, 1978; pp. 97–146. [CrossRef]
49.
Hair, J.F.; Black, W.C.; Balbin, B.J.; Tatham, R.L. Multivariate Data Analysis, 6th ed.; Pearson Prentice-Hall: Upper Saddle River NJ,
USA, 2006; ISBN 0130329290.
50. Dancey, C.P.; Reidy, J. Statistics without Maths for Psychology; Pearson Education: London, UK, 2007; ISBN 0273726021.
51.
Lung-Yut-Fong, A.; Lévy-Leduc, C.; Cappé, O. Homogeneity and change-point detection tests for multivariate data using rank
statistics. J. SociétéFr. Stat. 2015,156, 133–162.
52.
Hoyle, R.H. 16—Confirmatory Factor Analysis. In Handbook of Applied Multivariate Statistics and Mathematical Modeling; Tinsley,
H.E.A., Brown, S.D., Eds.; Academic Press: San Diego, CA, USA, 2000; pp. 465–497. [CrossRef]
53.
Hu, L.; Bentler, P.M. Cutoff Criteria for Fit Indexes in Covariance Structure Analysis: Conventional Criteria versus New
Alternatives. Struct. Equ. Model. Multidiscip. J. 1999,6, 1–55. [CrossRef]