SciELO - Scientific Electronic Library Online

 
vol.36 número3USE OF CONTINUED ITERATION ON THE REDUCTION OF ITERATIONS OF THE INTERIOR POINT METHODA NON-ARCHIMEDEAN DEA MODEL TO ASSESS GROUP COMPARISONS índice de autoresíndice de assuntospesquisa de artigos
Home Pagelista alfabética de periódicos  

Serviços Personalizados

Journal

Artigo

Indicadores

Links relacionados

Compartilhar


Pesquisa Operacional

versão impressa ISSN 0101-7438versão On-line ISSN 1678-5142

Pesqui. Oper. vol.36 no.3 Rio de Janeiro set./dez. 2016

http://dx.doi.org/10.1590/0101-7438.2016.036.03.0503 

Articles

EMPIRICAL COMPARISON OF THE MULTIDIMENSIONAL MODELS OF ITEM RESPONSE THEORY IN E-COMMERCE

Rafael Tezza1  * 

Antonio Cezar Bornia2 

Débora Spenassato3 

Andréa Cristina Trierweiller4 

1Centro de Ciências da Administração e Socioeconômicas, Universidade do Estado de Santa Catarina, Av. Madre Benvenuta, 2037, 88035-001 Itacorubí, Florianópolis, SC, Brasil. E-mail: rafael.tezza@udesc.br

2Departamento de Engenharia de Produção e Sistemas, Universidade Federal de Santa Catarina, Caixa Postal 476, Campus Universitário, 88010-970 Florianópolis, SC, Brasil. E-mail: cezar.bornia@ufsc.br

3Instituto de Matemática, Estatística e Física, Universidade Federal do Rio Grande, Campus Carreiros, Av. Itália Km 8, 96201-900 Rio Grande, RS, Brasil. E-mail: debospenassato@gmail.com

4Departamento de Tecnologias da Informação e Comunicação, Universidade Federal de Santa Catarina, Rua Prefeito Walter Belinzoni, 219, 88900-017 Araranguá, SC, Brasil. E-mail: andreatri@gmail.com


ABSTRACT

The measurement of latent traits within the organizational field such as quality, effectiveness and learning has been conducted in several formats using a wide variety of quantitative methods, including Item Response Theory that consistently increased in organizational studies. The purpose of this article is to compare the hierarchical and non-hierarchical structures of three multidimensional models of ItemResponse Theory, based on the interface quality measurement in e-commerce sites. We compared the multiple unidimensional, compensatory multidimensional and bifactorial models, and also elaborated and applied 75 items in a sample of 441 e-commerce websites. As a result, we conducted a discussion of the latent construct, the quality in e-commerce and its multidimensional configuration to adjust and compare three multidimensional models.

Keywords: multidimensional Item Response Theory; latent trait; measurement; websites; organizations

1 INTRODUCTION

The quality of websites is a complex matter and sometimes difficult to be measured directly. Multiple features should be discussed from various angles that exceed the simple to use interface technical questions, including issues related to aesthetics, reliability, interactivity and other factors that involve non-technical issues. According to Nielsen (2007), technical issues such as usability continue to be necessary, but not a sufficient condition that provides the quality demanded for a website, in all contexts.

A variety of scales using various dimensions were suggested to measure the quality of service within electronic commerce (known as quality of e-service or e-quality). Many of them influenced the quality of the website itself (Bernado et al. 2012).

Existing studies discuss that website quality has a multidimensional construct, yet a consensus has not been reached on what dimensions are most relevant and what is the multidimensional nature or configuration (Zeithaml et al., 2002; Parasuraman et al., 2005; Fassnacht & Koese, 2006).

Item Response Theory (IRT) proceeds from the efforts to understand the relations between the dimensions and the creation of a comparable scale. IRT is composed of a set of probabilistic models that relate the latent trait of a respondent (θ), something that cannot be directly measured, to the probability that this responds to an item within a certain category (Lord, 1980).

The application of IRT requires some initial presumptions such as the determination of IRT model that is suitable to the set of data to be analyzed and the objectives of the study. In general, the selection of a satisfactory IRT model is based partially on how acceptable the model is to the data. Whenever the data do not suit the model, the properties of IRT will not be assured, this could happen in a case of invariance of the parameters in a population study, for instance. (Hambleton et al., 1991).

Choosing a specific multidimensional model depends mainly on the nature of the latent dimensions, the items and the how they are associated within the test. Different models of Multidimensional Item Response Theory (MIRT) assume different statistical relations between the latent dimensions and the performance of the respondent (Hartig & Hohler, 2009). Furthermore, it is important to understand the multidimensional structures involved in an analysis, which are the relations between latent dimensions and the items and the relations between the latent dimensions and the respondents.

From these structures, emerged the hierarchical multidimensional models that have a general factor and a specific factor to their superior dimension, and those that are not hierarchical, which have various dimensions that may or may not be correlated to clarify a general latent trait.

The purpose of this article is to compare three multidimensional models of IRT: hierarchic (bifactorial) and non-hierarchical (multiple unidimensional and compensatory multidimensional) to measure the quality of commercial websites. This approach is justified by the fact that the quality of websites has been treated as a multidimensional issue and Item Response Theory, both in its simple unidimensional modeling (e.g. Sijtsma, 2011; Tezza et al., 2001), and its more complex multidimensional modeling (e.g. Bernini et al., 2014) are presented as capable tools of measuring latent traits, revealing the relations between respondents and the items of a study, to position them in a single, comparable and valid scale for broader situations, to not depend directly from the group of respondents, resulting in a better understanding of the relationship between the dimensions and respondents.

2 THEORETICAL REFERENCES

2.1 Multidimensional Item Response Theory

The mathematical foundation of IRT is a function that relates the probability of a person responding to an item in a specific manner to the standing of that person on trait that item is measuring (Ostini & Nering, 2006). One of the underlying assumptions of IRT is that examinees are all using the same skill or same composition of multiple skills to respond to each of the items in a test. When item response data do not satisfy the unidimensionality assumption, Multidimensional Item Response Theory (MIRT) should be used to model the item-examinee interaction. MIRT enables modeling the interaction of items that are capable of discriminating betweenlevels of several different abilities and examinees that vary in their proficiencies in these abilities (Ackerman, 1994; Reckase, 2009).

The use of MIRT models to handle measuring problems in large-scale educational evaluation has been conducted since the early 1990s (Ackerman, 1992; Camilli, 1992; Embretson, 1991; Glas, 1992; Oshima & Miller, 1992; Reckase & McKinley, 1991). Nevertheless, according to Adams et al. (1997), Hartig & Höhler (2008), Rauch & Hartig (2010), the application of models in a practical test outside of the educational (e.g. Levy, 2011) and psychological (e.g. Reise et al., 2011; Wilson, 2013) fields are relatively rare. Within the organizational context, the application of IRT is recent, Tezza et al. (2011) applied the logistical model of two unidimensional parameters, Birnbaum (1968) measured the usability in commercial websites, Trierweiller et al. (2013) applied the same model to propose a scale for measuring the disclosure of information of environmental management practices in Brazilian industries, Bernini et al. (2014) used a multidimensional approach of IRT to investigate the heterogeneity in residents’ reaction to the tourism industry. Tay et al. (2011) applied a mixed model that considers latent variables of variables observed to measure union citizenship behavior with years of work experience and gender as covariates. Rivers et al. (2009) applied IRT to measure employees’ attitude in relation to the directorate of an organization and its overall communication with them. Other studies, such as Carter et al. (2011), LaHuis et al. (2011), Nye et al. (2010), Tay & Drasgow (2012) also applied IRT in the organizational context and Soares (2005) implemented it in socio-economic status index, demonstrating the viability of this tool as well as its potential in this field.

Depending on the final objectives and the structure of the data, MIRT can be considered a special case of multi-varied statistical analysis, especially factorial analysis or modeling of structural equations, or even, as an extension of the unidimensional IRT (Reckase, 2009).

Before presenting the models of MIRT, it is important to understand the multidimensional structures involved in a test: First, the relationship between latent dimensions and the items and then, the relationship between the latent dimensions and the respondents. As stated by Adams et al. (1997) and Hartig & Hohler (2009), the standard of relations between the dimensions and the items can be defined by a loading matrix with a simple structure (multidimensionality between the items) or by a structure of complex loading (multidimensionality between each item), and therefore it varies in complexity. Meanwhile, the standard of the relationship between the latent dimensions and the ability of the respondent has a compensatory or non-compensatory interaction.

Hartig & Hohler (2009) believe that the advantage of models that use the multidimensional structure between items is that they are less complex than the models that use multidimensionality within each item and the latent trace can be easily elucidated. Among these models, the estimate of the latent trace scores provides a simple performance measure of a specific set of items. In many cases, these measures will be highly correlated, because the items measure the same set of abilities. Nevertheless, the latent dimensions in the multidimensional model between items represent the necessary combination of all the abilities required to solve the respective items, regardless of how these abilities need to be integrated. Any overlapping is represented in the latent correlations. Hence, if the main interest of the study is to obtain descriptive measures of performance in areas of a determined content, the models between items are more suitable than the more complex models such as those of multidimensionality within each item (Hartig & Hohler, 2009).

The distinction between the multidimensionality within each item and between items is illustrated in Figure 1.

Figure 1 Distinction between the multidimensionality within each item and between items. 

Although the type of dimensionality is a basic concept, the distinction between multidimensionality between items and within each item is vital to the correct identification of MIRT models (Babcock, 2009).

From a practical point of view Traditional IRT models assume that the data can be explained by a single latent trait. However, this restriction makes these models inappropriate for data with multidimensional structure as shown in Figure 1 (Hambleton & Swaminathan, 1985).

Min (2003) summarizes three different conditions under which the application of the one dimensional model is suitable: (a) the examined capacity and characteristics vary in a dimension as a presumed model, (b) examining the ability varies only in one dimension even if the tested items are measuring more than one skill, (c) the ability of the examined capacity is different in multiple dimensions, but all items are measuring the same compound skills. In line with Lin (2008), there are other conditions that cannot be classified within these three mentioned and applying unidimensional models can be problematic. Studies have shown that when the multidimensional data are modeled based on the unidimensional assumption, measurement errors will increase and the consequences of the results are problematic (Ansley & Forsyth, 1985; Sireci et al., 1991; Ackerman, 1994; Reckase, 1995).

Before deciding which is the most appropriate multidimensional structure of IRT it is necessary to assess the dimensionality of the data set. There are some methods that can be used to make it possible like restricted information and full information methods.

Soares (2005), states that the restricted information method is the inspection of the eigenvalues of the tetrachoric correlation matrix. The factor analysis method of full information, instead of using the eigenvalues tetrachoric correlation matrix, creates a multidimensional model using the warheads curve (Bock & Aitkin, 1981; Bock et al., 1988), and is therefore an adaptation of the factor analysis traditional model, that considers the dimensional structure associated with continuous variables set for dichotomous data as in the following equation:

and e~N(0,ψ) with ψ diagonal, it is then that:

where σei is the standard deviation of ei .

The structure of the model can be seen as

Parameterizing (6) in the following:

where bi is the difficulty of the item, and aij is the specific discriminating of each dimension, serving as one of the bases for estimation methods of multidimensional models of IRT.

Reckase (1985) describes the compensatory multidimensional model of two parameters in the following manner:

(1)

where:

  • Uij = response of the individual (or website) j to item i (0 or 1);

  • aik = vector of the parameter of discrimination of the item i in the dimension k;

  • θjk = vector of the latent trait of the individual (or website) j in dimension m;

  • di = scale of the parameter of difficulty of item i.

The exponent of e in the model (1) can be expressed in the following manner:

(2)

Equation (2) shows that the exponent is a linear function of elements of θ with parameter d as the ordinate of the origin and the elements of vector a as the parameters of inclination/discrimination. One of the properties of this model is that the expression represented in the exponent defines a line in a space of k dimensions that can generate lines of equiprobability, meaning, this multidimensional form allows the existence of infinite linear combinations that result in the same exponent, thus generating, the same probability of accuracy. This property confirms the model of compensatory characteristic.

Gibbons & Hedeker (1992) developed general concepts based on the classic work of Holzinger & Swineford (1937) and propose the Full-information Bifactorial (FI Bifactorial) model for dichotomous data (Li & Rupp, 2011). This model consists of general factors and groups of factors or of independent dimensions. The model FI Bifactorial assumes the place of a general factor that involves all the items and two or more groups of factors (or dimensions) corresponding to specific subgroups (Gibbons et al., 2007).

Mathematically, the FI Bifactorial model considers cases in which for n items, there is a solution of s factors of which one of the factors is considered a general factor and s-1 is related to groups or factors. The bifactorial solution restricts each item i to having a different load of zero ai 1 over the primary dimension and a second load (aik , k = 2,...s) and not more than one of the s-1 groups of factors. For four hypothetical items, the standard bifactorial matrix can be represented in the following (Gibbons & Hedeker, 1992):

(3)

where the first column of the matrix represents the main factor and the second and third columns represent the group of specific factors.

As specified by Seo (2011), the dimensional structure in a bifactorial model is pre-determined through previous information. Therefore, the bifactorial model is a confirmatory model. In the perspective of a confirmatory approach, the model allows each item to have loads in a single general factor and only one specific factor for the group. This particularity reduces the number of parameters to be estimated and gives the model greater degrees of liberty. In addition, the bifactorial model can avoid the problem of estimating inter-factorial correlations, because the general factor contributes directly to all items, and the factors secondary to the remaining residual information after the calculation of the general factor are independent from each other. A particular quality of this model is the fact that the secondary factors are necessarily orthogonal between each other and in relation to the general factor (Gibbons & Hedeker, 1992).

For two binaries, the bifactorial model can be defined as a particularization of the compensatory multidimensional model presented in equation (1). In the case of the bifactorial model, the restriction of loading the parameters of discrimination is inserted as it can be seen in equation (4).

(4)

where θjg is the ability of individual j in the general factor, θjesk is the ability of the individual j in the specific factor k, aig represents the parameter of discrimination of item i in the general factor and aiesk represents the parameter of discrimination of the item i in the specific factor k and finally, di represents the scale of the parameter of difficulty of the item i referring to the general dimension and to the specific dimension k. In this model, and in the compensatory multidimensional model represented in equation (1), the responses are assumed to be statistically independent.

Gibbons et al. (2007) believe that the FI Bifactorial model is relevant whenever the items share a common characteristic. The presence of the subgroups of items typically introduces associations relevant to the test that cannot be verified with total attribution of the loading parameter to the general factor. In addition, according to the authors, this separation of factors improves the error of the estimates.

Gibbons & Hedeker (1992) and Gibbons et al. (2009) believe that the restrictions of the bifactorial model presented in matrix (3) lead to a greater simplification of the equations of probability because they require only the evaluation of the two integral dimensions, which (a) allow analyzing models with a larger number of factors (or dimensions), (b) allow a conditional dependency between subgroups of identified items, and (c) in many cases provide a more parsimonious solution than the full information item factor analysis due to its lack of restrictions. Gibbons et al. (2007) extended the bifactorial model to cases of Polytomous Items Response Theory.

Figure 2 contextualizes the FI Bifactorial model within some multidimensional structures.

Figure 2 Four possible structures of latent traits. Source: Reise et al. (2007). 

Model A is the standard unidimensional model in which the covariance between the responses to the items is explained by a common factor. In model B, the matrix of data contains more than a common dimension, although the dimensions are not correlated. This is a trivial case of multidimensionality and it is easy to solve, forming subscales and then separately adjusting to the unidimensional models of IRT for each subscale. This is essentially equivalent to assuming the dimensions as not correlated.

Model C also has more than one common factor among the items, although the factors are correlated. This representation is characterized as a non-hierarchical multidimensional model.

Finally, model D is a bifactorial model, meaning, there is a general factor, which explains the correlations between items, but in addition, there are also the called “group” factors (on the right side of the figure), which are trying to capture the covariance of the items that are independent of the covariance of the general factor. In other words, it is expressed in terms of quality measure of websites, it can suppose that the latent trait quality, considering its conceptual amplitude, represents a general factor explained by other factors (for example, usability, aesthetic, architecture of information, etc.), thus characterizing a suitability to model D (hierarchic multidimensional, etc.), therefore, the scope of this concept can lead to its dissolution into correlated subfactors - Model C (multidimensional non-hierarchical).

2.2 Quality on the web

Quality is not a new concept in management of information systems. Researchers and professionals demonstrate that they are aware of the need to improve information systems to react to external and internal pressure and face the critical challenges for their growth and survival (Aladwani, 1999; Aladwani & Palvia, 2002).

From the early 1980s until the late 1990s it was possible to find various studies that tried to conceptualize quality in information systems, demonstrating the concern among professionals and academics to understand and improve these systems. These studies concentrate on conceptualizing specific topics in this context, such as the quality of management of data (Kaplan et al., 1998; Wang et al., 1995), quality of information (King & Epstein, 1983; Haga & Zviram, 1994), quality of software (ISO9126, 1992; Schneidewind, 1992; Kitchenham & Pfleege, 1996), global quality of the system (Kettinger & Lee, 1994; Nelson, 1996), and others. These studies were much more focused on the system evaluation, its performance and its relation with specific users.

Since the mid-1990s, with the development and popularization of the Internet, practices and researchers have strived to define quality in the context of the Internet (for example, Barnes & Vidgen, 2000; Day, 1997; Lindroos, 1997; Xie et al., 1998; Loiacono et al., 2002). Lindroos (1997) uses the perspective of software quality to discuss differences between web-based information systems and conventional information systems. Olsina et al. (1999), proposed a quality model for university sites, called Website QEM based on user opinions. Barnes & Vidgen (2000), Loiacono et al. (2002), Parasuraman et al. (2005) and Ding et al. (2011) also developed similar models more focused on commercial sites. These and various other studies break the quality of the websites into several attributes. The creation of these models is based mainly on experiences over many years in development and maintenance of information and web systems. The validation of these models is made mainly by empiric studies, such as the analysis of data collected in tests with users, satisfaction surveys and interviews. Nevertheless, different types of information systems can have different quality requirements (Worwa & Stanik, 2010). For example, commercial and personal websites are web-based information systems. However, their quality requirements are different, mainly in terms of information security and information searching issues. Thus, any study about the quality issues on the web must have a clear delimitation of the limits of the analysis given the large scope of the theme.

This study fits into the classification of Cristobal et al. (2007) as a study of website quality and design. Within this scope, website quality is understood as the quality of an information system, in which, according to Loiacono et al. (2002), focuses on information storage, processing, presentation and transfer.

As a result, the concept of website quality adopted is that of a set of technical and non-technical characteristics, allowing the user to proceed to create their objectives on a website in an accessible, efficient and pleasant manner. Technical characteristics are understood as the usability/navigability and presentation of information and the accessibility and interactivity of the system (focus of this study). Non-technical characteristics are understood as design, aesthetics, visual and commercial appeal, reliability, hedonism and empathy.

3 METHOD

The methodological procedure used in this work, involved the characterization of the study, the preparation of the items, data collection and statistical methodology.

In terms of the characterization, the study is predominantly quantitative although it has a qualitative exploratory base with the objective of understanding the field of study about quality in commercial websites and serves as a base for the elaboration of the items in this study.

3.1 Instrument testing

The questionnaire (checklist) has 75 items linked to the quality of websites. The construction of the items was made by the association of concepts from the analysis of 191 articles collected in a systematic analysis of literature. For example, some of the recurring concepts were “information content” associated with “credibility”, “accuracy”, “completeness”, “utility” which is reinforced by Kim et al. (2005). These concepts associations support the following construction:

- Is there product basic information? (Information content + utility + credibility). The Appendix shows each item with its reference. The items had their content validated by three expert judges in the area. These items, although they are objective response items independent from the user perception, were based on previous studies that used tests with users and or satisfaction studies.

3.2 Data collection

The sample definition used the intentional sampling method, in order to draw a low, medium and high sample quality of the commercial sites used by Brazilian population. Accordingly, in addition to sites with the most diverse genres of market products, there was a variety of design styles, usability, aesthetics and layout contemplating from something relatively primary to overly demanding, which does not necessarily imply high or low quality, only guarantees diversity, a precondition of Item Response Theory. There is no consensus on the optimal sample size for use of item response theory (Downing, 2003; Wongtada & Rice, 2008). The data collection was conducted on 441 Brazilian commercial websites. 56 out of 75 items were collected manually and another 19 were collected automatically using the Achecker tool (http://achecker.ca/).

3.3 Analysis and discussion

The statistical analysis methodology primarily conducted the evaluation of the number of dimensions of the set of items followed by a verification of quality of each item and then a validation of the dimensions and the verification of the suitability of Item Response Theory multiple unidimensional models; compensatory multidimensional and bifactorial IRT. The dimensional analysis of the original data set (75 items) was made through factor analysis method of restricted information and the method of factor analysis of full information. In the first method, the number of dimensions observed was based on a tetrachoric correlation matrix and parallel analysis, which was used to Psych Package (Revelle, 2012) implemented on software R (R Core Team, 2012), because of its dichotomous responses, the dimensionality of the total set was also verified through the full information method. The approach used is described by Bock & Aitkin (1981) and Bock et al. (1988), in which the dichotomous items treatment and the estimation of the loading factor is achieved by the technique called factor analysis of full information, implemented on software R (R Core Team, 2012) in MIRT package (Chalmers, 2012) and flexMIRT software (Cai, 2012). In this method, the number of dimensions was evaluated based on two information criteria, the Bayesian Information Criterion (BIC) (Schwarz, 1978) and the Akaike Information(AIC) (Akaike, 1973). The use of this method for determining the number of dimensions isdiscussed by Bartolucci et al. (2012), Nylund et al. (2007) and Rost (1997) the suitability of the bi-factor model and compensatory model of MIRT was evaluated based on AIC and BIC information criteria. Table 1 shows the flowchart of the analysis and targets.

Table 1 Flowchart of the analysis and targets. 

Analysis Target
Tetrachoric correlation matrix Evaluation of the number of dimensions
Parallel analysis Evaluation of the number of dimensions
Factor analysis method of restricted information Evaluation of the number of dimensions and Analysis dimensional structure
Factor analysis method of full information Evaluation of the number of dimensions and Analysis dimensional structure
Unidimensional model of two parameters IRT Assessment of the adequacy of the unidimensional structure and multiple unidimensional structure
Multidimensional model of two parameters compensatory MIRT Assessment of the adequacy of the multidimensional structure of IRT
Bifactorial model IRT Assessment of the adequacy of the Bifactorial model IRT

4 RESULTS AND DISCUSSIONS

The analysis of dimensionality phase revealed the complexity of an analysis of this nature. Depending on the statistical technique used for this analysis, the results can diverge in terms of the number of dimensions. The analysis by the restricted information method suggested the existence of 9 dimensions (Table 2). Meanwhile, the complete information method suggested the existence of 3 dimensions (Table 3) while the parallel analysis technique indicated the existence of more than three dimensions (Fig. 1).

Table 2 Values specific to the tetrachoric correlation matrix. 

Dimension 1 2 3 4 5 6 7 8 9 10
Eigenvalue 11.0 5.7 4.4 3.9 3.4 3.1 2.7 2.5 2.3 2.2
Accumulated proportion of explained variation 14.6 22.2 28.0 33.3 37.8 41.9 45.5 48.8 51.9 54.8

Table 3 Comparison of models of one, two, three and four dimensions. 

χ2 Degrees of liberty p-value AIC BIC
Mod1 23049.49 23648.21
Mod2 542.489 74 <0.001 22655.00 23549.09
Mod3 335.394 73 <0.001 22465.61 23651.07
Mod4 219.207 72 <0.001 22390.40 23863.25
Mod5 138.058 71 <0.001 22394.34 24150.59
Mod6 -1848.45 70 <0.005 24382.79 26418.44
Mod7 47.51 69 0.9775 24473.28 26784.34
Mod8 25.239 68 0.9999 24584.04 27166.52

Table 2 indicates that the first eigenvalue is 11.0 which in a set of 75 items that represents 14.6% of the total variations explained by the first factor or first dimension. This result brings evidence that the construct should not be assumed to be unidimensional. In addition, if we use the criteria of proportion of explained variance with >50% we identify 9 dimensions. However, in the IRT context, the percentage of variance accounted exceeds Reckase’s (1979) rule of 20% for an item parameter to be considered stable. Taken together, it is reasonable to conclude that there are at least two dominants factors; this is sufficient to satisfy the IRT assumption (Bortolotti et al., 2013). Ventura et al. (2011) says that gold standard rules of thumb for deciding when a response matrix is “unidimensional enough” or multidimensional for IRT modeling do not exist (see Embretson & Reise, 2000), researchers generally seek a large ratio of the first to the next eigenvalues (e.g. >3 to 1). Thus, the ratio between the first and fifth eigenvalue have a value >3. The important criterion is whether if a dominant general factor running through the items exists or not. The way of exploring this issue, as discussed by Reise et al. (2007) and others (e.g., Immekus & Imbrie, 2008), is through an adequate bi-factor model and comparing the results with the unidimensional or multidimensional models (Ventura et al., 2011).

According to Chalmers (2012), the number of dimensions that generate a better adjustment to the data can be verified by comparing models using a generic variance analysis (ANOVA) implemented with the software R (R Core Team, 2013). The result returns the chi-square (χ 2) based on the test of verisimilitude, as well as the value AIC and BIC when comparing models. A comparison was made of nine models, the first assumed one dimension (Mod1); the second, two (Mod2); the third, three (Mod3); the fourth, four (Mod4); the fifth, five (Mod5); the sixth, six (Mod6); the seventh, seven (Mod7); the eighth, eight (Mod8); the ninth, nine (Mod9) dimensions. Table 3 presents the results.

Table 3 shows that, according to AIC, the best model is Mod4, since it has the lowest value of the statistic; BIC indicates Mod2. These two criteria were already compared in various studies (Dias, 2006; Nylund et al., 2007; McLachlan & Peel, 2000) and according to Bartolucci et al. (2012), AIC eventually tends to overestimate the number of dimensions. Meanwhile, in some cases, BIC can underestimate it.

The parallel analysis method indicates the existence of 24 dimensions. This conclusion can be verified in Figure 3 where the dotted line refers to the set of simulated data and the full line represents the real data. Note that there are 24 eigenvalues above the dotted line.

Figure 3 Result of the parallel analysis for the 75 items. 

Thus, the equation for these diverging results was based on the empiric analysis of dimensions and of the concepts of items associated to each one. Moreover, the number of dimensions definitions of the construct was based on the theoretical interpretation of each dimension in relation to the items associated to it, resulting in a four-dimensional structure. In this analysis, 31 items were identified that present commonality lower than 0.40 or factor loading less than 0.30; in all dimensions, were assumed to be uninformative items to construct, therefore, they were excluded from the analysis, leaving 44 items.

From a statistical point of view, these items are not correlated with other items, implying that if the goal is to measure quality of websites, these items theoretically are not associated with this goal. From a practical point of view, it can be seen that, in case of item 01, as discussed by Tezza et al. (2011), in which the same item was evaluated and discarded in a unidimensional construct, it may not feature a cumulative item. That is, the possibility of a pop-up opening for interaction with the website user is seen in the literature as bad for the quality and confuse the user (Storey et al., 2002; Petre et al., 2006; Nielsen, 2006). However, this feature is complex because it involves commercial and technological maturity issues and it may indicate that this feature purely evaluated as whether an opening pop-up window can or cannot evaluate linear or cumulatively with the quality of a website.

Theoretically, the four dimensions revealed themselves to be associated to concepts of navigability or user conduction-orientation, accessibility and reliability of the system, interactivity and presentation of information. These dimensions found in this work are related to the dimensions mentioned most often in the literature and that are directly related with the definition of quality on websites, which is a set of technical and non-technical characteristics of a web system, allowing the user to proceed to create their objectives on a website in an accessible, efficient and pleasant manner. Technical characteristics are understood to be usability-navigability, presentation of information, accessibility and interactivity of the system.

Based on the definition of a four-dimensional structure, an analysis of the multiple unidimensional, the compensatory multidimensional and the bifactorial models began based on Item Response Theory. The multiple unidimensional analysis, which subdivides the general set of items into unidimensional subsets based on the dimensions defined in the factorial analysis, proves to be more suitable than simply considering the unidimensional construct as a whole.

The unidimensional approach has some advantages and disadvantages. One advantage of a unidimensional approach or a multiple unidimensional approach is the ease of analysis and representation of the resulting scale. Nevertheless, one disadvantage of the supposition of unidimensionality in a multidimensional construct is the fact that the result will be a linear combination of the dimensions, which may not represent the reality. In addition, Ackerman (1991) shows that the estimation of the parameters of IRT model using the unidimensional mode when the data are multidimensional, tends to filter the dimensionality, that is, measuring a multidimensional ability on a unidimensional scale tends to generate larger values of unidimensional discrimination. On the use of a multiple unidimensional structure there is the inconvenience of generating different k scales that are theoretically not comparable in terms of parameters as media and standard deviation. This makes the joint analysis of all the items more difficult. The results of the parameter estimation in the unidimensional models ModU1 (the factorial analysis first dimension), ModU2 (the factorial analysis second dimension), ModU3 (the factorial analysis third dimensions) and ModU4 (the factorial analysis fourth dimension) can be seen in Tables 4, 5, 6 and 7, respectively.

Table 4 Parameters estimation of difficulty and discrimination assuming the unidimensional model of two parameters ModU1. 

Item Parameters
Discrimination Standard Error Difficulty Standard Error
57 1.77 0.26 1.16 0.13
58 6.68 3.79 0.80 0.20
60 1.93 1.01 -2.64 0.67
69 0.64 0.18 -1.52 0.39
71 1.00 0.29 -2.10 0.46
74 8.91 58.30 0.58 0.83
75 0.84 0.18 -0.59 0.16

Table 5 Parameters estimation of difficulty and discrimination assuming the unidimensional model of two ModU2 parameters. 

Item Parameters
Discrimination Standard Error Difficulty Standard Error
3 1.25 0.43 -3.04 0.74
6 1.85 0.68 -2.30 0.51
12 0.76 0.22 -2.65 0.68
22 0.89 0.32 -3.85 1.16
23 1.32 0.34 -2.58 0.50
32 0.94 0.20 -1.59 0.30
33 2.58 0.69 -2.08 0.27
35 0.97 0.22 0.38 0.14
37 1.65 0.34 -1.47 0.20
40 0.68 0.17 0.20 0.16
43 1.25 0.60 3.54 1.18
45 1.07 0.23 -1.28 0.24
48 2.10 0.46 -1.47 0.17
56 2.07 0.64 -2.74 0.53
73 0.00 0.00 0.00 0.00

Table 6 Parameters estimation of difficulty and discrimination assuming the unidimensional model of two parameters ModU3. 

Item Parameters
Discrimination Standard Error Difficulty Standard Error
21 3.00 1.18 -1.85 0.25
24 9.10 505.90 -2.23 2.95
25 1.62 0.43 -2.39 0.45
46 1.39 0.32 -0.79 0.15
47 1.15 0.47 -3.86 1.36
52 1.10 0.26 -1.49 0.30
55 1.00 0.65 -4.09 2.44
59 0.96 0.37 -3.04 0.97
61 0.57 0.17 -0.51 0.25
64 1.31 0.93 -3.40 1.83
65 1.50 0.72 -3.22 1.08
70 0.86 0.21 0.65 0.18

Table 7 Parameters estimation of difficulty and discrimination assuming the unidimensional model of two parameters ModU4. 

Item Parameters
Discrimination Standard Error Difficulty Standard Error
5 0.01 2.09 -317.42 ****
8 0.00 0.00 0.00 0.00
10 0.55 0.17 -3.53 1.01
19 0.86 0.19 -1.60 0.32
27 0.68 0.17 0.65 0.21
28 1.18 0.28 0.77 0.16
29 1.34 0.22 1.43 0.20
30 1.91 0.39 -0.97 0.15
32 1.10 0.23 -1.68 0.30
35 0.67 0.17 0.33 0.18
38 1.24 0.24 -0.73 0.16
66 0.74 0.41 -5.81 2.86

The items parameters estimation in model ModU4, unlike those of the other 3 models, become more instable than the estimation of the same items considering the general unidimensional model.

In a unidimensional analysis it is common to eliminate the items with problems of estimation and re-estimate the other items to verify if some change in the items of good quality exists, despite assuming the independence between items. This analysis was conducted in the four models and few variations were found in the estimates.

Meanwhile, the compensatory multidimensional model, in addition to proving to be statistically more suitable than the unidimensional model for this work, according to the analysis of Table 10, has greater possibilities for joint analysis, because it considers the construct as a whole particularized in dimensions. This joint analysis allows generating a series of measures related to the items, such as multidimensional discrimination, multidimensional difficulty and multidimensional location. In addition to measuring on the same scale the proficiencies (degree of quality) of the websites in each particular dimension.

Table 8 Parameters Estimation of discrimination for each dimension and parameters of multidimensional discrimination (MDISC), parameter of difficulty of scale (d) the parameter of multidimensional difficulty (MDIFF) and its respective standard error (se) for the 44 items, assuming the multidimensional model of two compensatory parameters. 

Item a1 Se a2 se a3 se a4 se MDISC d se MDIFF
03 0.30 1.03 1.52 3.50 0.07 39.49 -0.040 54.53 1.55 4.07 1.24 -2.62
05 -0.32 22.88 0.28 31.36 -0.40 34.43 -0.880 20.25 1.06 5.23 9.19 -4.95
06 1.09 1.70 1.91 15.07 0.30 54.56 -0.250 64.98 2.23 4.62 1.76 -2.07
08 -0.14 0.59 0.15 39.66 -0.41 35.75 -1.210 16.01 1.29 0.81 0.32 -0.63
10 -0.12 0.67 0.65 41.81 0.06 19.93 1.140 23.05 1.32 2.08 0.53 -1.58
12 -0.13 0.78 1.03 42.84 0.87 42.55 -0.680 23.18 1.52 2.49 0.58 -1.64
19 0.67 0.81 0.90 28.51 0.51 14.99 0.920 23.47 1.54 1.46 0.48 -0.95
21 1.43 1.46 0.89 38.57 1.89 15.48 0.970 30.04 2.71 5.12 2.41 -1.89
22 0.56 0.98 0.82 32.96 0.57 34.70 -0.590 19.83 1.29 3.75 1.02 -2.91
23 0.55 0.90 1.08 18.07 0.59 17.63 0.610 28.72 1.48 3.55 0.66 -2.40
24 11.44 177.69 5.62 257.53 16.62 283.21 7.090 291.11 22.11 46.63 719.38 -2.11
25 0.39 1.62 -0.11 99.18 2.82 24.08 -0.980 79.36 3.01 5.57 2.82 -1.85
27 0.06 0.56 0.62 40.23 -0.16 17.11 1.010 26.64 1.20 -0.60 0.34 0.50
28 0.15 0.72 0.04 52.30 0.42 41.15 1.560 10.25 1.62 -1.35 0.44 0.83
29 0.47 0.68 0.37 49.02 -0.36 23.97 1.150 22.52 1.34 -2.03 0.46 1.51
30 0.76 0.88 1.01 39.26 0.78 20.63 1.290 23.39 1.97 1.35 0.58 -0.69
32 0.62 0.65 0.65 18.39 0.65 10.41 0.610 14.00 1.27 1.58 0.42 -1.25
33 0.92 1.29 2.29 32.29 1.24 58.78 -0.010 61.27 2.76 5.55 2.87 -2.01
35 0.10 0.45 0.64 22.48 0.10 10.49 0.640 21.77 0.92 -0.35 0.30 0.38
37 0.62 0.77 1.45 11.39 -0.10 31.66 0.260 56.15 1.60 2.36 0.56 -1.47
38 0.95 0.70 0.47 33.23 0.70 21.06 1.100 10.75 1.68 0.67 0.48 -0.40
40 0.15 0.50 0.90 17.42 0.05 14.72 0.490 32.63 1.04 -0.15 0.28 0.14
43 0.01 0.86 1.06 9.27 -0.18 23.78 0.160 43.42 1.09 -4.20 1.18 3.86
45 0.28 0.66 1.13 25.77 0.05 17.52 0.710 40.90 1.36 1.49 0.41 -1.09
46 0.53 0.77 0.49 32.04 1.32 19.01 1.02 22.69 1.82 1.24 0.50 -0.68
47 0.60 3.15 1.07 48.33 1.16 41.31 -0.58 22.02 1.78 5.19 3.09 -2.91
48 0.90 1.04 1.62 20.43 0.79 28.94 0.67 44.95 2.12 3.06 0.80 -1.44
52 0.61 0.64 0.47 19.32 0.95 7.59 0.46 14.58 1.31 1.73 0.44 -1.32
55 0.45 0.92 0.23 41.62 1.16 16.83 -0.43 24.72 1.34 4.40 0.87 -3.29
56 1.91 4.34 1.81 18.87 0.93 41.42 0.33 49.98 2.81 6.88 16.99 -2.45
57 -2.08 0.72 -0.14 30.90 1.17 2.99 -0.03 35.93 2.39 -2.41 0.68 1.01
58 -50.47 398.55 8.56 309.51 6.92 181.90 10.11 217.53 52.64 -41.73 332.82 0.79
59 0.38 0.99 -1.19 27.54 1.34 46.79 0.69 76.93 1.96 3.80 1.45 -1.94
60 -2.74 4.63 -0.46 70.45 0.05 61.38 1.92 21.13 3.38 7.25 13.39 -2.15
61 -0.75 0.53 -0.40 21.09 0.95 25.62 0.62 38.81 1.42 0.38 0.28 -0.27
64 0.71 1.75 -1.92 38.71 1.75 56.92 0.33 115.42 2.71 6.39 3.54 -2.36
65 0.26 3.79 -2.53 56.83 2.89 84.56 0.98 160.92 3.97 8.67 20.84 -2.18
66 1.01 1.19 -0.14 27.72 0.74 29.10 0.94 24.34 1.57 4.87 2.89 -3.10
69 -0.86 0.56 0.32 12.34 -0.47 18.09 -0.40 23.44 1.11 1.11 0.37 -1.004
70 -0.35 0.44 -0.10 17.00 0.82 13.87 0.43 25.20 1.00 -0.57 0.27 0.573
71 -1.02 0.56 0.46 9.67 0.46 7.56 0.25 9.91 1.24 2.23 0.39 -1.81
73 0.82 11.08 -3.37 70.10 2.36 86.72 -0.22 185.21 4.20 10.71 37.53 -2.55
74 -4.25 1.69 0.60 25.59 1.23 10.37 0.62 19.79 4.51 -3.05 1.27 0.68
75 -0.79 0.37 0.29 15.64 0.33 13.21 -0.24 6.18 0.94 0.51 0.23 -0.54

Table 9 Estimate of the parameters of discrimination (a) for each dimension and in the general dimension and parameter of scale difficulty (d) assuming the bifactorial model. 

Item General a1 a2 a3 a4 d Item General a1 a2 a3 a4 d
3 0.531 0 0.374 0 0 2.116 43 0.277 0 2.097 0 0 -4.691
5 -0.483 0 0 0 -0.178 -0.178 45 0.627 0 0.274 0 0 0.793
6 0.755 0 0.676 0 0 2.342 46 0.999 0 0 0.584 0 0.69
8 -0.447 0 0 0 -0.839 0.538 47 0.672 0 0 0.402 0 2.591
10 0.522 0 0 0 0.747 1.293 48 1.089 0 0.722 0 0 1.79
12 0.106 0 0.831 0 0 1.418 52 0.709 0 0 0.253 0 0.956
19 0.948 0 0 0 0.021 0.832 55 0.388 0 0 0.741 0 2.519
21 1.804 0 0 1.104 0 3.364 56 1.31 0 0.321 0 0 3.426
22 0.233 0 0.73 0 0 2.2 57 -0.203 -1.064 0 0 0 -1.209
23 0.79 0 0.333 0 0 2.03 58 -0.513 -3.581 0 0 0 -2.966
24 6.115 0 0 4.418 0 12.924 59 0.243 0 0 0.9 0 1.925
25 0.385 0 0 0.997 0 2.252 60 -0.045 -1.175 0 0 0 2.986
27 0.482 0 0 0 0.151 -0.332 61 0.102 0 0 0.495 0 0.161
28 0.632 0 0 0 0.567 -0.768 64 0.12 0 0 1.796 0 3.903
29 0.505 0 0 0 0.218 -1.074 65 0.102 0 0 2.676 0 5.447
30 1.286 0 0 0 0.428 0.822 66 0.69 0 0 0 -0.155 2.572
32 0.805 0 0 0.068 0 0.916 69 -0.324 -0.387 0 0 0 0.616
33 1.146 0 1.127 0 0 3.182 70 0.228 0 0 0.411 0 -0.337
35 0.456 0 0.412 0 0 -0.233 71 0.051 -0.635 0 0 0 1.253
37 0.597 0 0.648 0 0 1.317 73 -0.103 0 -0.694 0 0 3.168
38 0.953 0 0 0 0.151 0.353 74 -0.253 -2.751 0 0 0 -1.829
40 0.519 0 0.076 0 0 -0.103 75 -0.091 -0.509 0 0 0 0.296

Table 10 Comparison between the unidimensional, multidimensional models of two compensatory parameters and of the bifactorial model based on the AIC and BIC, RMSEA and M2 information criteria. 

Model Log of maximum verisimilitude AIC BIC RMSEA M2
Unidimensional Model 13826 14002.7 14362.6 0.07 2844.39
MIRT Model 12616 13016.1 13833.9 0.04 1405.22
Bifactorial model 13200 13464.1 14003.8 0.05 1937.95

Table 8 shows the estimation of the parameters of the 44 items using the compensatory Item Response Theory multidimensional model by means of the flexMIRTTM software (Cai, 2012).

From this point of view, MIRT exceeds the unidimensional IRT due to the possibility of joint analysis of each item and of each respondent in each dimension, and as a consequence it is possible to identify the probability of possessing a certain characteristic in each website, based on the estimated quality of the parameters of the items. The multidimensional model of IRT offers great analysis and interpretations opportunities, nevertheless, these advantages result in an increase of complexity, particularly by working analytically in the vector space, and therefore with multiple geometric associations that are difficult to be visualized and interpreted in the traditional analytical forms.

The basic difference between the bifactorial model and the compensatory multidimensional model of IRT is that the first presents a general factor in which all the items are loaded and other specific orthogonal factors are analyzed.

The bifactorial model most widely diffused in the literature is the confirmatory nature model. Specific cases of exploratory analysis Jenrich & Bentler (2011) have been developed, although their practical application is still limited. Thus, the approach adopted in this work was the confirmatory one. To do so, the definition of the number of dimensions and of the groupings of the items defined in the factorial analysis were adopted in this work, assuming as a general dimension the quality of the commercial websites.

Table 9 shows estimates of the parameters of the bifactorial model, assuming the confirmatory structure based on the dimensions found in the factorial analysis.

In the bifactorial structure it is possible to verify to what degree the items are associated to the general factor, which in this case, is represented as the quality of a website. Analyzing the values of the loading parameter associated to the general factor, it is found that most of the items identified in the previous model as characteristics of the requirements of the system, such as accessibility, present a low load in the general factor, and in addition, some present a negative load, such as, items 57, 58, 60, 69, 74 and 75, this mathematically reflects the negative loading visualized both in the secondary factors and in the compensatory multidimensional model. Nevertheless, if we verify the secondary loadings in these items, there is a uniformity of the intensity and the direction of the parameters, and it is thus possible to indicate what inherent characteristics of the system have an orientation different from the characteristics associated to the organization of the information or direct navigation, and can then represent orthogonal or non-compensatory characteristics and that cannot be treated as part of a general factor. In particular, the fact that the bifactorial model assumes the orthogonality between the secondary dimensions and of those in relation to the general factor, limits the suitability of this model to the construct that clearly possesses a general factor ortohgonal to the other subdimensions, which is not the case of the construct in question. Whereas, it is found that the quality of commercial websites is not a characteristic that can be represented by a general dimension orthogonal to the other subdimensions, at least not for the construct developed in this study. Thus, the comparison of these three models suggests that quality of commercial websites is a non-unidimensional characteristic that can be divided into four compensatory dimensions.

The suitable version of the bifactorial model of IRT in relation to the compensatory model of MIRT, both assuming four dimensions, was evaluated based on BIC and AIC information criteria, Root Mean Square Error of Approximation (RMSEA) and statistical error M2. The results can be seen in Table 10.

It can be seen in Table 10, that in both the AIC and the BIC criteria, the compensatory multidimensional model in four dimensions (MIRT), had lower values than the bifactorial model and the unidimensional model, which indicates that this model is more suitable to the data than the bifactorial and unidimensional model. This verification is confirmed by RMSEA, which indicates a lower error for the MIRT model. The statistical error M2, available in the flexMirt software and discussed by Joe & Maydeu-Olivares (2010) and Liu & Maydeu-Olivares (2012), is similar to the chi-squared statistic and has been widely used in the verification of IRT models, given that, the lower its value, the more adjusted is the model in comparison to others. In this way, M2 indicates that the compensatory MIRT model is the most suitable of the three studied models.

As a practical interpretation of the multidimensional model of MIRT to four dimensions, Table 11 shows the estimate of the ability of the 4 first websites analyzed, in the normal scale N(0;1), that is, an average of zero and variance of one.

Table 11 Degree of multidimensional quality estimate of the 4 first websites of the sample based on the multidimensional model of two compensatory parameters. 

Website User orientation during navigation Accessibility and reliability of the system User control or user interaction with the system Presentation of information
001 0,105 0,130 -0,864 0,677
002 0,367 0,524 -0,753 0,040
003 -0,887 -0,328 -0,557 0,525
004 0,782 0,324 -0,879 0,314

We verified that the first website has a greater command of the items related to information presentation, thus requiring improving its quality, mainly in terms of user control or interactivity, which had below average control. On the website number 002, one can verify that there is a good command of system accessibility-reliability, but there was a need for improvement in terms of requirements of control partially by the user-interactivity and information presentation.

5 FINAL CONSIDERATIONS

In general, this work verified the viability of the use of Item Response Theory in the organizational context. The main contribution of this article is the consideration of the multidimensional structures, which is common in organizational evaluations. It was evident that in this application, that the unidimensional model (non-hierarchical structures) is not always the best choice. This mainly depends, on the nature of the items and on the characteristics of the respondents. A deeper analysis of each element is thus necessary.

This study also verified the unsuitability of the use of a general unidimensional model or of a multiple unidimensional model (both non-hierarchical structures), which utilizes unidimensional models to try to represent the general construct in the question - quality of e-commerce websites. The multidimensional model suitability of two compensatory parameters (non-hierarchical structures) was then found, as an additional analysis, based on the compensatory multidimensional model, the data adjustment to the bifactorial confirmatory model (hierarchical structures) was verified. This analysis showed that statistically, the multidimensional model, non-hierarchical structures aggregates more information to the construct when compared to the bifactorial model, hierarchical structures and to the unidimensional model. In this way, it is found that the bifactorial model does not represent more information to the construct requiring a possibly different approach than that considered in this work.

ACKNOWLEDGMENTS

We would like to thank the anonymous referees for very helpful suggestions, as well as the editorial team of the Pesquisa Operacional Journal.

REFERENCES

1 ACKERMAN TA. 1991. The Use of Unidimensional Parameter Estimates of Multidimensional Items in Adaptive Testing. Applied Psychological Measurement, 15: 13-24. [ Links ]

2 ACKERMAN TA. 1992. A didactic explanation of item bias, item impact, and item validity from a multidimensional perspective. Journal of Educational Measurement, 29: 67-91. [ Links ]

3 ACKERMAN TA. 1994. Using multidimensional item response theory to understand what items and tests are measuring. Applied Measurement in Education, 7: 255-278. [ Links ]

4 ADAMS RJ, WILSON M & WANG WC. 1997. The Multidimensional Random Coefficients Multinomial Logit Model. Applied Psychological Measurement, 21: 1-23. [ Links ]

5 AGNER L. 2012. Ergondesign e arquiterura de informação: trabalhando com o usuário. Rio de Janeiro: Quartet, 3a. edição. [ Links ]

6 AKAIKE H. 1973. Maximum likelihood identification of Gaussian autoregressive moving average models. Biometrika, 60: 255-265. [ Links ]

7 ALADWANI AM. 1999. Implications of some of the recent improvement philosophies for the management of the information systems organization. Industrial Management & Data Systems, 99: 33-39. [ Links ]

8 ALADWANI AM & PALVIA PC. 2002. Developing and validating an instrument for measuring user-perceived web quality. Information & Management, 39: 467-476. [ Links ]

9 ANSLEY TN & FORSYTH RA. 1985. An examination of the characteristics of unidimensional IRT parameter estimates derived from two-dimensional data. Applied Psychological Measurement, 9:37-48. [ Links ]

10 BARNES S & VIDGEN R. 2000. WebQual: an exploration of website quality. ECIS 2000 Proceedings, 74. [ Links ]

11 BARTOLUCCI F, MONTANARI G & PANDOLFI S. 2012. Dimensionality of the latent structure and item selection via latent class multidimensional irt models. Psychometrika, 77: 782-802. [ Links ]

12 BERNARDO M, MARIMON F & ALONSO-ALMEIDA MDM. 2012. Functional Quality And Hedonic Quality: A Study Of The Dimensions Of E-Service Quality In Online Travel Agencies. Information & Management, 497: 342-347. [ Links ]

13 BIRNBAUM A. 1968. Some Latent Trait Models and Their Use in Infering an Examiniee’s Ability. In: Lord FM & Novick MR. Statistical Theories of Mental Test Scores, MA: Addison-Wesley. goodness of fit test for the Rasch model. Psychometrika, 38: 123-140. [ Links ]

14 BERNINI C, MATTEUCCI M & MIGNANI S. 2015. Investigating heterogeneity in residents’ attitudes toward tourism with an IRT multidimensional approach. Quality & Quantity, 49: 805-826. [ Links ]

15 BOCK RD & AITKIN M. 1981. Maginal maximum likelihood estimation of item parameters: Application of an EM algorithm. Psychometrika, 46: 43-445. [ Links ]

16 BOCK RD & GIBBONS R , MURAKI E. 1988. Full information item factor analysis. Applied Psychological Measurement, 12: 261-280. [ Links ]

17 BORTOLOTTI SLV, TEZZA R, DE ANDRADE DF, BORNIA AC & DE SOUSA JÚNIOR AF. 2013. Relevance and advantages of using the item response theory. Quality & Quantity, 47: 2341-2360. [ Links ]

18 CAI L. 2012. FlexMIRTTM version 1.86: A numerical engine for multilevel item factor analysis and test scoring. [ Computer software]. Seattle, WA: Vector Psychometric Group. [ Links ]

19 CAI S & JUN M. 2003. Internet users’ perceptions of online service quality: a comparison of online buyers and information searchers. Managing Service Quality, 13: 504-519. [ Links ]

20 CAMILLI G. 1992. A conceptual analysis of differential item functioning in terms of a multidimensional item response model. Applied Psychological Measurement, 16: 129-147. [ Links ]

21 CARTER NT, DALAL DK, LAKE CJ, LIN BC & ZICKAR MJ. 2011. Using mixed-model item response theory to analyze organizational survey responses: An illustration using the job descriptive index. Organizational Research Methods, 14: 116-146. [ Links ]

22 CYBIS W. 2007. Ergonomia e Usabilidade: conhecimentos, métodos e aplicações/Walter Cybis,Adriana Holtz Betiol, Richard Faust. São Paulo: Novatec Editora. [ Links ]

23 CHALMERS RP. 2012. Mirt: A Multidimensional Item Response Theory Package for the R Environment. Journal of Statistical Software, 48: 1-29. [ Links ]

24 CRISTOBAL E, FLAVIAN C & GUINALIU M. 2007. Perceived e-service quality PeSQ: measurement validation and effects on consumer satisfaction and web site loyalty. Managing Service Quality, 17: 317-340. [ Links ]

25 COLESCA SE. 2007. An Assessment of the Quality of the Romanian Urban Web Sites. Informatica Economică, 42: 26-33. [ Links ]

26 DAY A. 1997. A model for monitoring Web site effectiveness. Internet Research: Electronic Networking Applications and Policy, 7: 1-9. [ Links ]

27 DIAS JG. 2006. Model selection for the binary latent class model: A Monte Carlo simulation. In: Data science and classification. Springer Berlin Heidelberg, p. 91-99. [ Links ]

28 DING DX & HU PJ , SHENG ORL. 2011. E-SELFQUAL: A scale for measuring online self-service quality. Journal of Business Research, 64: 508-515. [ Links ]

29 DOWNING SM. 2003. Item response theory: Applications of modern test theory in medical Education. Medical Education, 37: 739-745. [ Links ]

30 EMBRETSON SE. 1991. A multidimensional latent trait model for measuring learning and change. Psychometrika, 56: 495-515. [ Links ]

31 EMBRETSON S & REISE SP. 2000. Item Response Theory for Psychologists. New Jersey: Lawrence Erlbaum Associates, Inc. Publishers. [ Links ]

32 FASSNACHT M & KOESE I. 2006. Quality of electronic services conceptualizing and testing a hierarchical model. Journal of service research, 9: 19-37. [ Links ]

33 FERREIRA SBL & NUNES RR. 2008. E-Usabilidade, Rio de Janeiro: LTC. [ Links ]

34 GIBBONS RD, RUSH AJ & IMMEKUS JC. 2009. On the psychometric validity of the domains of the PDSQ: An illustration of the bi-factor item response theory model. Journal of psychiatric research, 43: 401-410. [ Links ]

35 GIBBONS RD & HEDEKER DR. 1992. Full-information item bi-factor analysis. Psychometrika, 57: 423-436. [ Links ]

36 GIBBONS RD, BOCK RD, HEDEKER D, WEISS DJ, SEGAWA E, BHAUMIK DK & STOVER A. 2007. Full-information item bifactor analysis of graded response data. Applied Psychological Measurement, 31: 4-19. [ Links ]

37 GLAS CAW. 1992. A Rasch model with a multivariate distribution of ability. Objective measurement: Theory into practice, 1: 236-258. [ Links ]

38 HAGA WJ & ZVIRAN M. 1994. Information systems effectiveness: research designs for causal inference. Information Systems Journal, 4: 141-166. [ Links ]

39 HAMBLETON RK. 1991. Fundamentals of item response theory. Vol. 2. Sage publications. [ Links ]

40 HAMBLETON RK & SWAMINATHAN H. 1985, Item response theory: Principles and applications. Norwell, MA: Kluwer Academic Publishers. [ Links ]

41 HARTIG J & HÖHLER J. 2009. Multidimensional IRT models for the assessment of competencies. Studies in Educational Evaluation, 35: 57-63. [ Links ]

42 HARTIG J & HÖHLER J. 2008. Representation of competencies in multidimensional IRT models with within-item and between-item multidimensionality. Zeitschrift für Psychologie/Journal ofPsychology, 216: 89-101. [ Links ]

43 HERNÁNDEZ B, JIMÉNEZ J & MARTÍN MJ. 2009. Key website factors in e-business strategy. International Journal of Information Management, 29: 362-371. [ Links ]

44 HOLZINGER KJ & SWINEFORD F. 1937. The bi-factor method. Psychometrika, 2: 41-54. [ Links ]

45 HUNG W-H & MCQUEEN RJ. 2004. Developing an Evaluation Instrument for eCommerce Web Sites from the First-Time Buyer’s Viewpoint. Electronic Journal of Information Systems Evaluation, 7: 31-42. [ Links ]

46 IBRAHIM EE, JOSEPH M & IBEH KI. 2006. Customers’ perception of electronic service delivery in the UK retail banking sector. International Journal of Bank Marketing, 24: 475-493. [ Links ]

47 INKPEN K, DEARMAN D & ARGUE R. 2006. Left-Handed Scrolling for Pen-Based Devices. International Journal of Human-Computer Interaction, 21: 91-108. [ Links ]

48 JENNRICH RI & BENTLER PM. Exploratory bi-factor analysis. Psychometrika, 76: 537-549. [ Links ]

49 ISO9126. 1992. Information Technology - Software Product Evaluation - Quality Characteristics and Guidelines for Their Use, International Organisation for Standardization, Geneva. [ Links ]

50 KAPLAN D, KRISHNAN R, PADMAN R & PETERS J. 1998. Assessing data quality in accounting information systems. Communications of the ACM, 41: 72-78. [ Links ]

51 KETTINGER WJ & LEE CC. 1994. Perceived service quality and user satisfaction with the information services function. Decision sciences, 25: 737-766. [ Links ]

52 KING WR & EPSTEIN BJ. 1983. Assessing information system value: An experimental study. Decision Sciences, 14: 34-45. [ Links ]

53 KIM DJ, SONG YI, BRAYNOV SB & RAO HR. 2005. A multidimensional trust formation model in B-to-C e-commerce: a conceptual framework and content analyses of academia/ practitioner perspectives. Decision Support Systems, 40: 143-165. [ Links ]

54 KITCHENHAM B & PFLEEGER SL. 1996. Software quality: The elusive target. IEEE software, 13: 12-21. [ Links ]

55 LAHUIS DM, CLARK P & O’BRIEN E. 2011. An examination of item response theory item fit indices for the graded response model. Organizational Research Methods, 14: 10-23. [ Links ]

56 LEVY R. 2011. Posterior predictive model checking for conjunctive multidimensionality in item response theory. Journal of Educational and Behavioral Statistics, 36: 672-694. [ Links ]

57 LI Y & RUPP AA. 2011. Performance of the S - Statistic for Full-Information Bifactor Models. Educational and Psychological Measurement, 71: 1-20. [ Links ]

58 LINDROOS K. 1997. Use quality and the world wide web. Information and Software Technology, 39: 827-836. [ Links ]

59 LIU CT, DU TC & TSAI H-H. 2009. A study of the service quality of general portals. Information & Management, 46: 52-56. [ Links ]

60 LOIACONO ET, WATSON RT & GOODHUE DL. 2002. WebQual: A measure of website quality. Marketing theory and applications, 13: 432-438. [ Links ]

61 LORD FM. 1980. Applications of item response theory to practical testing problems. Routledge. [ Links ]

62 LONG M & MCMELLON C. 2004. Exploring the determinants of retail service quality on the internet. Journal of Services Marketing, 18: 78-90. [ Links ]

63 MCLACHLAN GJ, BEAN RW & PEEL D. 2002. A mixture model-based approach to the clustering of microarray expression data. Bioinformatics, 18: 413-422. [ Links ]

64 MIN K-S. 2003 The Impact of Scale Dilation on the Quality of the Linking of Multidimensional Item Response Theory Calibrations. PhD Dissertation, Michigan State University, East Lansing, MI. [ Links ]

65 NELSON KG. 1996. Global information systems quality: key issues and challenges. Journal of Global Information Management, 4: 4-15. [ Links ]

66 NIELSEN J. 2000. Projetando websites. Gulf Professional Publishing. [ Links ]

67 NIELSEN J. 2007 Web 2.0 can be dangerous. Available from: <Available from: http://www.useit.com/alertbox/web-2.html >, acessado em 13/7/2012. [ Links ]

68 NIELSEN J & LORANGER H. 2006. Prioritizing web usability. Pearson Education. [ Links ]

69 NIELSEN J & TAHIR M. 2002. Homepage usability: 50 websites deconstructed. Indianapolis, In: New Riders. [ Links ]

70 NYE CD, NEWMAN DA & JOSEPH DL. 2010. Never Say “Always”? Extreme Item Wording Effects on Scalar Invariance and Item Response Curves. Organizational Research Methods, 13: 806-830. [ Links ]

71 NYLUND KL, ASPAROUHOV T & MUTHÉN BO. 2007. Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural equation modeling, 14: 535-569. [ Links ]

72 OLSINA L, GODOY D, LAFUENTE GJ & ROSSI G. 1999. Specifying Quality Characteristics and Attributes for Websites, in Proceedings of the First ICSE Workshop on Web Engineering, 16-17 May Los Angeles, USA. [ Links ]

73 OSHIMA TC & MILLER MD. 1992. Multidimensionality and item bias in item response theory. Applied Psychological Measurement, 16: 237-248. [ Links ]

74 OSTINI R & NERING ML. 2006. Polytomous item response theory models. No. 144. Sage. [ Links ]

75 PARASURAMAN A, ZEITHAML VA & MALHOTRA A. 2005. ES-QUAL a multiple-item scale for assessing electronic service quality. Journal of service research, 7: 213-233. [ Links ]

76 PETRE M, MINOCHA S & ROBERTS D. 2006. Usability Beyond the Website: an Empirically Grounded E-commerce Evaluation Instrument for the Total Customer Experience. Behaviour & Information Technology, 25: 189-203. [ Links ]

77 R CORE TEAM. 2013. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, http://www.R-project.org/. [ Links ]

78 RAUCH D & HARTIG J. 2010. Multiple-choice versus open-ended response formats of reading test items: A two-dimensional IRT analysis. Psychological Test and Assessment Modeling, 52: 354-379. [ Links ]

79 RECKASE MD. 1985. The difficulty of test items that measure more than one ability. Applied Psychological Measurement, 9: 401-412. [ Links ]

80 RECKASE MD & MCKINLEY RL. 1991. The discriminating power of items that measure more than one dimension. Applied Psychological Measurement, 15: 361-373. [ Links ]

81 RECKASE MD. 2009 Multidimensional Item Response Theory, Springer, New York - USA. [ Links ]

82 REISE SP, MORIZOT J & HAYS RD. 2007. The role of the bifactor model in resolving dimensionality issues in health outcomes measures. Quality of Life Research, 16: 19-31. [ Links ]

83 REISE SP, VENTURA J, KEEFE RS, BAADE LE, GOLD JM, GREEN MF & BILDER R. 2011. Bifactor and item response theory analyses of interviewer report scales of cognitive impairment in schizophrenia. Psychological assessment, 23: 245-261. [ Links ]

84 REVELLE W. 2012. Psych: Procedures for psychological, psychometric, and personality research. R package version 1.1-10 Retrieved from http://personality-project.org/r/psych.manual.pdf. [ Links ]

85 RIVERS DC, MEADE AW & FULLER WL. 2009. Examining question and context effects in organization survey data using item response theory. Organizational Research Methods, 12: 529-553. [ Links ]

86 ROST J. 1997. Logistic Mixture Models. In W.J. van der Linden and Hambleton (Ed.), Handbook of Modern Item Response Theory (pp. 449-463). New York: Springer. [ Links ]

87 ROTH S, SCHMUTZ P, PAUWELS S, BARGAS-AVILA J & OPWIS K. 2010. Mental models for web objects: where do users expect to find the most frequent objects in online shops, news portals, and company web pages? Interacting with Computers, 22: 140-15. [ Links ]

88 SCHNEIDEWIND NF. 1992. Methodology for validating software metrics. IEEE Transaction Software Engineering, 18: 410-422. [ Links ]

89 SEO DG. 2011. Application of the Bifactor Model to Computerized Adaptive Testing, Ph.D. thesis. University of Minnesota. [ Links ]

90 SIJTSMA K. 2011. Introduction to the measurement of psychological attributes. Measurement, 44: 1209-1219. [ Links ]

91 SIGNORE O. 2005. ‘A comprehensive model for Web sites quality’. Proc. of the Seventh IEEE International Symposium on Web Site Evolution (WSE’05), pp. 30-36. [ Links ]

92 SOARES TM. 2005. Utilização da teoria da resposta ao item na produção de indicadores sócio-econômicos. Pesquisa Operacional, 25: 83-112. [ Links ]

93 STOREY M-A, PHILLIPS B, MACZEWSKI M & WANG M. 2002. Evaluating the usability ofWeb-based learning tools. Educational Technology & Society, 5: 91-100. [ Links ]

94 SCHWARZ G. 1978. Estimating the dimension of a model. The annals of statistics, 6: 461-464. [ Links ]

95 TAY L, NEWMAN DA & VERMUNT JK. 2011. Using mixed-measurement item response theory with covariates MM-IRT-C to ascertain observed and unobserved measurement equivalence. Organizational Research Methods, 14: 147-176. [ Links ]

96 TAY L & DRASGOW F. 2012. Theoretical, Statistical, and Substantive Issues in the Assessment of Construct Dimensionality Accounting for the Item Response Process. Organizational Research Methods, 15: 363-384. [ Links ]

97 TEZZA R, BORNIA AC & ANDRADE DF. 2011. Measuring web usability using item response theory: Principles, features and opportunities. Interacting with Computers, 23: 167-175. [ Links ]

98 TRIERWEILLER AC, PEIXE BCS, TEZZA R, BORNIA AC & CAMPOS LM. 2013. Measuring environmental management disclosure in industries in Brazil with Item Response Theory. Journal of Cleaner Production, 47: 298-305. [ Links ]

99 VAN DER MERWE R & BEKKER J. 2003. A framework and methodology for evaluating e-commerce Web sites. Internet Research: Electronic Networking Applications and Policy, 13: 330-341. [ Links ]

100 W3C. 2008. Web content accessibility guidelines (wcag) 2.0. W3C Recommendation. Disponível em: Disponível em: http://www.w3.org/TR/WCAG20/ . Acessado em 25 de maio de 2015. [ Links ]

101 []WANG R & STOREY V. 1995. A. Firth, A framework for data quality research. IEEE Transactions on Knowledge and Data Engineering, 7: 623-640. [ Links ]

102 WILSON M. 2013. Using the concept of a measurement system to characterize measurement models used in psychometrics. Measurement, 46: 3766-3774. [ Links ]

103 WONGTADA N & RICE G. 2008. Multidimensional latent traits of perceived organizational innovation: Differences between Thai and Egyptian employees. Asia Pacific Journal of Management, 25: 537-562. [ Links ]

104 WORWA K & STANIK J. 2010. Quality of Web-based information systems. Journal of Internet Banking and Commerce, 15: 1-13. [ Links ]

105 YOO B & DONTHU N. 2001. Developing a scale to measure the perceived quality of Internet shopping sites (SITEQUAL). Quarterly Journal of Electronic Commerce, 2: 31-47. [ Links ]

106 XIE M, WANG H & GOH TN. 1998. Quality dimensions of Internet search engines. Journal of Information Science, 24: 365-372. [ Links ]

107 ZEITHAML VA, PARASURAMAN A & MALHOTRA A. 2002. Service quality delivery through web sites: a critical review of extant knowledge. Journal of the academy of marketing science, 30: 362-375. [ Links ]

108 ZERFASS A , HARTMANN B. 2005. The usability factor: Improving the quality of E-content. In E-Content (pp. 165-182). Springer Berlin Heidelberg. [ Links ]

Appendix

Item Description of the item Source
01 When opening the homepage do pop-up windows open? Storey et al. (2002); Petre et al. (2006); Tezza et al. (2011)
02 Does the homepage present the site’s main content areas (navigation)? Nielsen (2000)
03 Does the homepagemake clearwhat the site does (demonstrate the main products and or a brief description of its objective and or benefits that it offers), without needing the roll-bar? Nielsen (2000)
04 Does the homepage present a summary of the most important sales? Nielsen (2000)
05 Does the site have moving images that can distract the user? Colesca (2007)
06 Do the links for sales go directly to the sale announced? Nielsen (2000)
07 Are the menus in alphabetical order? Olsina et al. (2001)
08 Does the site have a cascademenu? Nielsen & Loranger (2006)
09 For navigation, is there a track on the left or links on top of the page? Nielsen & Tahir (2002)
10 Are sub-categories grouped? Nielsen & Tahir (2002)
11 Does the title of the window (browser) list the name of the site in the first place? Nielsen & Tahir (2002)
12 Are there information for telephone contact or an address? Nielsen (2000); Nielsen & Tahir (2002);
13 Do the page have a consistent visual appearance, that is do they always have the same visual appearance? Aladwani & Palvia (2002); Ferreira & Nunes (2008)
14 Does the site have the option to access it in other languages? Colesca (2007)
15 Do the clickable words (in color or underlined) have a distinct form when selected? Nielsen & Loranger (2006)
16 Do the field labels begin with a capital letter, and are the remaining letters in lower case? Aldwani & Palvia (2002)
17 Are the titles aligned to the left? Inkpen et al. (2006)
18 Are text paragraphs separated? Zerfass & Hartmann (2005)
19 Are apparently clickable words in fact clickable? Nielsen (2000)
20 Are the title for screens, windows and dialog boxes on top, centered or left justified? Inkpen et al. (2006)
21 Do all the pages have a search field? Nielsen (2000)
22 When there is rolling, are there design elements (in the initial screen) thaat appear with end of page markers? Nielsen & Loranger (2006)
23 Does the company logo in the upper left corner on all the site pages? Roth et al. (2010)
24 Is there a link with a single click that leads to the homepage? van der Merwe & Bekker (2003)
25 Does the site allow navigating its pages in only one window, that is, there is not opening of new windows in amid the navigation? Nielsen & Loranger (2006); Tezza et al. (2011)
26 Are there different colors for previously visited links? Nielsen & Tahir (2002)
27 Is there a list of frequently asked questions - FAQs? Colesca (2007); Hernández et al. (2007); Tezza et al. (2011)
28 When entering search terms in the search field does the search engine offer suggestions? Long & McMellon (2004)
29 Is the search system flexible in relation to the terms used by the user, that is, if the user types a term incorrectly, does the search system suggest a correction? Agner (2012)
30 Do the search results allow classification by other criteria in addition to cost? Nielsen (2000); Tezza et al. (2011)
31 Do long lists present indicators of communication, number of items and pages? Nielsen & Loranger (2006); Tezza et al. (2011)
32 Are page continuation items visible? Agner (2012)
33 Is the price of a product next to the image or link for the product? Hung & McQueen (2004)
34 Is it possible to expand the photos of products to visualize details? Novikova (2009); Nielsen & Loranger (2006); Tezza et al. (2011)
35 In products in which there is more than one perspective, is it possible to visualize all the perspectives? Nielsen & Tahir (2002); Tezza et al. (2011)
36 Are the groups of command buttons available in the column to the right, or on a line below the objects to which they are associated? Inkpen et al. (2006)
37 Is there sufficient information about the products (size, basic characteristics)? Signore (2005)
38 Is there a way for consumers to provide to insert feedback about the products? Agner (2012)
39 Are there background images for the text? Nielsen & Tahir (2002)
40 Does the site present related products at the end of the page? Nielsen (2000)
41 Is there an option to share the page on social networks? Agner (2012)
42 Is there an on-line help option? Hernández et al. (2007)
43 Does the site have multimedia for product presentation? Aladwani & Palvia (2002)
44 Does the company offer a free service, such as free shipping? Nielsen & Tahir (2002)
45 Is there an indication that the site is safe at the time of making the purchase? Nielsen & Tahir (2002)
46 When filling in the forms, can the user visualize the next steps in the interface? Nielsen (2000)
47 Does the site have other payment forms in addition to a credit card? Yoo & Donthu (2001); Hung & McQueen (2004)
48 Is it possible to know the total cost before registering (including shipping costs)? Nielsen & Loranger (2006)
49 Are the fields used for the forms to be filled in by the user grouped linearly, avoiding unnecessary spaces? Inkpen et al. (2006); Tezza et al. (2011)
50 When filling in a form, are options that are not valid or not available visibly deactivated to avoid errors? Nielsen (2000)
51 When filling in a form, is information provided about how to complete the form? Cybis (2007); Nielsen & Loranger (2006); Tezza et al. (2011)
52 Are required data differentiated from optional data in a visually clear manner? Cybis (2007); Nielsen & Loranger (2006)
53 Is it possible to make a purchase without registering (which includes user name and password)? Nielsen & Loranger (2006); Tezza et al. (2011)
54 Does the system provide audio signals when there is a problem with the data entered? Nielsen & Loranger (2006); Tezza et al. (2011)
55 Are the error messages free of abbreviations and or codes generated by the operating system? Cybis (2007)
56 Can any user action be taken back with the UNDO or BACK option? Nielsen & Loranger (2006); Tezza et al. (2011)
57 Does all the non-textual content that is presented to the user have an alternative in text form that serves an equivalent purpose? W3C (2008)
58 Can the information, structure and relations broadcast through the presentation be determined in a programmatic form or are they available in the text? W3C (2008)
59 Is there another visual form of presenting information, beside color, to indicate an action, request a response or distinguish a visual element? W3C (2008)
60 Does the visual presentation of text and images have a relation of contrast of at least 4.5:1? W3C (2008)
61 Except for captions and text images, can the text be resized up to 200 percent without support technology, and without losing content or functionality? W3C (2008)
62 Does the visual presentation of text and images have a contrast ratio of at least 7:1? W3C (2008)
63 Are all the content functions operable through a keyboard interface without a need for any space of time between each individual item entered, except when the subjacent function requires entering data that depends on the sequence of actions by the user and not only on final points? W3C (2008)
64 For each time limit defined by the content, is there control by the user? W3C (2008)
65 For information in movement, in an intermittent mode, in shifting or automatic updating, is there an option for user control? W3C (2008)
66 Is a mechanism available to ignore blockages of content that are repeated on various Web pages? W3C (2008)
67 Do the Web pages have titles that describe the topic or finality? W3C (2008)
68 Can the finality of each link be determined based only on the text of the link or based on the text of the link together with the respective context of the determined link in a programmatic form, except when the finality of the link is ambiguous for users in general? W3C (2008)
69 Do the headers and the tags describe the topic or the finality? W3C (2008)
70 Can the pre-defined human language for each Web page be determined in a programmatic manner? W3C (2008)
71 Does changing the definition of a component of the user interface automatically provoke a change of context, at least when the user has been warned about this situation before using the component? W3C (2008)
72 Are the components that have the same functionality in a set of Web pages identified in a consistent manner? W3C (2008)
73 If an input error is automatically detected, is the item that has the error identified and is the error described to the user in text? W3C (2008)
74 Are labels or instructions provided when the content requires inputting data by the part of the user? W3C (2008)
75 In the content implemented using languages for marking, do the elements have complete marks at the beginning and end, are the elements fit according to the respective specifications, do the elements have duplicated attributes, and are all the IDs exclusive? W3C (2008)

Received: April 20, 2015; Accepted: October 20, 2016

*Corresponding author.

Creative Commons License This is an open-access article distributed under the terms of the Creative Commons Attribution License