3.2. Questionnaire and AHP Application
displays the set of criteria that we selected based on the literature review, the interviews to the managers and the academic experts’ advice. We used these criteria to elaborate the questionnaire four our AHP application.
We have a double purpose: first, to evaluate the perception of a group of Costa Rican firms and public institutions with respect to the selected criteria and, second, to measure the propensity of those organizations to adopt a specific environmental certification according to this set of criteria. Among all the available certifications, this study focusses on CN for its relevance for the sake of pursuing sustainability in Costa Rica. As an alternative, we take ISO 14001, which as an important and well-established certification oriented to the EMS of the company.
We conducted a four-level-AHP exercise as shown in Figure 1
. The first level (“Goal
”) is the organization’s objective to choose an environmental certification. The second level refers to the general-purpose criteria (or simply “Criteria
”) that we consider relevant for the decision. According to our classification, these are the economic-strategic aspects, on the one hand, and environmental sustainability, on the one hand. The third level (“Sub-criteria
”) disaggregates the general-purpose criteria into more specific aspects. We refer to the latter as “sub-criteria” to differentiate them form the aggregate “criteria” on the second level. The lower level (“Alternatives
”) refers to the environment certifications that the respondents will evaluate in terms of the criteria and sub-criteria.
To conduct the exercise, we identified a group of companies holding the CN certification, the ISO 14001 certification, or both. For the sake of completeness, we also included some companies that did not hold any of them. Apart from companies, we also addressed some public institutions to check if the latter had somewhat different perceptions and preferences than the former. We identified the CN companies and institutions from the Climate Change Department [29
] and the ISO 14001 organizations from the Institute of Technical Standards of Costa Rica [66
]. In the case of non-certified companies, we searched the emails’ contacts on their webpages.
In 24 May 2019, we invited 171 companies and 12 certified public institutions to complete the questionnaire on the Google’s survey platform. In the group of companies, 58 of them did not have the CN or ISO 14001 certification, 62 were CN but not ISO 14001 certified, 33 were 14001 but not CN certified and 18 had both certifications. With respect to public institutions, six of them were CN certified, 2 were ISO 14001 certified and four did not have any of both certifications.
In October 2019, 22 companies ‘managers had completed the questionnaire correctly; four of them were CN certified, four ISO 14001 certified, seven had both certifications and seven had neither. In addition, two respondents were from public institutions (a university and a governmental department), both of which were CN but not ISO 14001 certified. Appendix A Table A2
lists the organizations’ features and the positions of the respondents.
Following standard AHP methodology, the questionnaire sets pairwise comparisons of elements (criteria, sub-criteria, or alternatives) belonging to the same level with respect to their contribution to the immediate superior level. It is based on Satty’s [11
] scale, which allows us to convert the qualitative judgments into numerical values (see Table 2
At the beginning of the questionnaire, we explained the structure of the questions and the Saaty scale. Then, we asked the respondents for their pairwise judgments within each level of the study, i.e., about the importance level of criteria (with respect to the goal of selecting a VEC), sub-criteria (with respect to each of the general criteria) and the relative merit of the alternatives, CN and ISO 14001 (with respect to each of the sub-criteria). Figure 2
shows three examples of the questions presented to the managers for levels 2, 3, and 4. The rest of questions had the same structure.
The individual responses give rise to the individual Satty’s comparison matrices. Then, we use the geometric mean to combine the individual matrices and obtain the consensus pairwise comparison matrices [67
]. These combined matrices are the ones that we use to compute the consensus priority weights for each level of the study, first for the whole group of respondents, and then to different subgroups (see Section 4.1
and Section 4.2
below). Before computing the priority weights, we computed the consistency ratios (CR
) of each of the relevant consensus comparison matrices. The consistency ratio is defined as
, where CI is the consistency index of each matrix and RI
is the consistency index of a random matrix of the same size. See Saaty [11
] for details about the calculation of the consistency indexes (CI
) and the average consistency values (RI
) of randomly generated matrices. The consistency ratios of all the consensus comparison matrices that we use in the study (for the whole group and the subgroups) are well below 0.10, which is the threshold value recommended by Saaty [11
We calculated the priority weights for the criteria, sub-criteria and alternatives using a variant of the traditional eigenvector method. For each level, we multiply the associated comparison matrix iteratively by itself. In each iteration, we add up the elements of each row of the matrix and normalize the resulting vector yielding an approximation to the first eigenvector of the initial matrix [70
]. The process stops when the approximate eigenvector obtained in one iteration does not change significantly (up to four decimal places) from the previous iteration. The result is taken as the vector of relative importance or priority weights; see e.g., [71
We denote the criteria eigenvector (level 2) as VC. It indicates the weights or relative importance of economic-strategic and environmental sustainability criteria. In level 3, we have two eigenvectors: One for the economic-strategic sub-criteria, denoted as VEC, and one for the environmental sustainability sub-criteria, denoted as VSC. By combining levels 2 and 3 we can obtain the eigenvectors representing the global contributions of each sub-criterion to the goal of the study. Thus, the global eigenvector associated with the economic-strategic sub-criteria (VGEC) is obtained as follows: , where is the weight given to the economic-strategic criterion in level 2, i.e., the first element of VC. Similarly, the global eigenvector of all the sustainability sub-criteria (VGSC) is obtained as follows: , where is the weight given to the environmental sustainability criterion (the second element of VC).
At level 4, we have six two-components eigenvectors, VAi (i = 1, …, 6), each of one indicates the weight or relative score of the alternatives in terms of a given sub-criterion. Specifically, we get four eigenvectors related to the economic-strategic sub-criteria that can be grouped as and 2 eigenvectors related to the environmental sustainability sub-criteria, . We can order the alternatives according only to the economic-strategic criteria by computing or only to the environmental sustainability criteria by computing . Finally, the globally preferred alternative can be determined by computing the global weight vector as follows: .