*3.4. Demographic Variables as Significant Antecedents of the Wellbeing—Performance Classification.*

As indicated previously, we used multidimensional logistic regression to explain cluster membership based on the demographic covariates. The odds ratios for all the models are displayed in Tables 3–6. An odds ratio greater than 1 implies that a person in a given category has greater odds of belonging to a cluster than a person in the reference category (in the case of variables with 2 categories) or than the average of all the subjects in the study (in the case of variables with more than 2 categories). An odds ratio below 1 suggests reduced odds. We identified different demographic predictors when different operationalizations of wellbeing (hedonic-eudaimonic) and performance (self- or supervisor-evaluated) are considered.


**Table 3.** Multinomial logistic regression analysis of factors associated with the clusters. Model 1: Hedonic (H) Performance employee

 *p p* \*\*\* *p* ≤ 0.001.

#### *IJERPH* **2019**, *16*, 479



 cluster is in brackets; Cluster 1: h E- h PE; Cluster 2: ml E-mh PE; Cluster 3: mh E-ml PE; Cluster 4: l E-l PE; OR: odds ratio; CI: confidence interval; *p* ≤ 0.05, *p* ≤ 0.01,\*\*\**p*≤0.001.


Reference cluster is in brackets; Cluster 1: h E- h PE; Cluster 2: h E-l PE; Cluster 3: l E-l PE; OR: odds ratio; CI: confidence interval; \* *p* ≤ 0.05, \*\* *p* ≤ 0.01, \*\*\* *p* ≤ 0.001.

#### 3.4.1. Multidimensional Logistic Regression: Model 1 (H-PE)

The multinomial logistic regression analyses identified five predictors that explain cluster membership: the organization's sector, gender, seniority in the position, educational level, and occupational category (see Table 3). The results show that the model has a good fit (−2 log LR = 679.06, X<sup>2</sup> = 129.83, df = 24, *<sup>p</sup>* ≤ 0.001) (with LR being the likelihood ratio). The probability of having high wellbeing and high performance is greater in the production sector and for managers. The probability of having medium low wellbeing and medium high performance is greater in the production sector, for people with more than 5 years of seniority, and for technicians/administrative work. The probability of having medium high wellbeing and medium low performance is greater in the services sector, for people with less than 5 years of seniority, with professional training or high school, and for technicians/administrative workers. Finally, the probability of having low wellbeing and low performance is greater in the services sector, for men, with no education or compulsory education, and for technicians/administrative work.

Comparing Clusters 1 (high levels) and 4 (low levels), the production sector, women, people with a university degree, and managers are more likely to be in Cluster 1, whereas the services sector, men, people with no education or compulsory education, and technicians/administrative workers are more likely to be in Cluster 4.

#### 3.4.2. Multidimensional Logistic Regression: Model 2 (E-PE)

The multinomial logistic regression analyses identified five predictors that explain cluster membership: the organization's sector, gender, age, educational level, and occupational category (see Table 4). The results show that the model has a good fit (−2 log LR = 777.45, X<sup>2</sup> = 99.68, df = 27, *p* ≤ 0.001) The probability of having high wellbeing and high performance is greater in the production sector, women, and managers. The probability of having medium low wellbeing and medium high performance is greater for men, people over 50 years old, and unqualified manual workers or technicians/administrative workers. The probability of having medium high wellbeing and medium low performance is greater for women, and for unqualified manual workers or technicians/administrative workers. Finally, the probability of having low wellbeing and low performance is greater for the services sector, men, people with no education or compulsory education, and technicians/administrative workers.

Comparing Clusters 1 (high levels) and 4 (low levels), results are similar to those in Operationalization 1. The production sector, women, people with university degrees, and managers are more likely to be in Cluster 1, whereas the services sector, men, people with no education or compulsory education, and technicians/administrative workers are more likely to be in Cluster 4.
