SciELO - Scientific Electronic Library Online

 
vol.20 issue2Association between symptoms of anorexia and bulimia nervosa and suicidal behavior in school children of Boyacá, ColombiaThe effect of correspondence between descriptions of contingencies and contingencies on the conduct of election under the self-control paradigm author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand

Journal

Article

Indicators

Related links

  • On index processCited by Google
  • Have no similar articlesSimilars in SciELO
  • On index processSimilars in Google

Share


Acta Colombiana de Psicología

Print version ISSN 0123-9155

Act.Colom.Psicol. vol.20 no.2 Bogotá May/Aug. 2017

https://doi.org/10.14718/ACP.2017.20.2.10 

Artículos

Comparison of subjective health between personality prototypes extracted from general population of Mexico

Comparação da saúde subjetiva entre protótipos de personalidade em população geral do México

Pedro Solís-Cámara1  * 

Rosa Martha Meda Lara2 

Bernardo Moreno Jiménez3 

Andrés Palomera Chávez2 

Pedro Juárez Rodríguez2 

1 Centro Universitario de la Ciénega, Universidad de Guadalajara

2 Centro Universitario de Ciencias de la Salud, Universidad de Guadalajara

3 Facultad de Psicología, Universidad Autónoma de Madrid

* , Universidad Autónoma de Madrid, Av. Universidad #1115. CP 47820, Ocotlán, Jalisco, México. Teléfono: (+52-33)3686-2267, psolis@mexis.com


Abstract

The objective of this study was to extract personality prototypes from general population of Mexico and to compare subjective health indicators between these prototypes. Participants were 994 individuals (aged 14 to 63 years). Five personality traits (Neuroticism, Extraversion, Openness, Conscientiousness, and Agreeableness) were assessed with the NEO-FFI. Subjective indicators of health (self-rated health and psychological distress) were assessed with a question regarding health status and the GHQ-12. To verify the consistency of the prototypes, the sample was divided into two age groups, young (14 to 25 years) and mature (26 to 63 years.). Three stable prototypes were recovered from both groups. Resilient individuals (low neuroticism and high in other traits) had the best subjective health; the Non-Resilient individuals (high neuroticism and low in other traits) had the worst subjective health; and Self-Disciplined individuals (high conscientiousness and medium scores in other traits) were in the middle of these extremes in subjective health. Self-discipline and resilience were most discriminative in terms of subjective health. Findings are discussed in terms of the generalization of prototypes across cultures.

Key words: Self disciplined; non-resilient; personality; prototypes; resilient; health

Resumo

O objetivo deste estudo foi obter protótipos da personalidade em população geral do México e comparar a saúde subjetiva entre eles. No total, participaram 994 indivíduos de 14 a 63 anos. Avaliaram-se cinco traços de personalidade (extroversão, neuroticismo, abertura à experiência, conscienciosidade e amabilidade) com o NEO-FFI, e a saúde subjetiva com o GHQ-12, e uma pergunta sobre o estado de saúde. Para corroborar a consistência dos protótipos, dividiu-se a amostra em dois grupos de idade: jovem (de 14 a 25 anos) e adulto (de 26 a 63 anos). Como resultado, obtiveram-se três protótipos em ambos os grupos: as pessoas resilientes -baixo neuroticismo e alto nos demais traços-, os que tiveram a melhor saúde subjetiva; as não resilientes -alto neuroticismo e baixo no restante dos traços-, que apresentaram a pior saúde subjetiva, e as pessoas disciplinadas -alto em conscienciosidade e média nos outros traços-, que apresentaram uma saúde subjetiva intermediária em comparação com os outros protótipos. Finalmente, a autodisciplina e a resiliência foram os traços que melhor discriminaram a boa saúde. Os achados são discutidos em termos da generalização de protótipos através de culturas.

Palavras-chave: disciplinado; não resiliente; personalidade; protótipos; resiliente; saúde

Resumen

El objetivo de este estudio fue obtener prototipos de personalidad en población general de México y comparar la salud subjetiva entre ellos. En total, participaron 994 individuos de 14 a 63 años de edad. Se evaluaron cinco rasgos de personalidad (Neuroticismo, Extraversión, Apertura, Responsabilidad y Amabilidad) con el NEO-FFI, y la salud subjetiva con el GHQ-12 y una pregunta sobre el estado de salud. Para corroborar la consistencia de los prototipos se dividió la muestra en dos grupos de edad: joven (de 14 a 25 años) y maduro (de 26 a 63 años). Y como resultado se recuperaron tres prototipos en ambos grupos: las personas Resilientes -bajo neuroticismo y alto en el resto de los rasgos-, quienes tuvieron la mejor salud subjetiva; las No-Resilientes -alto neuroticismo y bajo en el resto de los rasgos-, que presentaron la peor salud subjetiva; y las personas Disciplinadas -alto en responsabilidad y promedio en los otros rasgos-, que presentaron una salud subjetiva intermedia en comparación con los otros prototipos. Finalmente, la autodisciplina y la resiliencia fueron los rasgos que discriminaron mejor la buena salud. Los hallazgos se discuten en términos de la generalización de prototipos a través de culturas.

Palabras clave: Disciplinado; no-resiliente; personalidad; prototipos; resiliente; salud

INTRODUCTION

One of the main proposals for the study of personality which has achieved considerable support is the taxonomy of personality represented by the Big Five Traits model, also known as Five-Factor Model (FFM; Garcia, Aluja & Garcia, 2004; Hoyle, 2010). The NEO Personality Inventory (NEO-PI-R) (Smith, Williams & Segerstrom, 2015) and a short version of the same known as the NEO-FFI, are proba bly the most well-known instruments to assess personality traits within the Five-Factor Model. These two instruments, characterized as self-reports assess the following dimensions: Neuroticism (N: anxiety, hostility, depression, social anxiety, impulsivity and vulnerability to stress); Extraversion (E: warmth, gregariousness, assertiveness, activity, excitement seeking and positive emotions); Openness to Experience (O: fantasy, aesthetics, feelings, actions, ideas and values); Agreeableness (A: trust, straightforwardness, altruism, compliance, modesty and tendermindedness); and, finally, Consciousness (C: order, dutifulness, achievement striving and self-discipline) (Costa & McCrae, 2008).

Personality studies within the Five-Factor Model can be divided into two broad approaches (De Fruyt, Mervielde & Van Leeuwen, 2002). The first one, known as variable-centred approach, conceptualizes personality in terms of individual differences among people; in this proposal, each one of the dimensions of the NEO-FFI describes differences across individuals with any number of variables as the pri mary research goal (e.g., Roberts, Kuncel, Shiner, Caspi & Goldberg 2007). The second approach is the typological, or person-centred, which is founded on the configuration or set of attributes that define each individual, and based on this notion, it can group individuals with a similar personality pattern and explore the relationships with the variables of interest (Xie, Chen, Lei, Xing & Zhang, 2016).

According to Donellan and Robins (2010), the Resilient, Overcontrolled, and Undercontrolled personality types were identified for the first time by Robins, John, Caspi, Moffitt, and Stouthamer-Loeber (1996). Studies have found that the three types are replicable and have been reported to have external validity to predict behavior problems, as well as being useful in research and for applied purposes (e.g., Asendorpf, Borkenau, Ostendopf & Van Aken, 2001; De Fruyt et al., 2002; Van Leeuwen, De Fruyt & Mervielde, 2004).

In terms of the outstanding traits of the types, Resilient people show low scores in Neuroticism and relatively high scores in the rest of the dimensions; these people have been described as emotionally stable, assertive and kind toward others. Overcontrolled individuals have a high score in Neuroticism and low in Extraversion, with relatively low or average scores in the other dimensions; these people have been described as emotionally brittle (anxious or tense), and introverted. The characteristics of the Undercontrolled individuals are low scores in Agreeableness and Consciousness, and mean scores in the other dimensions; they are described as people with lack of self-control, are excitement seekers and oriented to actions (antisocial behavior). Multiple findings confirm that these personality types are associated with psychological and social variables in all developmental stages of individuals (Akse, Hale, Engels, Raaijmakers & Meeus, 2007; Hart, Burock, London, Atkins & Bonilla-Santiago, 2005; Klimstra, Hale, Raaijmakers, Branje & Meeus, 2010; Steca, Alessandri & Caprara, 2010; Van Leeuwen et al., 2004).

In spite of the above, the person-centred approach to personality description is just emerging, particularly in Latin-American countries and, moreover, it is no stranger to controversy. This paper does not intend to elaborate on all outstanding issues since there are excellent reviews on the subject (Caspi, 1998; Donellan & Robins, 2010; Herzberg & Roth, 2006; Miller, 2012), but it seeks to explore the most remarkable contradictions and the possible reasons for them. Thus, it is necessary to consider the following:

  1. Some authors have pointed out the lack of consistency within the personality types. For example, Herzberg and Roth (2006), in their analysis of seven studies, show that unlike other dimensions, Neuroticism is consistent in Resilent and Overcontrolled people, but not so the other traits. For example, they found that traits such as Extraversion and Openness, of the Undercontrolled type, varied greatly in z-scores (from < 0.05 to > 0.50). This lack of consistency within a personality type is a problem, and this variability can be worse if researchers are using the same label for what may be a different type, thus pretending homogeneity of types which is far from being perfect (Herzberg & Roth, 2006). Related to the above, Donellan and Robins (2010) report studies that found a "Nondesirable" type, instead of the Undercontrolled which resembles the reverse type of Resilient -high score in Neuroticims and low in the rest of the traits-. But it seems that it would be more appropriate to name it Non-Resilent, as suggested by Zawadzki and Strelau (2003). Configurations with four and five types have also been reported (Isler, Liu, Sibley & Fletcher, 2016); for example, with five types which include the Confident -rela tively high scores in Openness and Extraversion- (Herzberg & Roth, 2006) and in other cases, the Reserved or Discrete -with low scores in Openness- (Roth & von Collani, 2007) whereas others include the Reserved and Ordinary- mean scores in all traits - (Kinnunen et al., 2012).

  2. A problem that can be related to the above is the use of cluster analysis. This has been the preferred method, particularly with Ward analysis, followed by K-Means procedure (Akse, Hale, Engels, Raaijmakers & Meeus, 2004; Chapman & Goldberg, 2011; Steca et al., 2010). However, Donellan and Robins (2010) argue that when using that method,multiple analyses of the data are not carried out, or subsamples are not obtained. For example, in the studies reviewed by Herzberg and Roth (2006), the samples ranged from 156 to 786 individuals, but in a number of studies these were divided in half with the purpose of demonstrating the criterion of replicability of the types. In doing so, the samples were much lower than 500, and it is argued (Steinley, 2003) that they are therefore inadequate to obtain stable classifications. In addition, several samples did not correspond to the general population, nor were homogeneous in one variable (e.g., university students). This can lead to greater or lesser prevalence of persona lity disorders affecting the recovery of personality types (Donellan & Robins, 2010).

  3. It is also of interest to consider that not all dimensions that supposedly characterize each type behave the same through studies. For the Resilient type, most of the studies agree in a low score in Neuroticism and high in the rest of the dimensions. But other works indicate mean scores in all dimensions and low in Neuroticism (Van Leeuwen et al., 2004); or relatively high in Extraversion and Consciousness but low in Neuroticism (Ekehammar & Akrami, 2003). For the Overcontrolled type they coincide in high score in Neuroticism and low in Extraversion; but other studies indicate mean or low scores in the rest of the traits (Donellan & Robins, 2010), or high in Neuroticism and Consciousness, and low in Extraversion and Openness (Herzberg, 2009). For the Undercontrolled type there is a great variability and only coincides in low score in Consciousness and in some cases in Agreeableness, with some other exceptions (Ekehammar & Akrami, 2003) or in greater detail, & Robins (2010) and Herzberg & Roth (2006). As a result, these findings generate uncertainty about the characteristics of the personality types. In addition, an aspect missing in the studies is about the quantitative criterion to identify high or low scores. This is important because, as happens in a study which does specify standard deviations below or above the arithmetic mean of the traits to classify the types (De Fruyt et al., 2002), it prevents inaccurate interpretations about what it means "slightly low or relatively high scores".

  4. Researchers have argued that contradictory results are likely due to other sample characteristics, such as sex, age, and culture (Herzberg & Roth, 2006). Regarding sex, no consistent evidence was found that this variable should be controlled. In terms of age, there are few studies, like the one by De Fruyt et al. (2002) where the three types of personality were not recovered in samples of different ages. Other investigators argue that since the traits are related to age, it is likely that this variable has an effect on the prototypes (Strus, Cieciuch & Rowinski, 2014). Regarding culture, in Philippines' university students, two of the three most common types were recovered in both women and men, but investigators extracted other types also. Avdeyeva and Church (2005) conclude that support for cross-cultural generalization is partial. No studies were found in the Latin American population, although in Spain the three types have been replicated with a student sample, but not with one of the general population (Boehm, Asendorpf & Avia, 2002). Recently, Alessandri et al. (2014) recovered the three types from university students (n's from 235 to 499), although with some variations; anyhow, the authors considered that the transcultural generalization was relatively good among the countries studied (Spain, USA, Italy and Poland). Either way, it is premature to draw conclusions about the effect of culture since most of the studies have been performed in industrialized countries, particularly in Europe. In contrast, in the first study in Chinese adolescents, Xie et al. (2016) only recovered the Resilient and Undercontrolledtypes (the last has the lowest consistency in European studies), and two others, named Ordinary and Withdrawn.

  5. Finally, an important distinction on the conceptuali zation of personality types as labels or as categories that are found in nature, is discussed. The generalized perspective looks at the types as appropriate labels to summarize the information of personality and it is recognized that does not reflect biological categories, but rather relative distinctions between people. For these and other reasons (see Donellan & Robins, 2010) it seems more appropriate to use the term prototype rather than type, as has been suggested by several authors (Weir & Gjerde, 2002). Prototype indicates preemi nence, synonym of the first mold or model, and should not be confused with a consolidated or final status. However, in research these labels are interchangeably referred to as types, profiles or prototypes of personality.

Based on the above, the present study addresses two questions regarding whether it is possible to derive the most common personality prototypes from the general popula tion of Mexico and whether it is possible to corroborate differences in subjective health among the recovered pro totypes. Despite the fact that several studies support the cross-cultural generalization of the FFM, at present this is questioned (see: Church, 2016). And, given the above considerations, to replicate the prototype studies is of great importance in Latin-American countries, because, as far as it is known, no evidence has been obtained in favor of or against such generalization. In this context, it seems premature to establish hypothesis; however, it would be expected to recover at least the three most common pro totypes (Resilient, Overcontrolled, and Undercontrolled) and perhaps others.

On the other hand, it is important to compare subjective health indicators by prototypes, considering that studies in developed countries have found that people with high Neuroticism perceive themselves as having a poorer health status and more physical illnesses than individuals with lower Neuroticism. In general, Neuroticism has been as sociated with a negative impact on health in both subjective and objective measures; while Consciousness has been associated with positive outcomes (Lahey, 2009; Roberts, Kuncel, Shiner, Caspi & Goldberg, 2007).

For example, Kinnunen et al. (2012) studied the per sonality prototypes and health of304 adults. They included objective measures (e.g., body mass, blood pressure, etc.) and subjective health indicators. The latter were assessed in three ways: the health status was obtained by means of a question regarding how participants would describe their health in the last year; through a health interview to assess physical symptoms, and with the GHQ-12 as a measure of psychological distress. Prototypes were formed based on the NEO-FFI scores using a latent profile analysis (clus ters of traits) and were referred as Resilient (low N-high in the other traits), Overcontrolled (high in N-low in E), Undercontrolled (high in O and E-low in C and N) Reserved (low in N, E, O, and A-high in C) and Ordinary (profile zero in the latent model for all of the traits). Regarding significant differences, the Resilient people had the best subjective health. Overcontrolled individuals had the poor est subjective health. The other three profiles were in the middle of these two extremes.

In summary, the objective of the present study was to extract personality prototypes from the general population of Mexico and to compare subjective health indicators among them. It was planned to include, in addition to the GHQ-12 as a measure of health, a question similar to the one made by Kinnunen et al. (2012). It should be added that analysis by sex was considered due to the evidences that indicate that men have a better self-perceived health than women, who more frequently present physical and psychological symptoms (Kaleta, Polanska, Dziankowska-Zaborszcsyk, Hanke & Drygas, 2009). Analysis by sex for the prototypes were not performed because previous findings showed no consistent sex-differences (Borkenau, Hrebickova, Kuppens, Realo & Allik, 2013). It is expected that the results provided in this study offer knowledge on the subject in Latin-America for the first time.

METHOD

This is a cross-sectional descriptive study.

Participants

Bearing in mind that the state-of-the-art studies on prototypes were taken into account, it was planned to ob tain a larger sample and split it into two age groups, thus generating a young group and a mature group. A sample of 1440 participants was calculated by using a non-probability cluster sampling, using sex and chronological age as crite ria. The study had a 69% response rate, so the sample was composed of 994 individuals living in 13 counties in four different states of Mexico (i.e., Jalisco, Veracruz, Colima and Michoacán).

Ages ranged from 14 to 63 years. Men (n = 413) had a mean age of 31.6 yrs (SD = 13.1) and women (n = 581) 28.3 yrs (SD = 11.8). The percentages of the major demo graphic variables were:

  • Marital status: 64.6 single, 29.0 married, 3.7 with a partner, 1.7 divorced, and 1.0 widowed;

  • Educational level: 1.1 without studies, 5.1 elementary school, 9.2 junior high school, 25.3 high school, 14.4 technical school, 38.6 bachelor's degree, and 6.3 graduate or postgraduate degree;

  • Current occupation: 51.7 student, 9.4 blue-collar worker, 10.2 house-keeper, 9.6 trader, 3.4 technician, and 15.8 professional.

None of the participants reported a history of mental or psychiatric disorders and all of them signed the respective letter of informed consent. For minors (7.5%), their parents signed an informed consent.

Instruments

Socio-demographic questionnaire Data on age, marital status, educational level, and the current occupation were collected. In addition, the participant was requested to answer the question: "Do you have any health problems?"

General Health Questionnaire (GHQ-12). The GHQ-12 evaluates the self-perceived health as well as the general health status (Goldberg and Williams,1988). It is recognized for its usefulness as a screening measure of the subjective general health status and some mental dysfunctions. It is presented in a Likert-type scale with four answer choices: never (0), sometimes (1), almost always (2) and always (3). In this study, the 12 items were divided based on the wor ding in positive or negative format, with six items in each subscale; the higher the score, the greater the dysfunction. For descriptive purposes internal consistencies (Cronbach's alpha) are reported for the full scale (.80), for positive items (.85) and for negative ones (. 82). Other authors have reported alpha values of .81, .85, and .82, respectively (Solís-Cámara et al., 2016). The version used was the same with which the factorial structure of the GHQ-12 was demonstrated in general population of Mexico (Solís-Cámara et al., 2016).

It should be mentioned that the GHQ is probably the most widely recognized instrument in the world to assess the degree of perceived health and the 12-item version has been identified as the most widely used because of its ease of administration (Sánchez-López & Dresch, 2008). Is a screening tool for mental health that is significantly associated with multiple physical disorders (González et al., 2012, Tuuliainen / Kirsi Sipilä, Mäki, Könönen & Suominen, 2015), and has been the recommended version for Health surveys (McDowell, 2006). However, it seems necessary to take into account that Hankins (2008) showed that factorial analyzes that support the multidimensional nature of GHQ-12 do not take into account that negatively formatted items generate spurious factor divisions (two or three) due to negative format responses. Studies in many countries, including Mexico, have confirmed this error of the method in GHQ-12, and have suggested, among other options, to separate positive and negative items from the instrument to have a general screening method for subjec tive health or psychological distress (Smith, Oluboyede, West, Hewison & House, 2013; Solís-Cámara, Meda Lara, Moreno-Jiménez & Juárez, 2016; Urzúa, Caqueo-Urizar, Bargsted & Irarrázaval, 2015).

NEO-Five Factors Inventory (NEO-FFI (Costa & McCrae, 2008). The Spanish version of the instrument (Costa & McCrae, 2008) was used because on the date the study was conducted, there was no version of it validated in Mexico, and it was only until recently when a 30 items version was proposed (Meda Lara, Moreno-Jimenez, Garcia, Palomera Chávez & Mariscal de Santiago, 2015). The NEO-FFI eva luates the dimensions ofthe FFM: Neuroticism, Extraversion, Openness to experience, Agreeableness and Consciousness (N, E, O, A, C, respectively). It is suitable for adolescents and adults, and consists of 60 items, divided into five scales with 12 items each.

The NEO-FFI is presented in a five-point Likert type format: strongly agree, agree, neutral, disagree and strongly disagree; high scores indicate greater expression of the trait. Cronbach's alpha coefficients obtained in this study were: N (.78), and E (.81), O (.73), A (.71), C (.81). If .75 is considered as acceptable reliability, three dimensions meet the criterion and two are slightly below it (Hernández, Fernández & Baptista, 2010). Other studies report .86, .77, .73, .68 and 0.81 (Costa & McCrae, 2008); or .83, .78, .67, .67 and .84, respectively (Martinez & Cassaretto, 2011).

It should be mentioned that the NEO-FFI, was chosen to assess the personality, and the conglomerate analysis method, to recover the prototypes. In addition, the quan titative criteria by De Fruyt et al. (2002) were taken into account to identify the prototype scores. In general, the Resilient was identified by a low score in N (half standard deviation below average) and high scores (+ V DE) in the other dimensions; The Overcontrolled by a high score in N (+ V DE), low in E (-V DE) and average in the rest of the dimensions; And the Undercontrolled by a low score in C (between -V and 1 DE) and in A (-V DE) and average in the other dimensions.

Procedure

The instruments and the letter of informed consent were converted to electronic format with the SurveyMonkey software; these were sent to the email address of potential participants who were mainly students and university tea chers, housewives, blue-collar workers and traders residing in the states of Mexico mentioned above. For this study, the 2004 Declaration of Helsinki and the Ethics Code of Psychologists (Mexican Psychological Society, 2007) were taken into account, ensuring the integrity of the participants and the confidential use of the data. The project was endor sed by the Ethics Board of the University of Guadalajara.

Statistical Analysis

Data analyses were performed with the SPSS-21 sta tistical package. Internal consistencies (Cronbach's alpha) of the NEO-FFI scales and, separately, for the positive and negative items of the GHQ-12, were obtained. Skewness and kurtosis of the study's main variables were analyzed to know their normal distribution fit.

In order to achieve replicability of the cluster solution, the sample was split into two age groups (young and mature). A full two-step clustering analysis was perfor med with each group. The prototypes were derived by applying a two step clustering procedure, which assumed variables as typified (N, E, O, A and C), measurement of the Euclidean distances, with automatic number of clusters (up to 15), and with management of noise (25%). In order to carry out the hierarchical analysis method of Ward, the silhouette measure of cohesion and separation was used, with quality cluster criterion between >0.5 and the closest to 1.0 (i.e., good quality). After this, the non-hierarchical K-means clustering procedure, without updated means, was performed.The prototypes retained after the two-step clustering analysis were analyzed via Chi2 to establish size differences and sex distribution. Factorial ANOVA and univariate analysis were included to determine effects by age, gender and their interaction. Post-hoc analyses included a conservative method (Scheffe test) and a suitable one to correct for multiple comparisons (Hochberg's GT2).

Finally, in order to analyze prototype differences in sub jective health, raw scores were converted into z scores, like other studies have done (eg., Kinnunen et al., 2012). GHQ scores were analyzed by gender via the Student t test for independent samples, including the Levene's test for equality of variances, the standard error and confidence intervals.

RESULTS

This section presents a descriptive analyses of the sample divided by age group, followed by an analyses to determine the number of clusters to retain. Three prototypes are discussed in detail in terms of the five dimensions of personality. Finally, comparisons in subjective health by prototypes and sex in the two age groups are presented.

The sample was split into two groups, according to their mean age: young (14 to 25 yrs; M = 19.9, SD = 2.43) and mature (26 to 63 yrs; M = 41.2, SD = 9.42). The young group (YG; n = 541) was formed by 340 women (62.8%) and 201 men (37.2%) and the mature group (MG; n = 453) was formed by 241 women (53.2%) and 212 men (46.8%). Table 1 shows, for descriptive purposes, means, skewness, and kurtosis of the main variables for the total sample and by age group. Skewness and kurtosis of the majority of the variables indicate they fit the normal distribution and suggest that age concentrated slightly in a young age and most answers to the question on health status ("do you have any health problems?") focused on the response "no" (71%), particularly in the YG that shows a mean of 1.80. On the other hand, slight positive skewness is observed for GHQ-negative items, which indicates that the values grouped around low scores.

Table 1 Means, deviations, skewness and kurtosis of the variables for the total sample and by group 

Forming personality prototypes

The procedure was similar for both groups. In the young group (YG) the two step clustering procedure indicated two solutions with cohesion and comparison indicating good quality of the clusters (from >.05 and close to 1.0). The Ward method supported two-, and three-clusters solutions; the solution with two prototypes included one with a high score in N and low scores in the other traits, and the other was the opposite. The three-cluster solution was retained for further analysis rather than the two-cluster solution because it was thought that the latter was less enriching.

In the mature group (GM) the two step clustering proce dure indicated four solutions with cohesion and comparison indicating good quality of the clusters (from >.05 and close to 1.0). The method of Ward supported three-, four-, and five-cluster solutions. Four- and five-cluster solutions were similar in three prototypes, but the other two had high or low scores in all traits. Finally, the three-cluster solution was retained for further analysis rather than the four and five-cluster solutions as it offered a more consistent solution with at least two of the three common prototypes and was similar in both age groups.

Description of the prototypes

Table 2 presents the YG scores in the NEO-FFI by proto types. The Resilient prototype was characterized by a score with a standard deviation below the mean in Neuroticism and high scores in Extraversion (+1 SD), Openness (+1/2 SD), Agreeableness (+1/2 SD), and Conscientiousness (+3/4 SD). The second prototype scored half-standard deviation above the mean in Neuroticism and had low scores in Extraversion (-1/2 SD), Openness (-1/2 SD), Agreeableness (-1/2 SD), and Conscientiousness (-3/4 SD), and therefore was named Non-Resilient prototype. The third prototype had average scores (not even half SD below or above the mean) in all dimensions, with the exception of the high score in Conscientiousness (almost +3/4 SD), and was interpreted as Self Disciplined prototype. This because Conscientiousness comprises traits such as order, organi zation and self-discipline.

Table 2 Scores of personality traits by prototypes in the young group 

The size of the prototypes was different X 2(2) = 56.0, p < .000), with 28.3% of Resilient, 48.2% of Non-Resilient and 23.5% of Self Disciplined. Women were over-represen ted: in the Resilient prototype there were 99 women and 54 men; in the Non--Resilient prototype there were 148 women and 113 men and in the Self Disciplined, 93 women and 34 men. There were no age differences between prototypes (F(2, 538) = 0.88,p > .05), nor meaningful interaction of age by prototype and sex (F(2, 541) = 0.81,p > .05).

Table 3 presents the mean scores of the MG by perso nality prototypes. The Resilient prototype had the lowest score in Neuroticism (-1 SD), and high scores in the rest of the dimensions, all of them approximately with a standard deviation above the mean. The Non-Resilient prototype had the highest score in Neuroticism (+1/2), and low scores on Extraversion (-3/4 SD), Openness (-1/2 SD), Agreeableness (-1/2 SD), and Conscientiousness (-3/4 SD). The Self

Disciplined prototype was characterized by a pattern of scores around the mean or slightly above it (< V SD) in all dimensions, with the exception of the high score in Conscientiousness (+1/2 SD).

Table 3 Scores of personality traits by prototypes in the mature group 

The size of the prototypes for the MG was clearly di fferent, with 18.1% of Resilient, 44.8% of Non-Resilient and 37.1% of Self Disciplined. There were fewer women (33) than men (49) in the Resilient prototype (X2(1) = 3.12, p = .07). In the Non-Resilient prototype, women (n = 118) were significantly over represented (X2(1) = 5.36, p < .05), in comparison with men (n = 85). In the Self Disciplined prototype, sex (90 women and 78 men) was distributed more evenly (X2(1) = 0.86, p = .35). Age differences were observed between prototypes (F(2, 453) = 3.75,p < .05) and by sex (F(1, 453) = 10.0, p < .01), but no significant interaction was found (F(2, 453) = 2.63, p > .05). Post-hoc analysis (Scheffe and Hochberg tests) were statistically significant (p < .05); Self Disciplined individuals were younger (M= 39.6, SD = 8.67) than Non-Resilient individuals (M = 42.4, SD = 9.6). In addition, women (M = 39.5) were younger than men (M = 42.6).

Differences in health by prototypes

In order to enhance comparability of subjective health indicators, raw scores on health status as well as on positive and negative items of the GHQ-12 were standardized (z-scores). These z-scores were compared by prototypes and sex for both age groups. For ANOVA statistically signifi cant results, a conservative post-hoc test (Scheffe test) and Levene's test for equality of variances were performed on the data. Table 4 shows the results of these comparisons. It is important to clarify that the comparisons are mentioned as low or high scores with the intention of facilitating the communication of the results; however, it is known that z-scores refer to number of units below (negative) or above the mean (positive). The first statistically significant result is presented with examples to further clarify the subsequent.

Table 4 Differences in subjective health by prototypes and gender in young and mature groups 

For the YG, no significant differences were found to the question concerning the health status. Regarding differences found in GHQ-positive and negative items, these were similar but clearer for the positive items than for the negative ones, and only these analyses are presented. Resilients had the lowest score in GHQ-positive (see Table 4); that is to say, the greatest number of units or deviations (-0.72) below the mean compared with the Self Disciplined and Non-Resilient individuals; these results indicated better general health or less perceived psychological distress. In addition, Self Disciplined individuals also had greater number of units (-0.19) below the mean than the Non-Resilients (0.29).

The sex difference (Table 4) was due to the fact that men (n = 201) had a lower score in the positive items (M = -.436, Standard Error = .072; IC 95% = -.576/-.295) than women (n = 340; M = -.112, Standard error = .050; IC 95% = -.210/-014). That is to say, women perceived more health problems. Independent samples t-tests were conducted to identify the interactions prototype-sex in the positive items. The scores for men (M = -.716, SD = .87) were lower than those for women (M = -.000, SD = .86) only in the Self Disciplined prototype (t(125) = -4.11, p < .001; mean difference = -.716; IC 95% for the difference = -1.06/-.371); thus, women in this prototype perceived more health problems.

Due to the sex differences previously found, intra-sex comparisons were done in order to explore more in depth these effects. Differences in women were found among the Resilients (M = -.672, SD = .82) and Non-Resilients (M= .335, SD = .97; (t(245) = -8.420, p < .001) and the Self Disciplined women (M = -.000, SD = .86; t(190) = -5.488, p < .001); there was also a difference between these last two prototypes (t(239) = 2.702, p < .01); and, in all of these comparisons, the Non-Resilients showed the worst subjective health, but the Resilients had a significantly lower score than the other prototypes and, therefore, better subjective health. In the case of men, Resilients (M = -.815, SD = .79) and Self Disciplined individuals (M = -.716, SD = .87) had similar scores (t(86) = 0.543, p = .58). And both Resilients (t(165) = 7.04, p < .001) and Self Disciplined men (t(145) = 5.23, p < .001) had lower scores than Non-Resilients (M = .224, DE = .93). For this reason, Non-Resilient men also showed a poorer subjective health.

In the comparison of the prototypes regarding health indicators, the ANOVA for the MG resulted in no significant effects by sex or interactions (prototype x sex). Significant differences by prototypes were found to the question con cerning the health status. Since the question was answered with a yes or a no, the positive responses indicate a larger negative z-score (see Table 4). The Levene test for equality of variance was significant (F(5, 447) = 8,148,p < .001). Resilient individuals (M= -.122, Standard Error = .121; IC 95% = -.360/.116) had a score similar to Self Disciplined individuals (M = -.113, Standard Error = .083; IC 95% = -.276/.050), and only Resilients scored significantly lower (p < .05) than Non-Resilients (M = -.390, Standard Error = .076; IC 95% = -.540/-.240), both with the Scheffe and the Hochberg test. Although this difference appears to be marginal, the percentages of the "yes" responses to the health question, by prototypes, were clearly different: Resilients (15.5), Non-Resilients (46.8) and Self Disciplined individuals (32.0).

Table 4 also presents the differences by prototypes for the MG. Resilients had the lowest score in GHQ-positive, the greatest number of units or deviations (-0.56) below the mean compared with the Self Disciplined individuals and the Non-Resilient ones. This indicates the Resilients had the best self-rated health, or, if prefered, the lowest perceived distress. In addition, the Self Disciplined individuals also had a greater number of units (-0.06) below the mean than the Non-Resilients, whose units were above the mean (0.57).

DISCUSSION

The objective of this study was to extract personality prototypes from the general population of Mexico and to compare subjective health indicators among them. The common method to find prototypes has been cluster analysis. Considering that this method is dependent on the sample characteristics, the use of a larger sample splitted into two groups or the inclusion of two samples are common criteria to confirm the prototypes; this strategy is regarded as a type of replication or validation of the recovered prototypes (e.g., Herzberg & Roth, 2006). In addition, although age effects on the prototypes are not consistent, to minimize their potential effects on the composition of the sample for this study, this was divided arbitrarily into two large age groups: the young group (YG) in a narrow range (14-25 years) and a mature group (MG) with a wide range of ages (26-63 years).

Three prototypes were consistent in both groups. The Undercontrolled was not found which, as we saw earlier, is the least consistent. Two of the recovered prototypes correspond to the commonly in a large number of coun tries, particularly in Europe (Alessandri et al., 2014; Donellan & Robins, 2010; Herzberg & Roth, 2006). In both age groups, the Resilient was characterized by a low score in Neuroticism and high scores in the dimensions of Openness to experience, Agreeableness and, above all, in Extraversion and Conscientiousness; this configuration is similar to other studies (Herzberg, 2009; Herzberg & Roth, 2006; Kinnunen et al., 2012).

The prototype traditionally known as Overcontrolled has been labeled in other studies as Non-Desirable (Barbaranelli, 2002). But in this study it was preferred to name it Non-Resilient, as proposed by Zawadzki and Strelau (2003), because it is clearly the opposite of the Resilient (Alessandri et al., 2014), and because it shows a low score in Conscientiousness, which has been related to self-control (Strus et al., 2014); this prototype configuration is similar to others (e.g., Ekehammar & Akrami, 2003).

Note: for "health status" to more negative score, the groups agreed to have health problems. For GHQ to greater positive score, the groups presented greater health problem. ***.001; **.01; *.05.

The third prototype obtained in this study presented scores closer to the average or slightly high (they did not reach V SD above the mean) in all dimensions, with the exception of Conscientiousness, where the score was higher, and hence it is named Self Disciplined prototype; the trait scores of this group were intermediate to those of the other two prototypes. Although it would seem at first glance that this prototype can be confused with "the authentic" Overcontrolled, because of its high score in Conscientiousness (between +1/2 and +3/4 DE), this does not comply with the low score on Extraversion, that most of the studies support (see revisions: Donellan & Robins, 2010; Herzberg & Roth, 2006). In fact, many studies that indicate having recovered the Overcontrolled prototype do not show agreement on the score of Conscientiousness; in some studies the scores were low (e.g., Barbaranelli, 2002; De Fruyt et al., 2002), in others around the mean (e.g., Asendorpf et al., 2001; Boehm et al., 2002), and in other studies the scores were higher (e.g., Van Leeuwen et al., 2004). Undoubtedly, this is one of the reasons for the contradictory findings mentioned extensively in this work.

In the young group the percentages of Resilients and Self Disciplined people were similar, but in Non-Resilients, the percentage was double. However, the distribution of prototypes in the mature group was very different. A mini mum percentage corresponded to Resilient individuals, but again the Non-Resilient concentrated the majority of the people, with an intermediate percentage of Self Disciplined individuals. In some studies the prototypes were homoge neously distributed, one third for each (Herzberg, 2009), but this is not supported by other studies (Asendorpf et al., 2001; Bohem et al., 2002; Kinnunen et al., 2012; Van Leeuwen et al., 2004).

Until today, the reason for the heterogeneity in the size of the prototypes can be awarded vaguely to the sample characteristics. In any case, not only the size but the number of recovered prototypes add to the results that, throughout more than 20 years, have promoted the argument that it is premature to generalize prototypes in multiple cultures. Recent studies support the "traditional" solution of three clusters (Alessandri et al., 2014), as well as of four (Isler et al., 2016).

There is still not enough evidence to explain why there are differences in the prototypes' distribution between men and women. This is probably due to the heterogeneity of the results (Herzberg, 2009). In this work, the only consistent result was that women were over-represented in the Non-Resilient prototype, in both groups. And, for age, there were no consistent results. A result that draws attention to this study is the small size of the Resilient prototype in the MG. It should be recalled that in this group, from two to five clusters were extracted, but in the four- and five- cluster solutions, the Resilient prototype remained composed by the same individuals (data not shown). In addition, the four- and five- cluster solutions had a configuration with either high or low scores in all traits, which does not allow a clear interpretation. In any case, trying to explain the reason of findings such as the low number of Resilients compared with the high number of Non-Resilients as well as the finding of the Self Disciplined individuals, is a complex task in the absence of similar studies. However, there is an ethno-psychological theory for the Mexican culture with empirical evidence of personality types (Díaz-Guerrero, 2012), which, facing these results, seems appropriate to consider.

Díaz-Guerrero (2012) mentions eight personality types, product of a cultural pattern which favors authoritarian relationships and an important confusion between love and power (Díaz-Guerrero, 2012, p. 76). Four of these types describe the vast majority of Mexican people according to the evidence collected over more than 25 years: the passive andobedient-affiliative, the active self-affirmative rebel, the one with active internal control, and the one with passive external control.

It may be interesting to compare the characteristics attributed to these specific personality types of Mexicans with the prototypes proposed with an Universalist model, such as the FFM (Costa & McCrae, 2008). The obedient-affiliative is described as methodical, disciplined, orderly, plan their things, patient and cautious, and also displays little impulsivity, spontaneity and excitability. It is worth noting that the characteristics of this type are very similar to those of the prototype that was named Self Disciplined. particularly because in the latter the central feature is the self-discipline or self-control.

The active self-affirmative rebel shows characteristics of aggressiveness and dominance, and also exhibits impulsivi-ty, emotional problems, anxiety, and adaptation problems; this type is very similar to the Non-Resilient prototype. The active internal control type, which is regarded as the least common of the Mexican types, is characteristic of committed individuals, ordered, less aggressive, and, above all, who exhibit more internal resources to address the problems. These personality traits resemble the Resilient prototype. Díaz-Guerrero (2012) indicates that the passive external control type resembles the active self-affirmative rebel, but with lower intellectual and academic skills. However, with data from the present study it is not possible to diffe rentiate this type.

The prototypes recovered in this study were compared by subjective health indicators in two ways: First, through the question: "do you suffer from any health problem?," and second,with the GHQ-12 scores. There were no im portant differences in response to the question on health status in the young group, since the majority answered not having problems, as would be expected given their age. The Non-Resilients of the mature group, on the contrary, reported more health problems than the Resilients. As for the GHQ-12, the differences were significant for positive and negative items, although they were clearer for positive items in both groups. Thus, it is confirmed that the persona lity scores of Resilients, in both sexes, are associated with better subjective health, and the opposite configuration of the Non-Resilient or Overcontrolled, according to other studies (Chapman & Goldberg, 2011), is related to the worst health. To avoid confusion, it is important to clarify that Kinnunen et al. (2012) named Overcontrolled the prototype with similar configuration to the one called Non-Resilient prototype in this study.

Finally, it is concluded that the person-centred research is an important complement to the research of individual differences. To our knowledge, this is the first study of personality prototypes carried out in Latin America,which makes it difficult to compare the prototypes recovered in this study. However, the fact of having the Resilient and the Non-Resilient prototypes, as well as having recovered them in two age groups provides confidence in its consistency. More interesting is to have found the prototype named Self Disciplined, but more studies are needed to confirm this.

A limitation of this study is not having included objective health measures (e.g. body mass index, blood pressure, among others), but resources were not available to do so. However, person-centred studies favors the development of useful models to relate and/or predict behavior in real life, such as the health status of individuals. Thus, the fact that the prototypes showed differences in subjective health may be useful knowledge to be considered in the actions necessary to prevent non-transmittable diseases that adver sely affect the Latin American population.

REFERENCES

Akse, J., Hale, W. W. III, Engels, R. C. M. E., Raaijmakers, Q. A. W. & Meeus, W. H. J. (2004). Personality, perceived parental rejection and problem behavior in adolescence. So cial Psychiatry and Psychiatric Epidemiology, 39, 980-988. [ Links ]

Akse, J., Hale, W W. III., Engels, R. C. M. E., Raaijmakers, Q. A. W. & Meeus, W. H. J. (2007). Co-occurrence of depres sion and delinquency in personality types. European Jour nal of Personality, 21, 235-256. [ Links ]

Alessandri, G., Vecchione, M., Donnellan, B. M., Eisenberg, N., Caprara, G. V. & Cieciuch, J. (2014). On the cross-cultu ral replicability of the resilient, undercontrolled, and over-controlled personality types. Journal of Personality, 82(4), 340-353. [ Links ]

Asendorpf, J. B., Borkenau, P., Ostendorpf, F. & van Aken, M. A. G. (2001). Carving personality description at its joints: Confirmation of three replicable personality prototypes for both child and adults. EuropeanJournal of Personality , 15, 169-198. doi: 10.1002/per.408. [ Links ]

Avdeyeva, T. V. & Church, A. T. (2005). The cross-cultural gen-eralizability of personality types: A Philippine study. Euro peanJournal of Personality , 19, 475-499. [ Links ]

Barbaranelli, C. (2002). Evaluating cluster analysis solutions: An application to the Italian NEO Personality Inventory. EuropeanJournal of Personality , 16, s43-s55. [ Links ]

Boehm, B., Asendorpf, J. B. & Avia, M. D. (2002). Replicable types and subtypes of personality: Spanish NEO-PI sam ples. EuropeanJournal of Personality , 16, s25-s41. [ Links ]

Borkenau, P., Hrebícková, M., Kuppens, P., Realo, A. & Al-lik, J. (2013). Sex differences in variability in personality: A study in four samples. Journal of Personality , 81(1), 49-60. doi: 10.1111/j.1467-6494.2012.00784.x. [ Links ]

Caspi, A. (1998). Personality development accross the life course. En W. Damon, & N. Eisenberg (Eds.), Handbook of child psychology: Vol. 3. Social, emotional, and personality development (pp. 311-388). New York: Wiley. [ Links ]

Chapman, B. P. & Goldberg, L. G. (2011). Replicability and 40-year predictive power of childhood ARC types. Journal of Personality and Social Psychology, 101(3), 593-606. doi: 10.1037/a0024289. [ Links ]

Costa, P. T. Jr. & McCrae, R. (2008). Inventario de personali dad NEO Revisado (NEO PI-R). Inventario NEO Reducido de Cinco Factores (NEO-FFI). Manual Profesional (3ra. edición revisada y ampliada). Madrid, España: TEA Edi ciones S. A. [ Links ]

Church, A. T. (2016). Personality traits across cultures. Current Opinion in Psychology, 8, 22-30. [ Links ]

De Fruyt, F., Mervielde, I. & Van Leeuwen, K. (2002). The con sistency of personality type classification across samples and five-factor measures. EuropeanJournal of Personality , 16, s57-s72. doi: 10.1002/per.444. [ Links ]

Díaz-Guerrero, R. (2012). Psicología del Mexicano. Descu brimiento de la etnopsicología (6ta. ed., reimpresión). México, DF: Trillas. [ Links ]

Donnellan, M. B. & Robins, R. W. (2010). Resilient, overcon-trolled, and undercontrolled personality types: Issues and controversies. Social and Personality Psychology Compass 3, 1-14. doi: 10.1111/j.1751-9004.2010.00313.x. [ Links ]

Ekehammar, B. & Akrami, N. (2003). The relation between per sonality and prejudice: A variable- and a person-centered approach. EuropeanJournal of Personality , 17, 449-464. doi: 10.1002/per.494. [ Links ]

García, O. Aluja, A. & García, L. F. (2004). Psychometric prop erties of the Goldberg's 50 Personality Markers for the Big Five Model: A study in Spanish language. European Jour nal of Psychological Assessment, 20(4), 310-319. [ Links ]

Goldberg, D. P. & Williams, P. (1988). A user's guide to the General Health Questionnaire. Windsor, UK: NFER-Nelson. [ Links ]

González, S., Tello, J., Silva, P., Lüders, C., Butelmann, S., Fristch, R., Solar, F., et al. (2012). Calidad de vida en pa cientes con discapacidad motora según factores sociode-mográficos y salud mental. Revista Chilena de Neuro-Psiquiatría, 50(1), 23-34. [ Links ]

Hankins, M. (2008). The reliability of the twelve-item General Health Questionnaire (GHQ-12) under realistic assump tions. BMC Public Health, 8, 355. [ Links ]

Hart, D., Burock, D., London, B., Atkins, R. & Bonilla-Santi-ago, G. (2005). The relation of personality types to physi ological, behavioural, and cognitive processes. EuropeanJournal of Personality , 19, 391-407. [ Links ]

Hernández, R., Fernández, C. & Baptista, P. (2010). Metodología de la investigación. (5ta. ed.). Peru: McGraw-Hill. [ Links ]

Herzberg, P. Y. (2009). Beyond ''accident-proneness": Using five-factor model prototypes to predict driving behavior. Journal of Research in Personality, 43, 1096-1100. doi: 10.1016/j.jrp.2009.08.008. [ Links ]

Herzberg, P. Y. & Roth, M. (2006). Beyond resilients, under-controllers, and overcontrollers? An extension of personal ity prototype research. EuropeanJournal of Personality , 20, 5-28. doi: 10.1002/per.557. [ Links ]

Hoyle, R. H. (2010). Personality and self-regulation. En Rick H. Hoyle (Ed.), Handbook of personality and self-regula tion (pp. 1-18). West Sussex, UK: Wiley-Blackwell. [ Links ]

Isler, L., Liu, J. H., Sibley, C. G. & Fletcher, G. J. O. (2016). Self-Regulation and personality profiles: Empirical devel opment, longitudinal stability and predictive ability. Euro peanJournal of Personality , 30(3), 274-287. doi: 10.1002/per.2054. [ Links ]

Kaleta, D., Polariska, K., Dziankowska-Zaborszcsyk, E., Han ke, W. & Drygas, W. (2009). Factors influencing self-per ception of health status. Central European Journal of Pub lic Health, 17(3), 122-127. [ Links ]

Kinnunen, M. L., Metsäpelto, R. L., Feldt, T., Kokko, K., Tolvanen, A., Kinnunen, U., Leppänen, E., et al. (2012). Personality profiles and health: Longitudinal evidence among Finnish adults. Scandinavian Journal of Psychol ogy 53, 512-522. http://dx.doi.org/10.1111/j.14679450.2012.00969.x. [ Links ]

Klimstra, T. A., Hale, W. W.. III, Raaijmakers, Q. A. W., Branje, S. J. T. & Meeus, W. H. J. (2010). A developmental typol ogy of adolescent personality. European Journal of Person ality, 24, 309-323. [ Links ]

Lahey, B. B. (2009). Public health significance of neuroticism. American Psychologist, 64, 241-256. [ Links ]

Martínez, P. & Cassaretto, M. (2011). Validación del Inventario de los Cinco Factores NEO-FFI en español en estudiantes universitarios peruanos. Revista Mexicana de Psicología, 28, 63-74. [ Links ]

Meda Lara, R. M., Moreno-Jiménez, B., García, L. F., Palomera Chávez, A. & Mariscal de Santiago, M. V. (2015). Validez factorial del NEO-FFI en una muestra mexicana: propuesta de una versión reducida. Revista Mexicana de Psicología, 32(1), 57-67. [ Links ]

McDowell, I. (2006). Measuring health - a guide to rating scales and questionnaires (3aed.). New York: Oxford Press. [ Links ]

Miller, J. D. (2012). Five-factor model personality disorder pro totypes: A review of their development, validity, and com parison to alternative approaches. Journal of Personality , 80(6), 1565-1591. [ Links ]

Roberts, B. W., Kuncel, N., Shiner, R., Caspi, A. & Goldberg, L. R. (2007). The power of personality: The comparative validity of personality traits, socioeconomic status, and cognitive ability for predicting important life outcomes. Perspectives on Psychological Science, 2, 313-345. [ Links ]

Robins, R. W., John, O. P., Caspi, A., Moffitt, T. E. & Stoutham-er-Loeber, M. (1996). Resilient, overcontrolled, and under-controlled boys: Three replicable personality types. Journal of Personality and Social Psychology, 70, 157-171. [ Links ]

Roth, M. & von Collani, G. (2007). A head-to-head comparison of big five types and traits in the prediction of social atti tudes: Further evidence for a five-cluster typology. Journal of Individual Differences, 28, 138-149. [ Links ]

Sánchez-López, M. P. & Dresch, V. (2008). The 12-item Gen eral Health Questionnaire (GHQ-12): Reliability, external validity and factor structure in the Spanish population. Psicothema, 20(4), 839-843. [ Links ]

Smith, T. W., Williams, P. G. & Segerstrom, S. C. (2015). Per sonality and physical health. En M. Mikulincer & P. R. Sha ver (Editors-in-Chief), Handbook of personality and social psychology: Vol. 4. Personality processes and individual differences (pp. 639-661. Washington: American Psycho logical Association. http://dx.doi.org/10.1037/14343-029. [ Links ]

Smith, A. B., Oluboyede, Y., West, R., Hewison, J. & House, A. O. (2013). The factor structure of the GHQ-12: the interac tion between item phrasing, variance and levels of distress. Quality of Life Research, 22, 145-152. [ Links ]

Sociedad Mexicana de Psicología. (2007). Código Ético del Psicólogo. México: Trillas. [ Links ]

Solís-Cámara, P., Meda Lara, R. M., Moreno-Jiménez, B. & Juárez, P. (2016). Estructura factorial del Cuestionario de Salud General GHQ-12 en población general de México. Salud & Sociedad, 7(1), 62-76. [ Links ]

Steca, P., Alessandri, G. & Caprara, G. V. (2010). The utility of a well-known personality typology in studying successful aging: Resilients, undercontrollers, and overcontrollers in old age. Personality and Individual Differences, 48, 442- 446. [ Links ]

Steinley, D. (2003). Local optima in K-means clustering: What you don't know may hurt you. Psychological Methods, 8, 294-304. [ Links ]

Strus, W., Cieciuch, J. & Rowinski, T. (2014). The circumplex of personality metatraits: A synthesizing model of personal ity based on the big five. Review of General Psychology 18, (4), 273-286. http://dx.doi.org/10.1037/gpr0000017. [ Links ]

Tuuliainen/Kirsi Sipilä, L., Mäki, P., Könönen, M. & Suominen, A. L. (2015). Association between clinical signs of tempo-romandibular disorders and psychological distress among an adult Finnish population. Journal of Oral & Facial Pain andHeadache, 29(4), 370-377. doi: 10.11607/ofph.1439. [ Links ]

Urzúa, A., Caqueo-Urízar, A., Bargsted, M. & Irarrázaval, M. (2015). ¿Afecta la forma de puntuación la estructura facto rial de GHQ-12? Cadernos de Saúde Pública, 31(6), 1305- 1312. [ Links ]

Van Leeuwen, K., De Fruyt, F. & Mervielde, I. (2004). A lon gitudinal study of the utility of the resilient, overcontro-lled, and undercontrolled personality types as predictors of children's and adolescents' problem behavior. International Journal of Behavioral Development, 28, 210-220. Reco vered from http://www.tandf.co.uk/journals/pp/01650254. html. [ Links ]

Weir, R. C. & Gjerde, P. F. (2002). Preschool personality pro totypes: Internal coherence, cross-study replicability, and developmental outcomes in adolescence. Personality and Social Psychology Bulletin, 28, 1229-1241. [ Links ]

Xie, X., Chen,W., Lei, L., Xing, C. & Zhang, Y. (2016). The relationship between personality types and prosocial beha vior and aggression in Chinese adolescents. Personality and Individual Differences , 95, 56-61. [ Links ]

Zawadzki, B. & Strelau, J. (2003). Trzy podstawowe typy czy cztery struktury temperamentu? [Three basic types or four structures of temperament?]. Czasopismo Psychologiczne, 6, 271-285. [ Links ]

Referencia: Solís-Cámara R., P., Meda Lara, R.M., Moreno Jiménez, B., Palomera Chávez, A. & Juárez Rodríguez, P. (2017). Comparación de la salud subjetiva entre prototi pos de personalidad recuperados en población general de México. Acta colombiana de Psicología, 20(2), 214-226. doi: http://www.dx.doi.org/10.14718/ACP.2017.20.2.10

Received: August 29, 2016; Revised: January 16, 2017; Accepted: May 05, 2017

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