Perfectionism and school anxiety: More evidence about the 2 × 2 model of perfectionism in an Ecuadorian population

Abstract

The purpose of this study was to examine the 2 × 2 model of dispositional perfectionism via a person-oriented approach in a sample of 1588 Ecuadorian students aged 12- to 18-years-old (M = 14.83, SD = 1.86). The Child and Adolescent Perfectionism Scale (CAPS) and the School Anxiety Inventory (IAES) were employed. The statistical analysis included a non-hierarchical cluster analysis, analysis of variance (ANOVA), effect size (Cohen's d index), and bivariate and partial correlations. Four perfectionism profiles were obtained by using cluster analysis: Non-Perfectionism, Mixed Perfectionism, Pure Perfectionistic Strivings (Pure PS) and Pure Perfectionistic Concerns (Pure PC). Mixed Perfectionism and Pure Perfectionistic Strivings were the most maladaptive profiles whereas Non-Perfectionism was the most adaptive profile. Pure PC did not significantly differ from Mixed Perfectionism and Pure PS, with the exception of the mean scores on Anxiety about Academic Evaluation and Cognitive Anxiety. Results from bivariate and partial correlations revealed that both Perfectionistic Strivings and Perfectionistic Concerns showed shared relationships with all the dimensions of IAES. However, only Perfectionistic Concerns showed a consistent pattern of unique relationships with school anxiety. The current study provides evidenced about the importance of implementing strategies to prevent school anxiety in Ecuadorian population focused on Perfectionistic Concerns.

Keywords

2 × 2 model of perfectionism cluster analysis Ecuadorian teenagers school anxiety

Perfectionism is a multidimensional personality disposition characterized by the pursuit of perfection and the establishment of excessively high standards of performance, combined with too critical evaluation of one's behavior, as well as beliefs about perfectionistic demands and criticisms from others (Hewitt & Flett, 2004; Stoeber, 2018a). The two-factor theory of perfectionism defends that it is possible to classify any perfectionistic dimension in two higher order dimensions: One adaptive and another maladaptive, commonly labeled as Perfectionistic Strivings (PS) and Perfectionistic Concerns (PC), respectively (Stoeber, 2018a). On the basis of this theory, the 2 × 2 model of dispositional perfectionism (Gaudreau & Thompson, 2010) distinguishes four subtypes of perfectionism from the combination of high and low levels of both dimensions: Non-Perfectionism (low PS and PC), Pure PC (low PS and high PC), Pure PS (high PS and low PC) and Mixed Perfectionism (high PS and PC). Additionally, the 2 × 2 model establishes four hypotheses about the results of each profile in terms of psychological adjustment: Pure PS is the most adaptive subtype (Hypothesis 1), Pure PC is the most maladaptive (Hypothesis 2), and Mixed Perfectionism is more adaptive than Pure PC (Hypothesis 3) but more maladaptive than Pure PS (Hypothesis 4).

Perfectionism across cultures

The question about ethnic, cultural, and national differences in the study of perfectionism is under-researched (Stoeber, 2018b). Practically all the accumulated knowledge about perfectionism is based on research carried out with Caucasian populations from North America or the UK. However, it cannot be assumed that the conclusions obtained in one culture can be necessarily extrapolated to others.

Several studies have shown the existence of certain differences in the way in which perfectionism is linked to various psychological variables of maladjustment that can be attributed to cultural mechanisms, such as collectivism/individualism or parental style (see Dibartodo & Rendón, 2012, for a review). Thus, it is possible that certain dimensions of perfectionism, such as PC, characterized by the evaluative concerns of the environment, are especially sensitive to these cultural factors. For example, it is possible that the fear of disappointing the family group is greater in collectivist cultures than in those characterized by greater independence. However, at the same time, the negative consequences of these social pressures could be compensated for by the feeling of support and affection of the family group (Vicent, Inglés, Gonzálvez, Sanmartín, Ortega-Sandoval, & García-Fernández, 2017). In any case, it is necessary to promote research on perfectionism in other ethnic groups, not only to look for possible cultural differences but also for similarities. Thus, the biases in the publication of studies that only find significant differences between cultures will be avoided (Stoeber, 2018b).

Cultural differences in the 2 × 2 model of dispositional perfectionism

After almost a decade from the development of the 2 × 2 model, a considerable number of studies have been published in order to test it. However, the model presents limitations in terms of generalization to other cultures because its hypotheses were developed on North-American populations and most of the subsequent studies have focused on Canadian (e.g., Kljajic, Gaudreau, & Franche, 2017; Lagoutina, 2017), American (e.g., Gong, Fletcher, & Paulson, 2017; Speirs Neumeister, Fletcher, & Burney, 2015), or English samples (e.g., Hill and Davis, 2014; Nordin-Bates, Raedeke, & Madigan, 2017).

Currently, there are a few but increasing number of works that offer empirical evidence about the application of the 2 × 2 model in non-English speaking populations. Franche, Gaudreau, and Miranda (2012) were the first authors to detect the existence of cultural differences in the results for each subtype in terms of academic adjustment when comparing two samples of Canadian university students of Asian and Caucasian origin. Afterwards, Li, Hou, Chi, and Hager (2014) replicated the four perfectionist subtypes of the 2 × 2 model in a Chinese adult population by using cluster analysis. However, results did not support Hypotheses 2 and 3 of the model, because the Mixed group and Pure PC did not statistically differ on their burnout levels. Damian, Stoeber, Negru, and Baban (2014) analysed the association between each perfectionist subtype and positive and negative affect in a Romanian adolescent population by using hierarchical regression analysis. Authors obtained empirical support for the four hypotheses of the 2 × 2 model in the case of positive affect. Nevertheless, Mixed Perfectionism and Pure PC did not significantly differ in their association with negative affect, also rejecting Hypotheses 2 and 3 of the model.

To our knowledge, only three studies have tested the 2 × 2 model in Spanish populations. Méndez-Giménez, Cecchini-Estrada, and Fernández-Río (2015) employed hierarchical regression methodology to examine the 2 × 2 model in students of secondary education in the specific context of physical education, and analysed the link between each perfectionistic subtype and variables such as self-concept, affectivity, and satisfaction with life. Results supported the four hypotheses of the 2 × 2 model. In contrast, Inglés, García-Fernández, Vicent, Gonzálvez, and Sanmartín (2016) found differences between their results and the hypotheses of the model (especially Hypotheses 2 and 3) using cluster analysis in a children population aged 8- to 12-years-old. The most remarkable finding was that the Mixed subtype was even more maladaptive than Pure PC for most of the school anxiety factors evaluated, except for Victimization Anxiety and Social Evaluation Anxiety, whose differences did not reach statistical significance. Subsequently, also in a child population and using cluster analysis, Vicent, Inglés, Sanmartín, Gonzálvez, and García-Fernández (2017) observed that the Mixed profile was the most maladaptive in terms of physical and verbal aggression, anger and hostility, questioning Hypotheses 2 and 3 of the model.

Lastly, in a Latin-American population, Arana and Furlan (2016), on the basis of a clustering approach and employing a sample of Argentinian undergraduates, concluded that Mixed Perfectionism did not significantly differ from Pure PC or it obtained worse results in terms of anxiety and pre-exam coping. Similarly, Vicent, Inglés, Gonzálvez, et al., (2017) examined the 2 × 2 model in an Ecuadorian adolescent population and found that the profile Pure PC did not significantly differ from Mixed and Pure PS on anxiety, depression and stress scores.

To sum up, all known studies that have tested the 2 × 2 model in non-English samples, except for the work of Méndez-Giménez et al. (2015), have questioned the postulates of the model. Specifically, the majority of the studies agree on the fact that Mixed subtype either did not significantly differ from Pure PC or it reported higher maladaptive results. In this way, it is questioned the fact that Pure PC is the most damaging subtype of the four (i.e., Hypotheses 2 and 3). These differences could be explained due to the existence of certain sociocultural factors that influence the way in which perfectionism is manifested across culture (Dibartolo & Rendón, 2012). Thus, it is necessary to expand the field of study about the 2 × 2 model to other cultures and ethnic minorities because conclusions obtained in Caucasian populations may not be generalizable to other populations (Vicent, Inglés, Gonzálvez, et al., 2017).

As noted above, only two studies have analysed the 2 × 2 model in Latin American populations (Arana & Furlan, 2016; Vicent, Inglés, Gonzálvez, et al., 2017). Both studies agreed on the need to increase research on perfectionism in the Latino culture. Thus, this study aims to test the validity of the 2 × 2 model in an Ecuadorian adolescent population. Specifically, (a) this study intended to verify whether it is possible to find, through cluster analyses, four perfectionism profiles similar to those proposed by the 2 × 2 model. Taking into account the results of Vicent, Inglés, Gonzálvez, et al. (2017), it is expected that cluster analyses will result in a solution of four profiles, made by PS and PC levels, which will be consistent with those subtypes defined by the 2 × 2 model.

Secondly, (b) this study intended to examine the existence of differences on the mean scores of school anxiety reported by each perfectionist profile. Previous studies focused on the analysis of the relationship between perfectionism and test anxiety (e.g., Arana & Furlan, 2016; Finn & Guay, 2014; Mills & Blankstein, 2000; Stoeber, Feast, & Hayward, 2009; Tarklan & Alinezhad, 2016). However, apart from the school evaluation, there are other academic situations that can also generate anxiety, such as school punishment or peer aggression. According to Endler (1975), there are different features of anxiety linked to specific situations, so that each person will show a different profile of anxiety traits and behaviors depending on the different situational areas. Thus, it is important to consider the multidimensionality of the anxiety construct inside the school context because each child could manifest a different profile of traits and behaviors of school anxiety depending on the situation (Kearney, Eisen, & Silverman, 1995). In fact, different school anxiety situations have shown different patterns of development, gender differences (Morris, Finkelstein, & Fisher, 1976) and psychological outcomes (Delgado et al., 2018; Puklek, Inglés, Marzo, & García-Fernández, 2015), which supports the idea that although they are highly interconnected, they measure different constructs. In this study, we consider school anxiety as a broad construct that encompasses a pattern of cognitive, motor, and physiological anxious responses, caused by stressful school situations (i.e., academic failure and punishment, aggression, social evaluation, and academic evaluation) that students perceived as dangerous or threatening (Inglés, Gonzálvez, García-Fernández, Vicent, & Martínez-Monteagudo, 2015). Hence, these situations must be considered during the assessment of specific anxiety inside the academic context (García-Fernández, Inglés, Martínez-Monteagudo, Marzo, & Estévez, 2011). To our knowledge, only the study of Inglés et al. (2016) provides data about the relationship between perfectionism and school anxiety. As stated, this work was performed with a Spanish child population and questioned some of the 2 × 2 model hypotheses. The authors concluded that Mixed Perfectionism was linked with higher levels of school anxiety than the rest of profiles. However, the data offered by the study does not allow us to conclude which perfectionist facet (i.e., PS or PC) would explain the results obtained by each profile. In the same way, it is not possible to generalize the results obtained in the Spanish children to Ecuadorian adolescent populations.

In accordance with the study of Inglés et al. (2016) and previous empirical evidence in Latin American populations (Arana & Furlan, 2016; Vicent, Inglés, Gonzálvez et al., 2017), it is expected that the clustering solution will replicate the four perfectionist subtypes defined by the 2 × 2 model, but these profiles will not confirm Hypotheses 2 and 3 raised by Gaudreau and Thompson (2010). Specifically, it is expected that either Mixed Perfectionism will obtain significantly higher levels of school anxiety in comparison with Pure PC or both subtypes will not significantly differ between them.

Lastly, following the recommendations of Stoeber (2018b) and Stoeber and Gaudreau (2017), the bivariate and partial correlations between perfectionist dimensions (i.e., PS and PC) and school anxiety will be analysed in order to understand the shared and unique relationship between each perfectionist dimension and school anxiety factors. It is expected that PS and PC will show positive and significant bivariate correlations with school anxiety factors. In contrast, it is expected that positive and significant partial correlations between PC and school anxiety will be found, as well as non-significant correlations in the case of PS. These expectations are based on previous studies performed by Mills and Blankstein (2000) and Stoeber et al. (2009), who examined the unique and shared relationships between perfectionist dimensions and test anxiety.

Method

Participants

A sample consisting of 1588 students (39.1% females) was recruited by multistage random cluster sampling. Students were enrolled in 11 high schools from Quito (Ecuador). Participants aged 12- to 18-years-old (M = 14.83, SD = 1.86) respectively represented 11.8%, 17.8%, 17.6%, 14.2%, 15%, 14.8%, and 8.9% of the sample. Socio-economic distribution corresponded mainly to the average level (23% Medium-Low, 60% Medium, and 17% Medium-High).

Instruments

The Child and Adolescent Perfectionism Scale (CAPS)

The CAPS (Flett et al., 2016) is a 22-item self-report measure comprised by two dimensions: Self-Oriented Perfectionism (12 items; e.g., ‘I try to be perfect in everything I do’) and Socially Prescribed Perfectionism (10 items; e.g., ‘Other people always expect me to be perfect’). Participants responded to all items on a five-point Likert scale. In accordance with previous research, Self-Oriented Perfectionism and Socially Prescribed Perfectionism were used, respectively, as indicators of PS and PC (Stoeber, 2018a). The Spanish translation of the CAPS was employed (Castro et al., 2004).

A reliability generalization meta-analysis evidenced that the 22-item version of the CAPS is a reliable instrument to be employed with general research purposes (Vicent, Rubio-Aparicio, Sánchez-Meca, & Gonzálvez, 2018). Internal consistency, Cronbach's alpha, calculated for the present study was acceptable for the CAPS subscales (see Table 1), considering acceptable values ≥ 0.70, in accordance with the criteria of Nunnally (1978).

The School Anxiety Inventory (Inventario de Ansiedad Escolar; IAES)

The IAES (García-Fernández et al., 2011) is a self-report measure comprised by four situational factors: FI. Anxiety about Academic Failure and Punishment (eight items; ‘If I repeat the year’), FII. Anxiety about Aggression (six items; ‘If I am insulted or threatened’), FIII. Anxiety about Social Evaluation (five items; ‘If I go up to the blackboard’), FIV. Anxiety about Academic Evaluation (four items; ‘If I take a written exam’). These four situational factors are responded by students based on three response systems: Cognitive Anxiety (nine items; ‘I repeat in my head. I think of nothing else’), Behavioral Anxiety (five items; ‘My legs shake’), and Physiological Anxiety (five items; ‘My stomach begins to hurt’). Students have to indicate on a five-point Likert scale how often each answer is given in the situation to which it refers. Internal consistency, Cronbach's alpha, calculated for the present study was acceptable for the IAES subscales (see Table 1), considering acceptable values ≥ 0.70, in accordance with the criteria of Nunnally (1978).

Compliance with Ethical Standards

This research was conducted in accordance with the Helsinki Declaration as revised 1989. The parents or legal guardians of students included in our sample were requested to give their written consent, having previously informed them of the characteristics and goals of the research.

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Procedure

Invitations for the high schools to participate in the study were sent to the head of schools management. Parental or guardian research description statements and consent forms were distributed and collected. Students were advised that participation was voluntary and anonymous. Data were collected by a researcher during school hours, for approximately 45 minutes, in group format and under the supervision of one trained researcher.

Data analysis

We applied a non-hierarchical cluster analysis (quick cluster analysis) to establish a four perfectionism cluster solution. Unlike hierarchical methods, non-hierarchical methods allow researchers to specify a priori the number of desired groups, so the system offers a single cluster solution. Therefore, it allows researchers to test the adequacy of a previous model, such as the 2 × 2 model of dispositional perfectionism. In addition, this method allows an individual who was assigned to a group at a certain point in the process to be reassigned to another group in a later step if this fact optimizes the selection criterion (Clatworthy, Buick, Hankins, Weinman, & Horne, 2005). Moreover, quick cluster analysis is considered the most appropriate procedure to establish profiles in a large sample of subjects (Hair, Anderson, Tatham, & Black, 1998). In accordance with previous research, z scores below -0.5 were considered low levels, z scores between −0.5 and +0.5 were considered moderate, and z scores over +0.5 were considered high (Cumming & Duda, 2012; Inglés et al., 2016; Nordin-Bates et al., 2017; Vicent, Inglés, Gonzálvez et al., 2017; Vicent, Inglés, Sanmartín et al., 2017).

An analysis of variance (ANOVA) was conducted to examine the differences between the four perfectionism profiles in the school anxiety dimensions. Then, post hoc tests (Scheffé method) were carried out to determine between which groups there were statistically significant differences. Effect size (Cohen's d index) was calculated to obtain the magnitude of the observed differences. This index was interpreted based on Cohen's (1988) criteria: Small (values between 0.20 and 0.49), moderate (values between 0.50 and 0.79) and large (values higher than 0.80). Additionally, we computed bivariate and partial correlations between variables controlling for the overlap between the two forms of perfectionism (PS and PC). All data analyses were carried out with the SPSS/IBM 22.0 statistical package.

Results

Perfectionism clusters

The cluster analysis allowed for the identification of four profiles characterized by different levels of PS and PC (see Figure 1). The first cluster, Non-Perfectionism, included 330 adolescents (21.18%) with low levels of PS and PC. The second profile, Mixed Perfectionism, consisted of 393 participants (25.22%) having high levels in both perfectionism dimensions. The third cluster, Pure PC, was constituted by 408 participants (26.19%) with moderate scores on PS and low on PC. Finally, the fourth group, Pure PS, classified 427 subjects (27.41%) with moderate and low levels in PS and PC, respectively.

Figure 1.

Graphic representation of the perfectionism profiles through cluster analysis.

Inter-cluster differences in school anxiety

Results of the ANOVA indicated the existence of statistically significant differences between the four perfectionism profiles for all school anxiety dimensions (see Table 2) with the Mixed Perfectionism and Non-Perfectionism obtaining the highest and lowest means, respectively.

Table 1.

Reliability, bivariate and partial correlations between PS, PC and school anxiety dimensions.

	Bivariate correlations		Partial correlations		α	M	SD
	PS	PC	PS	PC	α	M	SD
Situational factors
FI. Anxiety about Academic Failure and Punishment	0.16***	0.16***	0.09**	0.09**	0.93	36.80	32.50
FII. Anxiety about Aggression	0.17***	0.22***	0.07*	0.16***	0.93	110.58	47.98
FIII. Anxiety about Social Evaluation	0.07**	0.18***	–0.03	0.16***	0.94	59.18	40.62
FIV. Anxiety about Academic Evaluation	0.12***	0.18***	0.04	0.14***	0.91	39.72	24.63
Response systems
Cognitive Anxiety	0.13***	0.19***	0.04	0.15***	0.85	49.84	25.52
Behavioral Anxiety	0.08**	0.15***	0.00	0.13***	0.71	22.43	14.89
Physiological Anxiety	0.12***	0.18***	0.04	0.13***	0.83	38.44	30.23
Total IAES	0.15***	0.21***	0.05	0.16***	0.97	0.97	128.15
α	0.71	0.75	–	–	–	–	–
M	39.43	32.87	–	–	–	–	–
SD	6.47	6.74	–	–	–	–	–

Notes: PC = Perfectionistic Concerns; PS = Perfectionistic Strivings.

*p < 0.05; **p < 0.01; ***p < 0.001.

Post hoc comparisons revealed that both Mixed Perfectionism and Pure PC obtained significantly higher mean scores than Non-perfectionism on all school anxiety dimensions. In contrast, Mixed Perfectionism did not statistically differ from Pure PC, whereas the comparisons between the mixed group and Pure PS were significant in the case of the three first factors of school anxiety, Physiological Anxiety and the total score of the IAES. Pure PC scored significantly higher than Pure PS only in FIV. Anxiety about Academic Evaluation and Cognitive Anxiety. Finally, the comparisons between Pure PS and Non-Perfectionism were significant for all school anxiety dimensions with the exception of FIII. Anxiety about Social Evaluation.

Effect sizes of these differences (Cohen's d index) ranged from 0.07 to 0.67 (see Table 3). Contrasts between Mixed Perfectionism and Non-Perfectionism obtained the largest effect sizes.

Table 2.

Means and standard deviations, effect size, and post hoc contrasts between mean school anxiety scores (IAES) obtained by the four clusters.

School anxiety dimensions	Non- Perfectionism		Mixed Perfectionism		Pure PC		Pure PS		Statistical significance
School anxiety dimensions	M	SD	M	SD	M	SD	M	SD	F _{(3, 1584)}	p
FI	27.00	27.79	43.41	34.20	37.90	33.50	36.08	31.26	16.13	<0.001
FII	91.29	45.61	122.82	47.75	112.65	47.78	111.25	46.40	27.78	<0.001
FIII	48.12	36.94	66.73	41.82	62.68	41.60	56.36	38.91	14.82	<0.001
FIV	32.41	22.49	45.02	25.46	40.36	25.03	39.27	23.63	16.34	<0.001
Cognitive Anxiety	41.31	23.20	55.31	26.05	50.97	26.37	49.30	23.90	19.38	<0.001
Behavioral Anxiety	18.64	13.46	24.44	15.14	23.33	15.22	22.13	14.41	10.41	<0.001
Physiological Anxiety	29.17	26.34	44.62	31.28	40.30	31.22	37.50	29.48	16.95	<0.001
Total IAES	198.84	116.62	277.99	129.55	253.57	131.24	242.96	122.12	24.66	<0.001

Note: FI = Anxiety about Academic Failure and Punishment; FII = FII. Anxiety about Aggression; FIII = FIII. Anxiety about Social Evaluation; FIV = FIV. Anxiety about Academic Evaluation; IAES = School Anxiety Inventory; Pure PC = Pure Perfectionistic Concerns; Pure PS = Pure Perfectionistic Strivings.

Table 3.

Cohen's d index to post hoc contrasts between the school anxiety mean scores obtained by the four perfectionist clusters.

	Mixed vs Pure PC	Mixed vs Pure PS	Mixed vs Non- Perfectionism	Pure PC vs Pure PS	Pure PC vs Non- Perfectionism	Non- Perfectionism vs Pure PS
FI	–	0.22	0.52	–	0.35	0.31
FII	–	0.25	0.67	–	0.46	0.43
FIII	–	0.26	0.47	–	0.37	–
FIV	–	–	0.52	0.06	0.33	0.30
Cognitive Anxiety	–	–	0.56	0.07	0.38	0.34
Behavioral Anxiety	–	–	0.40	–	0.32	0.25
Physiological Anxiety	–	0.23	0.53	–	0.38	0.30
Total IAES	–	0.26	0.64	–	0.44	0.37

Bivariate and partial correlations between perfectionistic dimensions and school anxiety

Table 1 presents the results of bivariate and partial correlations for each form of perfectionism controlling for the other. PC showed positive and significant correlations with school anxiety in both the bivariate and the partial correlations. In contrast, PS also showed positive bivariate correlations with all school anxiety dimensions. These correlations became non-significant when partial correlations were regarded and the overlap between PS and PC was controlled, with the exception of the two first factors (FI. Anxiety about Academic Failure and Punishment and FII. Anxiety about Aggression) whose correlations were significant and positive.

Discussion

Results of the cluster analysis showed a four-profile solution similar to those subtypes defined by the 2 × 2 model (Gaudreau & Thompson, 2010), although with some differences in terms of PS and PC levels obtained by Pure PS and Pure PC. Specifically, the 2 × 2 model defines Pure PS as the combination of high levels of PS and low levels of PC, whereas on the contrary, Pure PC was characterized by high and low levels of PC and PS, respectively. Thus, these two subtypes reflect the opposite combination of the two perfectionist dimensions. However, in accordance with our results, Pure PS and Pure PC profiles present more moderate levels of perfectionism than Mixed and Non-Perfectionism profiles. These differences between the 2 × 2 model and our results could be due to the analytical approach used. While it is true that the 2 × 2 model can be tested through a variable-centered and a person-centered approach (Gaudreau, 2013), its postulates were raised and have been replicated mainly by regression analysis (e.g., Crocker, Gaudreau, Mosewich, & Kljajic, 2014; Gong et al., 2017; Kljajic et al., 2017; Lagoutina, 2017). In fact, most of the investigations that resorted to a person-oriented approach obtained, in accordance with our results, that Pure subtypes reported less extreme z scores than the two remaining subtypes (e.g., Inglés et al., 2016; Nordin-Bates et al., 2017; Vicent, Inglés, Gonzálvez et al., 2017; Vicent, Inglés, Sanmartín et al., 2017). In this sense, it is important to remember that a person-oriented approach focus the study of perfectionism on patterns of beliefs and behaviors. Unlike the variable-oriented approach, which focuses on single variables, a person-oriented approach attempts to state the existence of different profiles of values on perfectionist dimensions, ‘thereby coming closer to the real person’ (Lundh, Saboonchi & Wangby 2008, p.348).

In terms of prevalence, clusters Pure PC (27.41%), Pure PS (26.19%) and Mixed Perfectionism (25.22%) included a similar percentage of subjects, with Pure PC the most numerous. Prevalence for each profile was similar to those previously obtained in an Ecuadorian adolescent population (Vicent, Inglés, Gonzálvez et al., 2017), and in Spanish children (Inglés et al., 2016; Vicent, Inglés, Sanmartín et al., 2017).

Regarding inter-cluster differences, in accordance with previous research in Latin American populations (Arana & Furlan, 2016; Vicent, Inglés, Gonzálvez et al., 2017), Mixed Perfectionism reported the highest levels of school anxiety. However, this profile did not significantly differ from Pure PC, rejecting Hypotheses 2 and 3 of the 2 × 2 model (Gaudreau & Thompson, 2010), which have established that Pure PC is the most maladaptive subtype of the four. Likewise, Pure PS obtained lower levels of school anxiety than Mixed Perfectionism according to Hypothesis 4 of the model, except for Cognitive Anxiety, Behavioral Anxiety and Anxiety about Academic Evaluation; and higher levels than Non-Perfectionism, contradicting Hypothesis 1. The fact that our results do not support the four hypotheses of the model (Gaudreau & Thompson, 2010) and most of previous literature that has tested the model in Caucasian North-American populations (e.g., Gaudreau, 2015; Gaudreau, Franche, & Gareau, 2016; Gong et al., 2017; Kljajic et al., 2017; Lagoutina, 2017; Taylor, Papay, Webb, & Reeve, 2016), could be due to the sociocultural differences that affect the perfectionist dimensions, as Vicent, Inglés, Gonzálvez et al. (2017) suggest; factors such as the role of family in Latin America, a culture which is clearly more collectivistic than the North-American (Oyserman & Wing-Sing, 2007). According to Stoeber (2018b), in order to find out which perfectionist dimension explains the inter-cluster differences, bivariate and partial correlations between PS and PC and school anxiety were examined. Both types of correlations have their own meaning and provide important information which must be interpreted in terms of shared (bivariate correlations) and unique (partial correlations) relationships (Stoeber & Gaudreau, 2017; Stoeber, Nolad, Mawenu, Henderson, & Kent, 2017). Specifically, positive and significant bivariate correlations, of a small magnitude, between PS and school anxiety were found. This result indicates that an individual with high levels of PS is likely to present higher school anxiety than another individual with low levels of PS. However, if two individuals present the same levels of PC, whoever presents higher PS would not be likely to report higher school anxiety scores than the subject with low PS (because when PC remained constant, PS showed non-significant partial correlations with school anxiety, except for the first two factors of the IAES). Concerning PC, positive and significant bivariate and partial correlations were observed, indicating a shared and unique relationship between PC and school anxiety. Similar conclusions were obtained by Mills and Blankstein (2000) and Stoeber et al. (2009) when partial and bivariate correlations between perfectionism and test anxiety were examined in Canadian and English undergraduates, respectively.

Thus, results from correlational analysis suggest that the high mean scores for school anxiety obtained by the Mixed profile may only be caused by high levels on PC (but not on PS). In fact, Mixed Perfectionism did not significantly differ from Pure PC in their mean levels of school anxiety, although this latter profile was characterized by low levels of PS whereas Mixed Perfectionism presented high levels of PS. It is recalled that, in this study, PC is characterized by cognitions about demands and criticisms of the environment, a perfectionist facet closely linked to psychopathology also in young populations (Morris & Lomax, 2014). The tendency to believe that people (such as parents, friends, teachers) expect us to have a perfect performance and they will harshly judge us when we do not reach the imposed bar generates high levels of school anxiety in students, which is justified by the fear of disappointing loved ones and being punished after failing to meet the others' expectations. Thus, students with high levels of PC tend to experience higher school anxiety regardless of their self-requirements and self-criticisms (i.e., PS).

Limitations and future lines of research

Given the cross-sectional design of this study, one of the main limitations is the impossibility of establishing causal relationships, an aspect that could be solved in the future using longitudinal designs. In the same way, the cluster analysis approach presents some limitations, including the fact of producing a cluster solution for a specific sample (Magidson & Vermunt, 2002). Therefore, it is important to note that this study served to replicate not only the cluster structure of Vicent, Inglés, Gonzálvez et al. (2017), but also the fact that both studies found perfectionism profiles in Ecuadorian adolescent populations indicative of the 2 × 2 model of perfectionism. Future research should also test the model in Ecuadorian populations using a variable-oriented approach (e.g. hierarchical regressions) as well as other clustering methods such as Latent Class Analysis and testing the existence of possible gender differences.

The available data in this study have been exclusively collected through self-report measures. Although self-report measures are generally used to assess perfectionism, they are not exempt from problems (Fernández-Montalvo & Echeburúa, 2006). These troubles could be solved by using a multisource and multimethod assessment. Likewise, this work employed as indicators of PS and PC the Self-Oriented Perfectionism and the Socially Prescribed Perfectionism subscales of the CAPS, respectively. The decision to use the CAPS was based on the fact that it is the most used and recommended evaluation instrument in child and adolescent populations (García-Fernández et al., 2016; Leone & Wade, 2017). However, as we have used Socially Prescribed Perfectionism as a measure of PC, we have limited perfectionistic concerns to those of an interpersonal nature. Therefore, it would be interesting in future to analyse if it is possible to replicate these same profiles of perfectionism using other scales of child and adolescent perfectionism, such as the Adaptive/Maladaptive Perfectionism Scale (Rice & Preusser, 2002).

Conclusions and practical implications

This work presents important practical implications. On the one hand, our study contributes to evidence that the four subtypes proposed by Gaudreau and Thompson (2010) are representative, although with some differences, of perfectionist subgroups of Ecuadorian adolescent students. On the other hand, this is the first work that examines the relationship between perfectionism and school anxiety in adolescents. Profiles characterized by high and moderate levels of PC (Mixed Perfectionism and Pure PC) obtained the most maladaptive results, with small and moderate effect sizes for the differences between both subtypes and Non-Perfectionism. It means that these differences are not only significant but they also represent practical and theoretical relevance. Taken together, both profiles (Mixed Perfectionism and Pure PC) represent more than 50% of the analysed population. So, there are a high percentage of Ecuadorian adolescents in risk of presenting anxious symptoms (cognitive, behavioural, and physiological) in different school situations such as failure and punishment, peers' aggression and social and academic evaluation. Hence, it is recommended that school psychologists and other educational professionals implement strategies to prevent school anxiety focused on perfectionism concerns resulted of beliefs about high demands and criticisms of the environment. Taking into account the intrapersonal nature of PC in this study, it is recommended, not only to modify the irrational beliefs of the perfectionists themselves, but also to reflect about criticisms and expectations that families and teachers transmit to young people. Flett and Hewitt (2014) propose that preventive strategies adopt a positive psychology focus and emphasize the development of resilience engaging parents and teachers. According to these authors, parents and teachers who are excessively demanding, over controlling and critical must learn to temper their tendencies, to model appropriate and to promote key meta-cognitive messages to mitigate perfectionistic concerns of our teenagers.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethical information

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author biographies

María Vicent, PhD, is an Assistant Professor in the Department of Developmental Psychology and Teaching at Alicante University (Spain). Her research focuses on perfectionism during childhood and adolescence.

Carolina Gonzálvez, PhD, is a Lecturer in the Department of Developmental Psychology and Teaching at Alicante University (Spain). Her research focuses on school refusal.

Ricardo Sanmartín, PhD, is an Assistant Professor in the Department of Developmental Psychology and Teaching at Alicante University (Spain). His research focuses on positive and negative affect.

Aitana Fernández-Sogorb is a PhD student in the Department of Developmental Psychology and Teaching of the University of Alicante (Spain). Her research centers on the effects of anxiety during childhood.

Nancy Isabel Cargua-García, PhD, is a Lecturer in the Central University of Ecuador. Her area of expertise is school anxiety.

José Manuel García-Fernández, PhD, is a Senior Lecturer of educational psychology in the Department of Developmental Psychology and Teaching at Alicante University. His research focuses on psychometrics and school anxiety.

References

Arana

F. G.

Furlan

(2016) Groups of perfectionists, test anxiety, and pre-exam coping in Argentine students. Personality and Individual Differences 90: 169–173. doi:10.1016/j.paid.2015.11.001.

Castro

Gila

Gual

Lahortiga

Saura

Toro

(2004) Perfectionism dimensions in children and adolescents with anorexia nervosa. Journal of Adolescent Health 35: 392–398. doi:10.1016/j.jadohealth.2003.11.094.

Clatworthy

Buick

Hankins

Weinman

Horne

(2005) The use and reporting of cluster analysis in health psychology: A review. British Journal of Health Psychology 10: 329–358. doi:10.1348/135910705X25697.

Cohen

(1988) Statistical power analysis for the behavioral sciences, Hillsdale, NJ: Erlbaum.

Crocker

P. R. E.

Gaudreau

Mosewich

A. D.

Kljajic

(2014) Perfectionism and the stress process in intercollegiate athletes: Examining the 2 × 2 model of perfectionism in sport competition. International Journal of Sport Psychology 45(4): 61–84. doi:10.7352/IJSP2014.45.325.

Cumming

Duda

J. L.

(2012) Profiles of perfectionism, body-related concerns, and indicators of psychological health in vocational dance students: An investigation of the 2 × 2 model of perfectionism. Psychology of Sport & Exercise 13: 729–738. doi:10.1016/j.psychsport.2012.05.004.

Damian, L. E., Stoeber, J., Negru, O., & Băban, A. (2014). Positive and negative affect in adolescents: An investigation of the 2 × 2 model of perfectionism. Cognition, Brain, Behavior. An interdisciplinary Journal, XVIII (1), 1–16.

Delgado

García-Fernández

J. M.

Martínez-Monteagudo

M. C.

Inglés

C. J.

Marzo

J. C.

La Greca

A. M.

Hugon

(2018) Social anxiety scale for adolescents and school anxiety inventory: psychometric properties in French adolescents. Child Psychiatry & Human Development 50(1): 13–26. doi:10.1007/s10578-018-0818-4.

Dibartolo

P. M.

Rendón

M. J.

(2012) A critical examination of the construct of perfectionism and its relationship to mental health in Asian and African Americans using a cross-cultural framework. Clinical Psychology 32(3): 139–152. doi:10.1016/j.cpr.2011.09.007.

10.

Endler

N. S.

(1975) A person-situation interaction model for anxiety. In: Spielberger

C. A.

Sarason

I. G.

(eds) Stress and anxiety, Washington, DC: Hemisphere, pp. 145–164.

11.

Fernández-Montalvo

Echeburúa

(2006) Uso y abuso de los autoinformes en la evaluación de los trastornos de personalidad. Revista de Psicopatología y Psicología Clínica 11(1): 1–12.

12.

Finn

Guay

M. C.

(2014) Perfectionism and university quota: Is there a link?. Canadian Journal of Behavioural Science 46(2): 252–261. doi:10.1037/a0029618.

13.

Flett

G. L.

Hewitt

P. L.

(2014) A proposed framework for preventing perfectionism and promoting resilience and mental health among vulnerable children and adolescents. Psychology in the Schools 51(9): 899–912. doi:10.1002/pits.21792.

14.

Flett

G. L.

Hewitt

P. L.

Besser

Vaillancourt

Boucher

Gale

(2016) The Child-Adolescent Perfectionism Scale: Development, psychometric properties, and associations with stress, distress, and psychiatric symptoms. Journal of Psychoeducational Assessment 34(7): 634–652. doi:10.1177/0734282916651381.

15.

Franche

Gaudreau

Miranda

(2012) The 2 × 2 model of perfectionism: A comparison across Asian Canadians and European Canadians. Journal of Counseling Psychology 59(4): 567–574. doi: 10.1037/a0028992.

16.

García-Fernández

J. M.

Inglés

C. J.

Martínez-Monteagudo

M. C.

Marzo

J. C.

Estévez

(2011) Inventario de ansiedad escolar: Validación en una muestra de estudiantes de educación secundaria. Psicothema 23(2): 301–307.

17.

García-Fernández

J. M.

Inglés

C. J.

Vicent

Gonzálvez

Gómez-Núñez

M. I.

Poveda-Serra

(2016) Perfeccionismo durante la infancia y la adolescencia. Análisis bibliométrico y temático (2004–2014). [Perfectionism during childhood and adolescence. Bibliometric and thematic analysis (2004–2014)]. Revista Iberoamericana de Psicología y Salud 7(2): 79–88. doi:10.1016/j.rips.2016.02.001.

18.

Gaudreau

(2013) The 2 × 2 model of perfectionism: commenting the critical comments and suggestions of Stoeber (2012). Personality & Individual Differences 55: 351–355. doi:10.1016/j.paid.2013.03.021.

19.

Gaudreau

(2015) Self-assessment of the four subtypes of perfectionism in the 2 × 2 model of perfectionism. Personality & Individual Differences 84: 52–62. doi:10.1016/j.paid.2014.10.039.

20.

Gaudreau

Franche

Gareau

(2016) A latent mediated moderation of perfectionism, motivation, and academic satisfaction: Advancing the 2 × 2 model of perfectionism trough substantive-methodological synergy. Journal of Psychoeducational Assessment 34(7): 688–701. doi:10.1177/0734282916651778.

21.

Gaudreau

Thompson

(2010) Testing a 2 × 2 model of dispositional perfectionism. Personality & Individual Differences 48: 532–537. doi:10.1016/j.paid.2009.11.031.

22.

Gong

Fletcher

K. L.

Paulson

S. F.

(2017) Perfectionism and emotional intelligence: A test of the 2 × 2 model of perfectionism. Personality & Individual Differences 106: 71–76. doi:10.1016/j.paid.2016.10.015.

23.

Hair

J. F.

Anderson

R. E.

Tatham

R. C.

Black

W. C.

(1998) Multivariate data analysis, (5th ed.). Upper Saddle River, NJ: Prentice-Hall.

24.

Hewitt, P. L., & Flett, G. L. (2004). Multidimensional perfectionism scale (MPS): Technical manual. Toronto: Multi-Health Systems.

25.

Hill

A. P.

Davis

P. A.

(2014) Perfectionism and emotion regulation in coaches: A test of the 2 × 2 model of dispositional perfectionism. Motivation & Emotion 38: 715–726. doi:10.1007/s11031-014-9404-7.

26.

Inglés

C. J.

García-Fernández

J. M.

Vicent

Gonzálvez

Sanmartín

(2016) Profiles of perfectionism and school anxiety: A review of the 2 × 2 model of dispositional perfectionism in child population. Frontiers in Psychology 7: 1403 doi:10.3389/fpsyg.2016.01403.

27.

Inglés

C. J.

Gonzálvez

García-Fernández

J. M.

Vicent

Martínez-Montegudo

M. C.

(2015) Current status of research on school refusal. European Journal of Educational Psychology 8: 37–52. doi:10.1016/j.ejeps.2015.10.005.

28.

Kearney

C. A.

Eisen

A. R.

Silverman

W. K.

(1995) The legend and myth of school phobia. School Psychology Quarterly 10: 65–85. doi: 10.1037/h0088293.

29.

Kljajic

Gaudreau

Franche

(2017) An investigation of the 2 × 2 model of perfectionism with burnout engagement, self-regulation, and academic achievement. Learning & Individual Differences 57: 103–113. doi:10.1016/j.lindif.2017.06.004.

30.

Lagoutina

(2017) Perfectionism and time perspective: An inquiry into the association between perfectionism subtypes and categorical time perspectives. Time & Society 26(3): 286–304. doi:10.1177/0961463X15577280.

31.

Leone

Wade

T. D.

(2017) Measuring perfectionism in children: A systematic review of the mental health literature. European Child Adolescent Psychiatry 27(5): 553–567. doi:10.1007/s00787-017-1078-8.

32.

Hou

Z. J.

Chi

H. Y.

Hager

M. J.

(2014) The mediating role of coping in the relationship between subtypes of perfectionism and job burnout: A test of the 2 × 2 model of perfectionism with employees in China. Personality & Individual Differences 58: 65–70. doi:10.1016/j.paid.2013.10.007.

33.

Lundh

Saboonchi

Wangby

(2008) The role of personal standards in clinically significant perfectionism. A person-oriented approach to the study of patterns of perfectionism. Cognitive Therapy & Research 32: 333–350. doi:10.1007/s10608 006-9109-7.

34.

Magidson

Vermunt

J. K.

(2002) Latent class models for clustering: A comparison with k-means. Canadian Journal of Marketing Research 20: 37–44.

35.

Méndez-Giménez

Cechinni-Estrada

J. A.

Fernández-Río

(2015) The 2 × 2 dispositional perfectionism model in adolescents: an assessment in the physical Education Context. Universitas Psychologica 13(4): 15–25. doi:10.11144/Javeriana.UPSY13-4.mpda.

36.

Mills

J. S.

Blankstein

K. R.

(2000) Perfectionism, intrinsic vs extrinsic motivation, and motivated strategies for learning: A multidimensional analysis of university students. Personality & Individual Differences 29: 1191–1204. doi: 10.1016/S0191-8869(00)00003-9.

37.

Morris

L. W.

Finkelstein

C. S.

Fisher

W. R.

(1976) Components of school anxiety: Developmental trends and sex differences. Journal of Genetic Psychology 128(1): 49–57. doi:10.1080/00221325.1976.10533971.

38.

Morris

Lomax

(2014) Review: Assessment, development, and treatment of childhood perfectionism: A systematic review. Child & Adolescent Mental Health 19(4): 225–234. doi:10.1111/camh.12067.

39.

Nordin-Bates

S. M.

Raedeke

T. D.

Madigan

D. J.

(2017) Perfectionism, burnout, and motivation in dance. A replication and test of the 2x2 model of perfectionism. Journal of Dance Medicine & Science 21(3): 115–122. doi:10.12678/1089-313X.21.3.115.

40.

Nunnally

J. C.

(1978) Psychometric theory, New York, NY: McGraw-Hill.

41.

Oyserman

Wing-Sing

(2007) Priming ‘culture’: Culture as situated cognition. In: Kitayama

Cohen

(eds) Handbook of cultural psychology, New York, NY: Guilford Press, pp. 255–282.

42.

Puklek

Inglés

C. J.

Marzo

J. C.

García-Fernández

J. M.

(2015) School anxiety inventory: reliability and validity in a sample of Slovenian adolescents. Psychology in the Schools 52(9): 860–873. doi:10.1002/pits.21862.

43.

Rice

K. G.

Preusser

K. J.

(2002) The Adaptative/Maladaptative Perfectionism Scale. Measurement and Evaluation in Counseling & Development 34(4): 210–222.

44.

Speirs-Neumeister

K. L.

Fletcher

K. L.

Burney

V. H.

(2015) Perfectionism and achievement motivation in high-ability students: An examination of the 2 × 2 model of perfectionism. Journal for the Education of the Gifted 38(3): 215–232. doi:10.1177/0162353215592502.

45.

Stoeber, J. (2018a). The psychology of perfectionism. An introduction. In J. Stoeber (Ed.), The psychology of perfectionism: Theory, research, applications (pp. 3–16). London: Routledge.

46.

Stoeber, J. (2018b). The psychology of perfectionism: Critical issues, open questions, and future directions. In J. Stoeber (Ed.), The psychology of perfectionism: theory, research, applications (pp. 333–352). London: Routledge.

47.

Stoeber

Feast

A. R.

Hayward

J. A.

(2009) Self-oriented and socially prescribed perfectionism: differential relationships with intrinsic and extrinsic motivation and test anxiety. Personality & Individual Differences 47: 423–428. doi:10.1016/j.paid.2009.04.014.

48.

Stoeber

Gaudreau

(2017) The advantages of partialling perfectionistic strivings and perfectionistic concerns: critical issures and recomendations. Personality & Individual Differences 104: 379–386. doi:10.1016/j.paid.2016.08.039.

49.

Stoeber

Noland

A. B.

Mawenu

T. W. N.

Henderson

T. M.

Kent

D. N. P.

(2017) Perfectionism, social disconnection, and interpersonal hostility: not all perfectionists don't play nicely with others. Personality & Individual Differences 119: 112–117. doi:10.1016/j.paid.2017.07.008.

50.

Tarklan

Alinezhad

M. J.

(2016) The relationship between attachment style and negative perfectionism with exam anxiety of students. Entomology & Applied Science Letters 3(2): 1–5. doi:10.6007/IJARP/v2-i1/1717.

51.

Taylor

J. J.

Papay

A. K.

Webb

J. B.

Reeve

C. L.

(2016) The good, the bad, and the interactive: Evaluative concerns perfectionism moderates the effect of personal strivings perfectionism on self-esteem. Personality & Individual Differences 95: 1–5. doi:10.1016/j.paid.2016.02.006.

52.

Vicent

Inglés

C. J.

Gonzálvez

Sanmartín

Ortega-Sandoval

V. N.

García-Fernández

J. M.

(2017) Testing the 2 × 2 model of perfectionism in Ecuadorian adolescent population. Journal of Health Psychology 1–7. doi:10.1177/1359105317733536.

53.

Vicent

Inglés

C. J.

Sanmartín

Gonzálvez

García-Fernández

J. M.

(2017) Perfectionism and aggression: Identifying risk profiles in children. Personality & Individual Differences 112: 106–112. doi:10.1016/j.paid.2017.02.061.

54.

Vicent

Rubio-Aparicio

Sánchez-Meca

Gonzálvez

(2018) A reliability generalization meta-analysis of the child and adolescent perfectionism scale. Journal of Affective Disorders 245(2019): 533–544. doi:10.1016/j.jad.2018.11.049.