Trait Perfectionism,L2 Anxiety,and Willingness to Communicate Among Adolescents: Evaluating the Roles of Worrying About Mistakes and Mistake Rumination

Abstract

The current study investigated the associations among trait perfectionism, worry about mistakes, mistake rumination, L2 anxiety, and willingness to communicate in adolescents. A key element was examining the psychometric properties of a new measure of worry about mistakes as well as the Mistake Rumination Scale and the short Child-Adolescent Perfectionism Scale and its two dimensions (i.e., self-oriented and socially prescribed perfectionism). Our participants were 954 Iranian adolescent English learners. Psychometric tests supported the psychometric properties and use of the mistakes and perfectionism measures. Correlation analyses indicated that both trait perfectionism dimensions were associated with mistake rumination and worry about mistakes, and socially prescribed perfectionism was associated with L2 anxiety. However, self-oriented perfectionism was associated with greater willingness to communicate. Further analyses showed that worry about mistakes and mistake rumination mediated the association that socially prescribed perfectionism had with L2 anxiety. The vulnerabilities of adolescent perfectionists are discussed in terms of the approach-avoidance conflict as it relates to worry about mistakes and pressures to be perfect. It is evident that adolescents who are vulnerable must learn cognitive and emotional regulation techniques to effectively manage mistakes and the pressure to be perfect.

Keywords

perfectionism mistake rumination language learning communication

Introduction

Perfectionism is a vulnerability factor among adolescents that can lead to a host of emotional and affective deficits and learning difficulties. Perfectionistic individuals tend to strive for flawless performance by setting very stringent standards for their performance, while often critically evaluating their behavior. Rather than seeking excellence and development, perfectionists suffer from an all-or-none mentality and when in extreme forms, perfectionists consider anything less than perfect as failure (Hewitt & Flett, 1991). When it comes to the capacity to learn new languages, this orientation can undermine the process of language learning since such perfectionist L2 learners tend to evade the trial-and-error phase of L2 learning (Dewaele, 2012). This avoidance tendency arises from perfectionists’ fear of making mistakes and their tendency to view mistakes as signs of failure. To make it worse, these individuals’ self-worth is typically perceived as contingent upon their perfect performance, and hence, making mistakes constitutes a threat to their ego and self-worth.

On the positive side, however, perfectionism does not necessarily lead to adjustment difficulties because it is a complex multidimensional personality disposition. Perfectionism may occasionally lead to positive outcomes such as positive emotions (Barabadi & Khajavy, 2020) and higher L2 achievement (Chasetareh et al., 2022). However, it should be borne in mind perfectionism in other circumstances is almost entirely problematic (see Stornelli et al., 2009). Most notably, socially prescribed perfectionism (i.e., demands to be perfect imposed on the self by other people) is consistently deleterious in terms of links with measures of distress and demotivation (see Flett et al., 2016a, 2022).

The negative aspects of perfectionism in the context of L2 learning can be studied from a cognitive perspective that highlights the debilitating effect of frequent automatic thoughts of learners, especially with regard to making mistakes (see Barabadi et al., 2022; Flett et al., 2016b). Individuals high in perfectionism show a tendency to brood about past mistakes, making them engage in a form of depressive rumination and causing self-control problems (Flett et al., 2020). Mistake rumination (MR) refers to people’s tendency to dwell on their mistakes and at the same time magnify such mistakes to the point that they cannot come to terms with such mistakes, filling them with a sense of disappointment and failure (Besser et al., 2004).

To our knowledge, no study thus far has investigated how two trait dimensions of perfectionism (Hewitt & Flett, 1991), namely, self-oriented perfectionism (SOP) and socially prescribed perfectionism (SPP) might be related to two cognitive dimensions of perfectionism, namely, mistake rumination (MR) and worry about mistakes (WAM). The current study investigated how these different dimensions of trait perfectionism and cognitive perfectionism contribute to L2 anxiety and Willingness to communicate (WTC).

The current paper examines both substantive and psychometric issues. The psychometric features of the Mistake Rumination Scale have been previously investigated (see Flett et al., 2020) but not in adolescents. This was one psychometric issue addressed in the current study. We also investigated the psychometric properties of the brief Child-Adolescent Perfectionism Scale (see Bento et al., 2019).

More broadly, to our knowledge, there are no measures in existence that directly measure worry about mistakes. Thus, another feature of this study was to create and evaluate a new measure. This new scale was developed on the basis of a comprehensive review of relevant literature on perfectionism, MR, and WAM. While worry about mistakes is a topic that is central to understanding the challenges inherent in language learning (see Flett et al., 2016b), it was our goal to create a general measure that could be used in various contexts and with various age groups to assess individual differences in worry about mistakes.

We investigated the following research questions:

1. Do the Mistake Rumination Scale and the newly created Worry About Mistake Scale have acceptable reliability and validity in the context of L2 learning?

2. Are higher levels of MR and WAM associated with socially prescribed perfectionism among L2 learners?

3. To what extent are SPP, MR, and WAM associated with L2 anxiety and WTC?

4. Do MR and WAM, given their anticipated close ties with language learning, mediate the relationship between trait perfectionism, L2 anxiety, and WTC?

We now consider relevant background information that pertains to the goals of this study. Specifically, we briefly discuss trait perfectionism, and its associated cognitive dimensions, namely, MR and WAM, and how these traits and cognitive aspects may contribute to L2 anxiety. We then introduce L2 willingness to communicate and consider how it may be hampered by this negative chain reaction of perfectionism.

Conceptual Background

The current study reflects our view that when it comes to specific types of learning and performance related to language learning and communication, it is important to consider broad personality dimensions but also more specific types of individual differences with greater and more immediate contextual relevance. This distinction seems particularly important to take into account when conceptualizing how perfectionism can undermine learning and performance. Flett and Hewitt (2022) recently examined how and when perfectionism can amount to too much pressure and undermine the achievement of otherwise capable learners. They described perfectionism in underachievers and how too much emphasis on needing to be perfect in children and adolescents who struggle with self-doubt can undermine their accomplishments and reinforce their focus on personal shortcomings and inadequacies.

The current study represents our initial attempt to explore the distinction between a more distal yet relevant personality orientation (i.e., trait perfectionism) and a more specific orientation toward mistakes. Specifically, as potential correlates of L2 anxiety, we assessed trait perfectionism with the brief Child-Adolescent Perfectionism Scale (CAPS; Bento et al., 2019) and mistake orientation in terms of measures of WAM and MR (see Flett et al., 2020). As indicated above, the version of the CAPS used here was the new brief version (see Bento et al., 2019). Also, as alluded to above, the WAM measure was developed for this study. The five items that comprise this new scale include items such as “I worry about how people will react when I make a mistake” and “Once I worry about making mistakes I cannot stop.” This element of mistake orientation taps into anticipatory concern about the possibility or likelihood of future mistakes and embraces the kind of orientation exhibited by test-anxious students prone to excessive worry (see Sarason, 1984). We felt a new measure was necessary because there is no existing measure created specifically to measure worry about mistakes. Unfortunately, as noted by Flett and Hewitt (2022), the concern over mistakes subscale from the Frost Multidimensional Perfectionism Scale (see Frost et al., 1990) has items that tap both mistake orientation and failure orientation and associated beliefs, and, as such, it does not assess worry about mistakes.

Our focus on mistake rumination is due to its presumed role in distress and vulnerability, including its potential contributions to anxious distress, worry, and fear of negative evaluation. Studies of the daily diary accounts of perfectionists (e.g., Frost et al., 1997) and how perfectionists react to not doing well in performance situations (Besser et al., 2004) have identified clear individual differences in the tendency to ruminate about mistakes. More general research with a qualitative focus shows that thinking about a past mistake is highly salient for many students (see Joubert et al., 2022). Flett et al. (2016a) called for a broader focus on mistake rumination and discussed it as a form of perfectionism reactivity. This emphasis on reactivity reflects the contention that much of the vulnerability of perfectionists is due to maladaptive and self-defeating ways of responding when things are not perfect (i.e., after making a mistake or after having failed on something important). The inability to stop thinking about mistakes is a key element of reactivity that can make perfectionism debilitating in terms of performance difficulties and emotional distress. This element was assessed in the current study with the Mistake Rumination Scale (see Flett et al., 2020). Mistake rumination should be particularly salient for adolescents who are invested in needing to be perfect.

The overarching goal of the current study was to test the hypothesis that trait perfectionism dimensions (self-oriented and socially prescribed perfectionism) and facets of mistake orientation (i.e., WAM and MR) are associated with higher levels of L2 anxiety and less WTC. Moreover, as more contextually relevant predictors, we also tested the hypothesis that WAM and MR would serve as mediators of trait perfectionism in terms of predicting concurrently measured levels of L2 anxiety and WTC. We also assessed English course scores to explore where these were linked with the other measures in the current study.

It is important to reiterate that while our main focus was on substantive issues, the current work also had an important unique focus on measurement and assessment issues. In addition to the two measures of mistake orientation, we also evaluated the psychometric characteristics of the brief 9-item CAPS when completed by a large sample of adolescents from Iran. It has four items that tap self-oriented perfectionism and five items that tap socially prescribed perfectionism (see Bento et al., 2019). SOP is a trait dimension that taps excessive striving, extreme self-standards, and demanding perfection from the self (see Hewitt & Flett, 1991). SPP is the perception, veridical or not, that others demand perfectionism from the self. Higher levels of SPP are very deleterious and destructive (see Flett et al., 2022). The brief CAPS was found to have sound psychometric characteristics when evaluated in two samples of adolescents from Portugal, but clearly, it needs to be re-evaluated.

Willingness to Communicate

Willingness to communicate (WTC) refers to “… a readiness to enter into discourse at a particular time with a specific person or persons, using L2” (MacIntyre et al., 1998, p, 547). This definition suggests that successful communication in L2 requires a sort of psychological readiness in addition to linguistic and communicative competence (Barabadi, et al., 2021; Chasetareh et al., 2022). Past research (Brauer et al., 2023; Shirvan et al., 2019) has clearly shown that L2 anxiety is the strongest negative predictor of L2 WTC, and one of the best ways to improve WTC is to help L2 learners overcome their anxiety (Dewaele & Dewaele, 2018). In a recent study, Barabadi et al. (2021) found that gelotophobia, or fear of being laughed at as a personality trait, is negatively associated with L2 WTC because such learners perceive their mistakes and their peers and teacher’s reactions to these mistakes in a negative way. So, L2 learners’ level of anxiety is dependent on how they perceive the communicative environment of the classroom (Brauer et al., 2023). As suggested by prior research, L2 learners’ perceptions of laughter, mistakes, and self-competence depend, to a large extent, on their personality traits like perfectionism (Barabadi & Khajavy, 2020), impostor phenomenon (Brauer et al., 2023), and gelotophobia (Barabadi et al., 2021). For example, Barabadi et al. (2022) identified that higher levels of concern over mistakes, as a negative dimension of perfectionism, and also higher levels of anxiety affected students in a way that they were less inclined toward participating in classroom activities and tasks. They explained that when students’ short-term memory is filled up with excessive and repetitive ruminations and worries about minor mistakes, they cannot concentrate on achieving their goals, and hence, they tend to avoid communicative situations. Following this line of research, in this study, we sought to understand how worry about mistakes and mistake rumination as two cognitive aspects of perfectionism are related to L2 WTC. These issues were examined in a large combined sample of students in junior high school and senior high school. Certain statistical analyses reported below took into account whether students were juniors or seniors.

Method

Participants and Procedure

Participants were 952 school students (489 females and 463 males) in Iran recruited from junior high schools (N = 536) and senior high schools (N = 416). Their age range was between 12 and 20 years (M = 15.04, SD = 1.72). Persian was their first language. English is a compulsory course at schools in Iran. Further, their self-reported proficiency in oral English was medium (M = 3.39, SD = 1.2; with 1 representing “no command” of oral English and 6 representing “full command” of oral English). Students’ participation in the study was voluntary.

Having obtained official permission from both the University of Bojnord and also the Educational Organization of North Khorasan, we spoke with the principals of schools in North Khorasan as well as language teachers in these schools about the main purposes of conducting this research. After obtaining their approval, two researchers collected data from the students during regular class hours in the paper-pencil form of the questionnaires. All scales were in the participants’ first language (e.g., Persian) and required about 20 min to be completed. It should be noted that the students were informed of the overall research purpose and they were told that participation was voluntary. Parental consent was also obtained and was required.

Instruments

CAPS (9 Items)

The short 9-item Child–Adolescent Perfectionism Scale (CAPS; Bento et al., 2019) assessed socially prescribed perfectionism (five items) and self-oriented perfectionism (four items). Both subscales have a 5-point Likert-type scale that ranges from 1 (strongly disagree) to 5 (strongly agree). A sample item from socially prescribed perfectionism is “Other people always expect me to be perfect.” A sample item from self-oriented perfectionism is “I try to be perfect in everything I do.”

Worry About Mistakes (5 Items)

The 5-item Worry About Mistakes Scale was developed for the current study. It utilizes a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). One sample item is “Once I start worrying about making mistakes, I cannot stop.” We administered these items to a pilot sample of 98 participants to evaluate the reliability of this measure. Our initial evaluation suggested that this new measure has acceptable internal consistency.

Mistake Rumination Scale

The 7-item Mistake Rumination Scale (MRS) was used to assess the cognitive aspects of perfectionism in terms of overthinking or ruminating about mistakes made in the past. The MRS was developed and validated by Flett et al. (2020). Items are responded according to a 4-point Likert scale ranging from 1 (not at all) to 4 (very much). Psychometric analyses described by Flett et al. (2020) indicate this scale has high internal consistency and validity. This scale requires participants to think about and report on a mistake they have made. One sample item from this scale is “To what extent do you still think about the mistake and wish it had gone better?”

Willingness to Communicate

L2 learners’ WTC was assessed using seven items from Khajavy et al. (2018). Originally developed by Weaver (2005), this questionnaire has a 5-point Likert-type scale that ranges from 1 (strongly disagree) to 5 (strongly agree). The items have high face validity. For instance, one sample item is “I am willing to speak English about a topic which is written on a board or a piece of paper.” Prior research that utilized the Persian translation of this questionnaire yielded ample evidence of its validity and reliability (Barabadi et al., 2021).

L2 Anxiety

Levels of L2 anxiety were assessed with eight items developed by Dewaele & MacIntyre (2016). The items are rated on a Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). This scale measures the extent to which language learners’ have L2 anxiety while they use English (e.g., “I get nervous and confused when I am speaking in my FL class”).

Results

We used SPSS and AMOS for all data analyses. Initially, data screening was conducted to ensure that data were appropriate for subsequent analyses. To this aim, normality, outliers, and missing data were checked. To examine normality, we used skewness and kurtosis values. For sample sizes larger than 300 participants, kurtosis and skewness absolute values should be less than 7 and 2, respectively (Kim, 2013). The skewness and kurtosis values of all the items lay within the acceptable range. It is worth noting that SPSS reports excess kurtosis, that is, absolute kurtosis minus three. As such, Table 1 only reports excess kurtosis which is less than four. Z-standardized values were used to inspect univariate outliers that range between −3 and +3. As a result, one item of SPP had 34 cases of outliers, and three items of SOP had 50 cases of outliers. We removed all these cases from the dataset for subsequent analyses. Then, the multivariate outlier was inspected using Mahalanobis D². As a result, we had to remove 22 cases from our data set since they had Mahalanobis D² with p < .001. To handle missing data, the expectation maximization (EM) approach implemented in SPSS was used. The rate of missing data was 8.2% for perfectionism items, 7.3% for WTC, 8.4 for L2 achievement, 9.1% for L2 anxiety, 5.7% for WAM, 6.2% for MR, and 11.4% for L2 achievement. EM is used for handling missing data when the data on a specific variable is missing at random. We used a series of t-tests and logistic regression to make sure that the missing data on our variables were random.

First, confirmatory factor analyses (CFAs) examined the construct validity of the scales. We conducted four separate CFAs for perfectionism, WTC, L2 anxiety, WAM, and MR. Structural equation modeling (SEM) was then used to examine how perfectionism predicts L2 achievement and WTC, followed by meditational analyses.

Model fit was tested based on different goodness of fit indices. In the present study, the Tucker–Lewis index (TLI), the comparative fit index (CFI), the root mean square error of approximation (RMSEA), and the standardized root mean square residual (SRMR) were used. For CFI and TLI, values above .90 and .95 show adequate and good fit, respectively (Hu & Bentler, 1999; Marsh et al., 2004). For RMSEA and SRMR, values less than .08 and .05 show adequate and good fit, respectively (Hu & Bentler, 1999; Marsh et al., 2004). These values should not be considered as golden rules given the complexity of data and models. To report effect sizes, we used Cohen’s f² (Cohen, 1992). The interpretation of Cohen’s f² is as follows: f² = 0.02 small effect; f² = 0.15 medium effect; and f² = 0.35 large effect (Cohen, 1992).

Measurement Models

First, we conducted CFA results for perfectionism, WAM, MR, L2 anxiety, and WTC (see Table 3). First, a two-factor model of perfectionism including SOP and SPP was tested. As mentioned before, we had to remove one item from the SPP subscale and one item from the SOP subscale because the omission of these items helped boost the reliability of the scales. The CFA model for these perfectionism dimensions showed a perfect fit to the data (see Figure 1). The factor loadings, as can be seen in Figure 1, range from .62 to .77.

Figure 1.

Measurement model for perfectionism.

Next, we tested a 2-factor model of WAM and MR as two state aspects of perfectionism that act as mediators between trait perfectionism (SOP and SPP) on the one hand and our outcome variables, namely, L2 anxiety, WTC, and L2 achievement, on the other hand. The CFA model for these two constructs showed a perfect fit to the data while retaining the original 5 items of WAM and 7 items of MR, with factor loadings ranging from .54 to .74 (see Figure 2).

Figure 2.

Measurement model for state aspects of perfectionism including WAM and MR.

Finally, we examined two separate CFA models for L2 anxiety and WTC. Having removed one item from the L2 anxiety scale because of lowering reliability, we tested a model based on 7 items. The model showed a very good fit to the data; item loadings ranged from .52 to .84 (see Figure 3). The initial CFA model for WTC did not show a very good fit to the data, but after connecting three error terms, the model showed a very good fit to the data. The loadings of the items ranged from .54 to .75 (see Figure 4).

Figure 3.

Measurement model for L2 anxiety.

Figure 4.

Measurement model for L2 WTC.

Table 1 presents the descriptive statistics for the measures for the junior high school students versus the senior high school students. We used Cronbach’s α and omega (ω) to evaluate internal consistency and estimate composite reliability.

Table 1.

Descriptive Statistics for All Variables.

	Α (ω)	Possible range	M	SD	Skewness	Kurtosis
WTC J	.85	1–5	26.97	5.69	.70	.24
WTC S	.84	1–5	25.82	5.96	.35	.37
SOP J	.68	1–5	17.51	2.29	.73	−.38
SOP S	.69	1–5	17.49	2.09	−.53	−.62
SPP J	.80	1–5	19.54	3.79	−.58	−.02
SPP S	.81	1–5	18.82	4.01	−.48	−.17
WAM J	.83	1–5	16.11	5.18	−.20	−.70
WAM S	.82	1–5	15.71	4.87	−.02	−.44
MR J	.80	1–5	19.06	4.91	−.32	−.43
MR S	.80	1–5	19.04	4.95	−.27	−.51
L2 Anxiety J	.85	1–5	24.65	5.56	.07	−.21
L2 Anxiety S	.83	1–5	24.45	5.23	.08	−.45
Achievement J		1–20	17.54	3.12	1.65	5.58
Achievement S		1–20	17.19	2.78	1.30	4.03

Supplementary analyses evaluated possible group differences for junior high school students (n = 536) versus senior high school students (n = 416). All measures were considered. The only significant group difference was on levels of socially prescribed perfectionism. Higher scores were found among junior high school students.

Correlation coefficients among the variables are presented in Table 2. SOP was not correlated significantly with WAM or MR, nor was it associated with L2 anxiety. However, it was associated positively with willingness to communicate and performance. SPP had positive and significant correlations with WAM and MR with the strongest association being with mistake rumination. There was also a weak negative association with L2 anxiety.

Table 2.

Latent Correlation Matrix Among All Variables.

	1	2	3	4	5	6	7
1. SOP	1.00
2. SPP	.27***	1.00
3. WAM	−.05	.18***	1.00
4. MR	.01	.34***	.77***	1.00
5. Anxiety	.07	−.12***	−.41***	−.50***	1.00
6. WTC	.21***	−.06	−.43***	−.51***	.59***	1.00
7. L2 Achiev	.21***	−.01	−.37***	−.45***	.60***	.81***	1.00

Note. *p < .05; ***p < .001.

Further, both WAM and MR had significant correlations with L2 anxiety. Both variables also had significant negative associations with willingness to communicate and L2 achievement.

Next, we conducted a hierarchical multiple regression to assess the effects of SOP and SPP on L2 anxiety after controlling for the influence of WAM and MR. WAM and MR were entered at Step 1, explaining 35% of the variance in L2 anxiety, F (2, 949) = 252, p < .001. The total variance accounted for did not change significantly after entering trait perfectionism (R squared change = .002, F change (2, 947) = 1.17, p = .31). In the final model, only WAM and MR were statistically significant, with the WAM having a higher beta value (beta = .58, p < .000) than MR (beta = .08, p < .02).

A similar analysis was conducted with WTC as the dependent variable. WAM and MR at Step 1 explained virtually no variance in L2 WTC. When trait perfectionism was entered in Step 2, the total variance accounted for by the model was 8% (F (4, 947) = 20.14, p < .001). SOP and SPP explained 7.5 % of the variance in L2 WTC, R squared change = .075, F change (2, 947) = 38.33, p < .001. In the final model, only WAM and SOP were statistically significant. WAM had a negative beta value (beta = −.12, p < .003 and SOP had a positive beta value (beta = .27, p < .001).

Testing Structural Models

We used one multi-group structural model to assess the mediating effects of WAM and MR as two state characteristics of perfectionistic students, as well as L2 anxiety on the relationship between two dimensions of perfectionism (e.g., SOP and SPP) and WTC. This model demonstrated a good fit to the data (see Table 3).

Table 3.

Goodness-of-Fit Indices for CFA and SEM Models.

Scales	ꭓ²	df	ꭓ²/df	GFI	CFI	TLI	RMSEA
Perfectionism	48.67	13	3.74	.99	.98	.97	.054
WAM and MR	143.56	53	2.7	.98	.98	.97	.042
Anxiety	90.44	14	6.46	.97	.97	.95	.075
WTC	78.28	11	7.11	.98	.97	.95	.08
SEM Model WTC	2083	967	2.15	.90	.91	.90	.035
SEM Model L2 Achiev	1225	315	3.89	.91	.90	.90	.055

As observed in Figure 5, for junior high school students (aged 14–16 years), SOP did not predict WAM and MR in a statistically significant way while it negatively predicted L2 anxiety (β = −.15, p < .012, R² = .02, f ² = 0.02, small effect). However, SPP could positively predict both WAM (β = .27, p < .000, R² = .07, f ² = 0.08, small effect) and MR (β = .27, p < .000, R² = .07, f ² = 0.08, small effect) while it did not predict L2 anxiety in a statistically significant way. The two perfectionism dimensions accounted for 9% of the variance in WAM and 8% of the variance in MR. Further, both WAM (β = .70, p < .000, R² = .49, f ² = 0.96, large effect) and MR (β = .15, p < .001, R² = .02, f ² = 0.02, small effect) were positive predictors of L2 anxiety. The two perfectionism dimensions and WAM and MR could account for 48% of the variance in L2 anxiety (R² = .48, f ² = 0.92, large effect). Finally, L2 anxiety predicted WTC negatively (β = −.45, p < .000, R² = .20, f ² = 0.25, medium effect).

Figure 5.

Structural model perfectionism, WAM, MR, anxiety, and WTC for junior high school students.

For the participants in the high school subsample, a similar pattern of results was observed: SOP did not predict WAM and MR in a statistically significant way while it negatively predicted L2 anxiety (β = −.14, p < .012, R² = .02, f ² = 0.02, small effect). Similar to the findings of the junior high school sample of participants, SPP could positively predict both WAM (β = .31, p < .000, R² = .07, f ² = 0.08, small effect) and MR (β = .32, p < .000, R² = .07, f ² = 0.08, small effect) while it did not predict L2 anxiety in a statistically significant way. The two dimensions of perfectionism accounted for 8% of the variance in WAM and an 8% of the variance in MR. Further, both WAM (β = .33, p < .000, R² = .49, f ² = 0.96, large effect) and MR (β = .13, p < .001, R² = .02, f ² = 0.02, small effect) could positively predict L2 anxiety. The two dimensions of perfectionism as well as WAM and MR could account for 44% of the variance in L2 anxiety (R² = .44, f ² = 0.92, large effect). In this model, L2 anxiety predicted WTC negatively (β = −.55, p < .000, R² = .20, f ² = 0.25, medium effect). Finally, we examined whether SOP and SPP could indirectly predict WTC through WAM, MR, and L2 anxiety for the whole sample. Mediation analyses indicated that while SOP could positively predict WTC via L2 anxiety (β = .05, p = .001), SPP could negatively predict WTC via WAM, MR, and then L2 anxiety (β = −.05, p = .006). In addition, in this subsample, WAM (β = −.18, p = .009) and MR (β = −.04, p = .01) negatively predicted WTC via L2 anxiety (Figure 6).

Figure 6.

Structural model perfectionism, WAM, MR, anxiety, and WTC for senior high school students.

The second structural model was run to assess the mediating effects of WAM and MR as two state characteristics of perfectionistic students, as well as L2 anxiety on the relationship between two dimensions of perfectionism (e.g., SOP and SPP) and L2 achievement. This model also demonstrated a good fit to the data (see Table 3). As observed in Figure 7, SOP did not predict WAM and MR in a statistically significant way while it negatively predicted L2 anxiety (β = −.12, p < .012, R² = .02, f ² = 0.02, small effect). However, SPP could positively predict both WAM (β = .30, p < .000, R² = .07, f ² = 0.08, small effect) and MR (β = .31, p < .000, R² = .07, f ² = 0.08, small effect) while it did not predict L2 anxiety in a statistically significant way. The two perfectionism dimensions could account for a 7% variance of WAM and MR. Further, both WAM (β = .64, p < .000, R² = .49, f ² = 0.96, large effect) and MR (β = .15, p < .001, R² = .02, f ² = 0.02, small effect) could positively predict L2 anxiety. The two perfectionism dimensions and WAM and MR accounted for 43% of the variance in L2 anxiety (R² = .43, f ² = 0.92, large effect). Finally, L2 anxiety predicted L2 achievement negatively (β = −.32, p < .000, R² = .20, f ² = 0.25, medium effect). Mediation analyses indicated that while SOP could positively predict L2 achievement via L2 anxiety (β = .06, p = .01), SPP could negatively predict WTC via WAM, MR, and then L2 anxiety (β = −.05, p = .008). Further, WAM (β = −.20, p = .007) and MR (β = −.04, p = .03) could negatively predict L2 achievement via L2 anxiety.

Figure 7.

Structural model perfectionism: WAM, MR, anxiety, and achievement for junior high school students.

Discussion

We begin by discussing the main psychometric findings and then the substantive findings involving key associations detected in the current study. At a psychometric level, the first focus was on the psychometric characteristics of the abbreviated CAPS (see Bento et al., 2019). Our results supported the use of this measure in Iranian adolescents but also demonstrated the need to further evaluate existing measures rather than presume they are generalizable. Our results indicated that the measure is sound, but one item can be dropped for each subscale, resulting in a 4-item measure of socially prescribed perfectionism with internal reliability of .80 and a 3-item measure of self-oriented perfectionism with lower internal reliability (i.e., an alpha coefficient of .68). These subscales were correlated at the construct level (r of .52) but not to the point that they are redundant with each other. Clear evidence was obtained in this sample to indicate that perfectionism is indeed multidimensional. Future research with participants from other countries is needed to determine whether the 9-item CAPS should be used or if it can be slightly modified, as was the case in the current study.

Another focus was the Mistake Rumination Scale, which was used here for the first time with adolescents. The MRS was confirmed to have a unidimensional structure with all 7 items being retained. A high level of internal consistency was found (alpha = .80), and as discussed below, the measure yielded findings that reflect construct validity.

The new measure of WAM proved to be unidimensional as intended and these highly face-valid items were also found to have high internal consistency (alpha = .82). It should not be too surprising to learn that it is possible to create a brief measure of WAM. Many students suffer from worry, and it should be the case that perfectionistic students in particular should be prone to WAM. The tendency to worry about mistakes likely applies broadly to many students who are learning a new language, and this should especially be the case when a mistake becomes public and known to peers. However, we view this new measure as a reflection of the general tendency to worry about making a mistake across many contexts. Of course, this assumption needs to be evaluated in future research with children, adolescents, and emerging adults in samples from a range of countries and backgrounds. We regard worry over mistakes as part of a general orientation reflecting a hypersensitivity to anticipate mistakes and react strongly when a mistake actually has occurred.

Regarding the correlational results, our analyses indicated a strong association between MR and WAM, as would be expected given that they are both measures of a deleterious orientation toward mistakes. This association signifies that not only do young people high in mistake rumination have a cognitive focus on past mistakes, but they also tend to have a cognitive preoccupation that is anticipatory and centered on the perceived possibility or likelihood of future mistakes. We suggest that if it had actually been assessed, many adolescents in our sample would be distinguished by a dread of making future mistakes that they perceive as certain or almost certain to eventually take place.

At the substantive level, this research examined the extent to which measures of trait perfectionism and mistake orientation are related to WTC and L2 anxiety. These communication measures had little overlap in the current study and both were modestly but significantly correlated in the expected directions with actual grades in English. Overall, the pattern of correlations attested to the need in this context to supplement trait perfectionism measures with more specific scales tapping facets of mistake orientation. Our analyses showed that SPP had a small but significant association with MR and WAM. SPP also had a small but significant link with greater L2 anxiety, while SOP was linked with greater rather than less WTC. This finding fits the general pattern of needing to examine in specific samples whether the drive and striving inherent in SOP may prove beneficial or problematic. More generally, not only this pattern of results has conceptual implications for model building, but also has important assessment implications when seeking to account for and understand the vulnerabilities of individual students.

The most striking findings in this study were the associations found between L2 anxiety and both MR and WAM with the link with worry about mistakes being especially robust. The clear implication here is that vulnerable students need to learn cognitive and emotional regulation techniques and focus on their ability to manage mistakes. Attempts at intervention should be more effective to the extent that they help students cognitively reappraise the meaning and significance of mistakes in ways that result in less threat and less dread. On a related note, there are bound to be some benefits associated with promoting a growth orientation in terms of mindset and a limited fixed mindset including the notion that mistakes reflect stable and unchanging deficiencies in the self (for a discussion, see Flett & Hewitt, 2022).

Structural model analyses yielded findings in keeping with our emphasis on the need to supplement distal measures of trait perfectionism with measures that seemingly relate more closely to the daily learning challenges adolescent students face. Separate analyses were conducted for junior high school and high school students. As the results suggest, WAM is a potentially important mediator in terms of the link between trait perfectionism and L2 anxiety. Further, while MR also has a clear role to play as a mediator, it should be more relevant once mistakes have actually been made; thus, it is logical that WAM would be more pertinent based on the current research design and focus when seeking to account for individual differences in L2 anxiety.

Conclusion

The current study sought to examine the associations among trait perfectionism dimensions, MR, WAM, L2 anxiety, and WTC. An equally important goal of this study was to examine the reliability and validity of the MR Scale and the WAM Scale in the context of L2 learning. The results of the study indicated that SPP, as the negative dimension of perfectionism, is positively correlated with MR and WAM, which is in line with the general notion that negative dimensions of perfectionism are associated with negative outcomes and processes. Moreover, MR and WAM are highly relevant to this context and seemingly mediate the relationship between SPP on the one hand and L2 anxiety and WTC on the other hand. This suggests that the imposed perfectionistic thoughts and behaviors from close others may involve language learners cognitively; hence, they focus more on their mistakes and their willingness to communicate decreases.

At a broader level, our results illustrate the need for a broader emphasis that includes the mistake orientation of students prone to perfectionism when they are being assessed in performance contexts. There has been a strong emphasis thus far on the negative emotional reactions to imperfections experienced by young people prone to perfectionism (see Stoeber & Rambow, 2007). The current results emphasize the need for an extended focus that includes the worries prior to performance, especially in terms of the cognitive experience of WAM. In many instances, this worry can prove overwhelming and can result in a self-fulfilling prophecy reflected by actually making mistakes and finding it quite difficult to concentrate. These worries are likely fueled by the fear of failure central to the approach-avoidance conflict experienced by adolescent perfectionists (see Flett & Hewitt, 2022). There is a clear need to assist students in developing social-emotional learning and emotion regulation skills that will limit the experience and impact of these worries and instead foster a degree of confidence that will boost their ability to engage in learning a language as well as other new things.

Limitations of the Current Study

The current study has some clear limitations. One key limitation was the cross-sectional design of this study, which limited the interferences of the causal relationship between the variables. As the researchers did not access any information on the nature of language learners’ past mistakes and also the real reason for their lower WTC, to this end, a further qualitative study can discuss whether those mistakes they ruminated about were only related to language learning or there were some other extraneous causes involved.

In addition, it cannot be presumed that the current results are generalizable. It is particularly important that the measures related to perfectionism and mistakes that were used in this study need to be evaluated in students of various ages who are from other countries.

In summary, the results of the current study indicated that to the extent that trait perfectionism is problematic in language learning, socially prescribed perfectionism is one trait dimension that is antithetical to language learning. However, more problematic are elevated levels of mistake rumination and worry about mistakes. Meaningful individual differences clearly exist in terms of worry about mistakes and mistake rumination in adolescents from Iran. It can be extrapolated from our findings that students learning new languages should profit from learning more adaptive ways of reacting to mistakes. It should prove illuminating in future research to examine the learning contexts and classroom cultures that promote these more adaptive ways of reacting so that making mistakes and learning from them can become and be regarded as a fundamental aspect of the learning process.

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.

Funding

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

ORCID iDs

Elyas Barabadi

Gordon L. Flett

Paul L. Hewitt

Seyyed Ayatollah Razmjoo

Mohsen Rahmani Tabar

Fatemeh Chasetareh

Appendix

References

Barabadi

Brauer

Proyer

R. T.

Tabar

M. R.

(2021). Examining the role of gelotophobia for willingness to communicate and second language achievement using self- and teacher ratings. Current Psychology, 42(6), 5095–5109. https://doi.org/10.1007/s12144-021-01846-w

Barabadi

Khajavy

G. H.

(2020). Perfectionism and foreign language achievement: The mediating role of emotions and achievement goals. Studies In Educational Evaluation, 65, 100874. https://doi.org/10.1016/j.stueduc.2020.100874

Barabadi

Khajavy

G. H.

Booth

J. R.

Rahmani Tabar

Vahdani Asadi

M. R.

(2022). The links between perfectionistic cognitions, L2 achievement, and willingness to communicate: Examining L2 anxiety as a mediator. Current Psychology, 42(35), 30878–30890. https://doi.org/10.1007/s12144-022-04114-7

Bento

C. F.

Pereira

Azevedo

Saraiva

Flett

G. L.

Hewitt

P. L.

Macedo

(2019). Development and validation of a short form of the child–adolescent perfectionism scale. Journal of Psychoeducational Assessment, 38(1), 26–36. https://doi.org/10.1177/0734282919879834

Besser

Flett

G. L.

Hewitt

P. L.

(2004). Perfectionism, cognition, and affect in response to performance failure vs. success. Journal of Rational-Emotive and Cognitive-Behavior Therapy, 22(4), 297–324. https://doi.org/10.1023/b:jore.0000047313.35872.5c

Brauer

Barabadi

Aghaee

Alrabai

Elahi Shirvan

Sendatzki

Vierow

L. M.

(2023). Impostor phenomenon and L2 willingness to communicate: Testing communication anxiety and perceived L2 competence as mediators. Frontiers in Psychology, 13, 1060091. https://doi.org/10.3389/fpsyg.2022.1060091

Chasetareh

Barabadi

Khajavy

G. H.

Flett

G. L.

(2022). Perfectionism and L2 achievement: The mediating roles of motivation and self-regulated learning among Iranian high school students. Journal of Psychoeducational Assessment, 41(1), 3–21. https://doi.org/10.1177/07342829221096916

Cohen

(1992). A power primer. Psychological Bulletin, 112(1), 155–159. https://doi.org/10.1037/0033-2909.112.1.155

Dewaele

J. M.

(2012). Personality in second language acquisition. In Chapelle

C. A.

(Ed.), The encyclopedia of applied linguistics. Malden: Wiley-Blackwell. https://doi.org/10.1002/9781405198431.wbeal0904

10.

Dewaele

(2018). Learner-internal and learner-external predictors of willingness to communicate in the FL classroom. Journal of the European Second Language Association, 2(1), 24. https://doi.org/10.22599/jesla.37

11.

Dewaele

J. M.

MacIntyre

P. D.

(2016). Foreign Language Enjoyment and Foreign Language Classroom Anxiety: The Right and Left Feet of the Language Learner. In MacIntyre

P. D.

Gregersen

Mercer

(Eds.), Positive psychology in SLA (pp. 215–236). Bristol/Blue Ridge Summit: Multilingual Matters. https://doi.org/10.21832/9781783095360-010

12.

Elahi Shirvan

Khajavy

G. H.

MacIntyre

P. D.

Taherian

(2019). A meta-analysis of L2 willingness to communicate and its three high-evidence correlates. Journal of Psycholinguistic Research, 48(6), 1241–1267. https://doi.org/10.1007/s10936-019-09656-9

13.

Flett

G. L.

Hewitt

P. L.

(2022). Perfectionism in childhood and adolescence: A developmental approach. American Psychological Association. https://doi.org/10.1037/0000289-000

14.

Flett

G. L.

Hewitt

P. L.

Besser

Vaillancourt

Boucher

Munro

Davidson

L. A.

Gale

(2016a). The child–adolescent perfectionism scale: Development, psychometric properties, and associations with stress, distress, and psychiatric symptoms. Journal of Psychoeducational Assessment, 34(7), 634–652. https://doi.org/10.1177/0734282916651381

15.

Flett

G. L.

Hewitt

P. L.

Nepon

Sherry

Smith

M. M.

(2022). The destructiveness and public health significance of socially prescribed perfectionism: A review, analysis, and conceptual extension. Clinical Psychology Review, 93(102130), 1–21. https://doi.org/10.1016/j.cpr.2022.102130

16.

Flett

G. L.

Hewitt

P. L.

Flett

K. D.

(2016b). Perfectionism in language learners: Review, conceptualization, and recommendations for teachers and school psychologists. Canadian Journal of School Psychology, 31(2), 75–101. https://doi.org/10.1177/0829573516638462

17.

Flett

G. L.

Nepon

Hewitt

P. L.

Zaki-Azat

Rose

Swiderski

(2020). The mistake rumination scale: Development, validation, and utility of a measure of cognitive perfectionism. Journal of Psychoeducational Assessment, 38(1), 84–98. https://doi.org/10.1177/0734282919879538

18.

Frost

R. O.

Marten

P. A.

Lahart

Rosenblate

(1990). The dimensions of perfectionism. Cognitive Therapy and Research, 14(5), 449–468. https://doi.org/10.1007/bf01172967

19.

Frost

R. O.

Trepanier

K. L.

Brown

E. J.

Heimberg

R. G.

Juster

H. R.

Makris

G. S.

Leung

A. W.

(1997). Self-monitoring of mistakes among subjects high and low in perfectionistic concern over mistakes. Cognitive Therapy and Research, 21(2), 209–222. https://doi.org/10.1023/a:1021884713550

20.

Hewitt

P. L.

Flett

G. L.

(1991). Perfectionism in the self and social contexts: Conceptualization, assessment, and association with psychopathology. Journal of Personality and Social Psychology, 60(3), 456–470. https://doi.org/10.1037/0022-3514.60.3.456

21.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. https://doi.org/10.1080/10705519909540118

22.

Joubert

A. E.

Moulds

M. L.

Werner-Seidler

Sharrock

M. J.

Popovic

Newby

J. M.

(2022). Understanding the experience of rumination and worry: A descriptive qualitative survey study. British Journal of Clinical Psychology, 61(4), 929–946. https://doi.org/10.1111/bjc.12367

23.

Khajavy

G. H.

MacIntyre

P. D.

Barabadi

(2018). Role of the emotions and classroom environment in willingness to communicat. Studies in Second Language Acquisition, 40(3), 605–624. https://doi.org/10.1017/s0272263117000304

24.

Kim

(2013). Statistical notes for clinical researchers: Assessing normal distribution (2) using skewness and kurtosis. Restorative dentistry & endodontics, 38(1), 52–54. https://doi.org/10.5395/rde.2013.38.1.52

25.

MacIntyre

P. D.

Clément

Dörnyei

Noels

K. A.

(1998). Conceptualizing willingness to communicate in a L2: A situational model of L2 confidence and affiliation. The Modern Language Journal, 82(4), 545–562. https://doi.org/10.1111/j.1540-4781.1998.tb05543.x

26.

Marsh

H. W.

Hau

Wen

(2004). In search of golden rules: Comment on hypothesis-testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler’s (1999) findings. Structural Equation Modeling: A Multidisciplinary Journal, 11(3), 320–341. https://doi.org/10.1207/s15328007sem1103_2

27.

Sarason

I. G.

(1984). Stress, anxiety, and cognitive interference: Reactions to tests. Journal of Personality and Social Psychology, 46(4), 929–938. https://doi.org/10.1037/0022-3514.46.4.929

28.

Stoeber

Rambow

(2007). Perfectionism in adolescent school students: Relations with motivation, achievement, and well-being. Personality and Individual Differences, 42(7), 1379–1389. https://doi.org/10.1016/j.paid.2006.10.015

29.

Stornelli

Flett

G. L.

Hewitt

P. L.

(2009). Perfectionism, achievement, and affect in children: A comparison of students from gifted, arts, and regular programs. Canadian Journal of School Psychology, 24(4), 267–283. https://doi.org/10.1177/0829573509342392

30.

Weaver

(2005). Using the Rasch model to develop a measure of second language learners’ willingness to communicate within a language classroom. Journal of Applied Measurement, 6(4), 396–415. https://pubmed.ncbi.nlm.nih.gov/16192663/