Introduction

Narrow definitions of giftedness (e.g., Sternberg 2004; Sternberg and Davidson 1986) recognize at least two distinct elements: (1) a natural potential, identified before its development during childhood, and (2) outstanding performance in tasks requiring cognitive abilities (Gagné 2003). However, broader visions of giftedness (e.g., Thompson and Oehlert 2010) consider gifted people beyond just their genetic influences. According to Heller and Perleth (2008), giftedness should not be measured only in relation to cognitive factors: there are creative, social, emotional abilities, as well as other elements that are implicit in the giftedness. The potential uniqueness of gifted children’s abilities could lead them to be socially vulnerable; they may experience fear, existential anxiety, and personal suffering, so that they may not be able to take advantage of their superior capacities (Seely 2004). Therefore, this vulnerability should be attended to the following: might mindfulness be a solution for decreasing such symptoms?

Gifted children think and behave differently from their peers, with these differences being observable in the early stages of the children’s development (Gross 2003). Research has also found that their development is asynchronous (e.g., Silverman 2005); for example, their cognitive development is faster (Walsh et al. 2010) and their psychomotor development is slower (Alsop 2003) than that of their peers. In other areas too, such as social and emotional development, the relative progress of gifted children has been found to be asynchronous (Daniels and Piechowski 2009), which is thought, in turn, to lead to them experiencing over-excitability (Tieso 2007). Over-excitability in gifted children is considered to be a mode of understanding and responding (Piechowski and Chucker 2011) that is different to the approach typically seen in non-gifted children. According to Piechowksi (1979), gifted children express emotions through an excess of energy or restlessness. These children require sensory stimulation (e.g., Daniels and Piechowski 2009), and their intellectual mode is characterized by analysis, logic, and a search for truth entailing continuous and intense intellectual stimulation (Piechowski and Chucker 2011). Their creative mode will often incorporate vivid dreams and strong visualizations of experience (Daniels and Piechowski 2009), while their emotional and spiritual modes are thought to be connected, with the children expressing feelings from empathy to loneliness, as well as experiencing strong cognitions regarding their spiritual concerns (Piechowski 2003). Gifted children’s over-excitability is also why a mindfulness-based intervention (MBI) might be of particular use to them.

Thus, gifted children exhibit several cognitive and affective differences in comparison to their peers that make them likely candidates to benefit from MBI. They are more perfectionistic (Guignard et al. 2012), they avoid taking risks (Silverman 1994), they experience more anxiety and are more inclined to worry (Schaerf 2016), and they are more socially withdrawn (Peterson et al. 2009) than their school-aged peers. Endeavoring to explain these differences, Webb et al. (2007) pointed out that gifted children conceal their abilities and interests in an attempt to fit in. Another reason why MBI could be especially useful for gifted children is connected to how these children perceive external pressures such as excessive academic demands, and the expectations of adults about their abilities (Zahn-Waxler et al. 2000), even though non-gifted individuals may have similar experiences (Cross and Cross 2015). Gifted children have been found to experience stress and depression symptoms (Fakolade and Archibong 2013) and may present disrupting behaviors, defiance, or seek attention from others (Neihart et al. 2002).

Most psychological interventions advanced in relation to gifted children have focused on attending to their additional stressor experiences, as discussed above (Haberlin 2015). Despite recent reports connecting giftedness and mindfulness (e.g., Sharp et al. 2017), such interventions generally have not included an existential element (Daniels and Piechowski 2009). From an early age, many gifted children show spiritual sensitivity (Thera 2014): they comprehend the universality of concepts such as forgiveness (Lovecky 1998), they develop a spiritual consciousness (Mayer 2000), and they have a systematic philosophy of life and death, positing momentous questions (Sisk 2008). Such spiritual requirements and a sense of self-consciousness (Mendaglio 2003) need to be supported and attended (Piechowski 2003). Machů and Morysová (2016) found that gifted children need to assume an influencing role, rather than be a passive observer of a given situation, especially with respect to the big existential questions. Moreover, their existential concerns lead them to experience fear about global issues such as war, violence, illness, and especially death (Derevensky and Coleman 1989; Lamont 2012).

Mindfulness has been defined as the process of increasing the quality of attention to the experience of the present moment (Kabat-Zinn 1990). The ability to evoke mental presence can be developed using various Buddhist meditation techniques (Nhat Hanh 1976) where mindfulness is conceived as a path leading to the cessation of personal suffering (Thera 2014). If we accept that gifted children experience both emotional and existential suffering, it might also be inferred that mindfulness would be a helpful tool for them and may reduce the adverse effects or increase the benefits of giftedness at school.

Gifted children can be taught skills to help them relax or to reduce stress so that they would be able to participate in school more successfully (Napoli, Krech, and Holley 2005). Several mindfulness-at-school studies have revealed the benefits, such as better academic achievement (Mrazek, Franklin, Phillips, Baird, and Schooler 2013), improved resilience (Bethell, Gombojav, Solloway, and Wissow 2016), a decrease in both stress and negative emotions (Zenner, Herrnleben-Kurz, and Walach 2014), less anxiety (Kallapiran, Koo, Kirubakaran, and Hancock 2015), a reduction in the intensity of depressive symptoms (Kemeny et al. 2012), and better social skills (Langer, Ulloa, Gangas, Rojas, and Krause 2015). Furthermore, mindfulness interventions have been shown to reverse brain patterns that were activated during biological stress (Young. 2011), help children become aware of unnecessary worries (Siegel 2010), provide them with the ability to identify erroneous thoughts and improve their emotional labeling abilities (Sharp, Niemiec, and Lawrence 2017), and promote resilience in facing adverse situations (Burton et al. 2010). Stress during childhood is a precursor of stress in adulthood because people retain learned stress patterns (Jensen et al. 2015).

According to Shonin and Van Gordon (2015), second-generation MBIs (hereafter, SG-MBIs) represent a novel approach to the study of, research into, and practice of mindfulness, as well as an opportunity to develop effective interventions that overcome the limitations of first-generation MBIs. Their study further asserted that (a) mindfulness should be taught in conjunction with other meditative practices and principles (i.e., impermanence, emptiness, loving kindness, and compassion), (b) ethics is a key component of an MBI, and (c) an MBI should be managed by instructors experienced in mindfulness (Van Gordon, Shonin, and Grifiths 2015).

Following several of Shonin and Van Gordon’s (2015) criteria, this pilot study aims to assess the impact of MBI for gifted children in decreasing symptoms of anxiety and depression, increasing emotional capacity to face experiential and existential demands, and improving social functioning. This paper presents the pretest, posttest, and 12-month follow-up findings with regard to the implementation of our “Atención plena para altas capacidades” (APAC) program in a sample of gifted children.

Method

Participants

The study’s participants (n = 22) were gifted children between the ages of 8 and 14 years (M = 11.36, SD = 1.89; 72.7% were males). All of the children had a documented diagnosis of giftedness from the Andalusian Autonomic Direction of School Education, and all belonged to an association of parents with gifted children in Campo de Gibraltar, southern Spain (ACES). All participants completed at least six of the seven sessions described below. Three boys and two girls missed one session. All of the children completed the pre- and the posttreatment measures. However, the 1-year follow-up measurement was only completed by 14 of the children.

Procedure

All of the children were recruited from ACES. Following an introductory session with gifted children’s parents, a total of 22 families agreed to participate in our study. The children’s parents received a schedule listing the APAC program’s sessions (see below) and the sessions’ contents, as well as an agreement detailing the conditions of their children’s participation in the study. All data pertaining to the participants were treated in accordance with this agreement and with full consent obtained from the children’s parents. All parents received an individualized report of their child for every assessment completed in the study. Seven weeks passed between the first (pretest) and the second phase (posttest); a year after that, 14 children (63.63% of the initial sample) completed the 12-month follow-up measures. A pretest, posttest, and 12-month follow-up study design was used to assess the APAC program’s effectiveness. Data were collected from participants at three time periods: (1) prior to starting the program, (2) at the program’s end, and (3) at a 1-year follow-up. The measures used to analyze the data gathered at these time periods were as follows.

The APAC Program

The APAC program consisted of eight 90-min sessions (one for the gifted children’s parents, seven for the gifted children) and followed the structure proposed for SG-MBIs Van Gordon et al. (2016). Table 1 presents a summary of the sessions. Further details are available on request.

Table 1 Principles and practices of the APAC Program (Mindfulness for Giftedness)
Full size table

Measures

CAMM

The Child and Adolescent Mindfulness Measure (CAMM) (Greco, Baer, and Smith 2011) is an awareness measure that consists of ten items designed to assess acceptance and mindfulness using a 5-point Likert scale (ranging from 0, never, to 4, always). Greco et al. carried out a validation study of the CAMM and found it to have a structure with a two-factor solution (thought suppression and psychological inflexibility). Bergomi, Tschacher, and Kupper (2013) described the CAMM’s items as reflecting (a) attendance to internal phenomena, (b) level of awareness of ongoing activities, and (c) judgmental thoughts and feelings regarding a non-accepting stance. Turanzas (2013) led a Spanish validation study, comprising a sample of 379 children (ages ranging from 12 to 15 years old) and found that the CAMM’s fifth and tenth items revealed poor factor loading and reliability. Turanzas’s eight-item Spanish version demonstrated good reliability (.78) and was positively and significantly related to other mindfulness measures. For the present study, we used this eight-item CAMM and used the average score (ranging from 0 to 4).

AFQ-Y

The Avoidance and Fusion Questionnaire for Youth (AFQ-Y8) is a shortened version of the full, 17-item AFQ-Y child-report measure developed by Greco, Lambert, and Baer (2008). It uses a 5-point Likert scale (0 = not true at all; 4 = very true) for rating and measuring young people’s psychological inflexibility, as characterized by cognitive fusion and experiential avoidance (Renshaw 2016). Items are based on the ACT model for human suffering (Coyne et al. 2011). A cutoff score of 15 (1.88 for average scores of an AFQ-Y8) is used for screening purposes to identify youths at risk for clinical-level depression and anxiety (Renshaw 2016). Research has shown that responses to the AFQ-Y8 have strong internal reliability as well as discriminant and convergent validity with several indicators of young people’s mental health problems (Greco et al. 2008; Greco et al. 2011). For this measure, we also used an average score of the eight items (ranging from 0 to 4).

CDI

The Children’s Depression Inventory (CDI) (Kovacs 1992) is one of the most frequently used instruments, as it is also a widely accepted method among experts in childhood depression, having proved to be significantly reliable from a psychometric point of view (Logan et al. 2013). It consists of 27 items with three response options (0 = absence of symptomatology 1 = mild symptomatology, or 2 = severe symptomatology). Half of the items begin with the option that reflects a higher severity of the symptom being investigated, and, for the other half, the presentation sequence is reversed. We used a validated Spanish version of the CDI (Figueras-Masip, Amador-Campos, Gomez-Benito, and del Barrio 2010), which featured five factors like the original: negative mood, ineffectiveness, interpersonal problems, anhedonia, and negative self-esteem (for a psychometric and comparative perspective, see Logan et al. 2013). This analysis used an averaged measure (ranging from 0 to 2).

STAIC

The State Trait Anxiety Inventory for Children (STAIC) (Spanish version; Spielberger, Goursch, and Lushene 1982) measures the state anxiety (how the child feels in a specific moment), identifying temporary anxiety situations (feelings of qualm, tension, and concern that flow and change in intensity). The STAIC also evaluates anxiety trait (the child’s typical feelings). A total of 40 items is used, 20 for each scale, with a 3-point Likert response model (1 = nothing, 2 = some, and 3 = a lot). For this study, we used an average score (ranging from 1 to 3).

PANAS-C

The Positive and Negative Affect Schedule for Children (PANAS-C) (PANASN, the Spanish version; Sandín 2004) is a version of the original PANAS (Watson et al. 1988) adapted for children, and it is one of most commonly used methods to measure affect (see Hughes and Kendall 2009). The Spanish version is a 20-item (versus 27-item) self-reporting questionnaire in which the items are scored on a 3-point Likert scale (ranging from 1 = very slightly or not at all to 3 = many times). Participants are instructed to indicate how often they have felt this particular way over the past two weeks. The PANAS-C’s subscales have demonstrated adequate internal consistency, with moderate convergent and discriminant validity (Sandín 2004). We also used an average score (from 1 to 3).

ESCQ

The Emotional Skills and Competence Questionnaire (ESCQ) (Takšić 2001) consists of 45 items based on Mayer and Salovey’s emotional intelligence (EI) model (Mayer, Caruso, and Salovey 2016), which defines emotional intelligence as a set of abilities (perceiving, using, understanding, and managing) for processing emotional information in order to attain better personal and social functioning. The ESCQ has three subscales: the perception and emotional understanding scale, the expression and emotional labeling scale, and the emotional control and management scale. Participants rate the items on a 6-point Likert scale (from 1 = never to 6 = always). This self-reporting EI measure has been cross-culturally validated and has shown good reliability rates (Faria et al. 2006). For the present research, we used an average score (ranging from 1 to 6).

Data Analyses

In order to evaluate the impact of the APAC program among gifted children, a repeated measures analysis of variance technique was applied to assess the impact of the treatment between the (1) pretreatment, (2) posttreatment, and (3) follow-up time periods. Where applicable, pairwise comparisons for the main effect were used, utilizing the Bonferroni correction when the sphericity assumption was not met. Results of this analysis are presented in Figs. 1, 2, 3, 4, and 5. To asses internal consistency, Cronbach’s alphas are reported.

Fig. 1
figure 1

CAMM (ranged 0–4) comparisons among pre, post, and 12-month follow-up

Full size image
Fig. 2
figure 2

AFQ (ranged 0–4) comparisons among pre, post, and 12-month follow-up

Full size image
Fig. 3
figure 3

CDI (ranged 0–2) comparisons among pre, post, and 12-month follow-up

Full size image
Fig. 4
figure 4

STAI-C (ranged 1–3) comparisons among pre, post, and 12-month follow-up

Full size image
Fig. 5
figure 5

PANAS-C (ranged 1–3) comparisons among pre, post, and 12-month follow-up

Full size image

Results

Table 2 shows the means and standard deviations for the pretreatment, posttreatment, and follow-up measures. In addition, internal consistency indexes, Cronbach’s alpha, are given where applicable.

Table 2 Descriptive statistics (figures are average scores) and reliability (Cronbach’s alpha) for all study measures
Full size table

The results showed significant differences, in the CAMM questionnaire data (see Fig. 1), between the posttreatment and follow-up assessment stages, F(2,26) = 4.15, p = .027, η2 = .242.

Also, there were differences in the participants’ levels of avoidance and fusion in the AFQ-Y8 results, V = .52; F(2,12) = 6.54, p = .012, η2 = .52. Applying the Huynh–Feldt (1976) correction of estimates of sphericity, the results showed significant differences between these levels, F(1.18,15.32) = 5.72, p = .03, while applying Mauchly’s (1940) sphericity test to validate the repeated measures analysis of variance findings indicated that the assumption of sphericity was violated, χ2(2) = 16.75, p < .01; therefore, multivariate analysis of variance findings is reported (ε = .571). Figure 2 presents these data.

In addition, there were significant differences in the CDI results between the time period groups (see Fig. 3), F(2,12) = 3.62, p < .05, η2 = .376. However, breaking down the results in line with the subscales, the results showed no differences in negative mood, ineffectiveness, interpersonal problems, anhedonia, or negative self-esteem between the pretreatment, posttreatment, and follow-up measures; hence, they are not shown in Fig. 3. Furthermore, the results showed no significant differences in low self-esteem between the assessment stages, F(2,26) = 2.65, p = .090, η2 = .169.

The results showed differences between levels of anxiety state, V = .56, F(2,12) = 7.83, p = .007, η2 = .329. Applying the Huynh–Feldt (Huynh and Feldt 1976) correction of estimates of sphericity, these results showed significant differences between the levels, F(1.14,18.47) = 5.72, p = .013. Mauchly’s sphericity test indicated that the assumption of sphericity was violated, χ2(2) = 8.47, p < .01; therefore, multivariate analysis of variance findings is reported (ε = .711). Correspondingly, these results showed that there were significant differences in anxiety trait between the pretreatment, posttreatment, and follow-up assessment stages, F(2,26) = 11.45, p < .001, η2 = .468. These data are shown in Fig. 4.

Regarding the PANAS-C data, Mauchly’s sphericity test indicated that the assumption of sphericity was violated, χ2(2) = 9.07, p = .01; therefore, multivariate analysis of variance findings is reported (ε = .653). The results showed no differences between the levels of positive or negative affect, V = .244, F(2,12) = 1.93, p = .187, η2 = .244. Applying the Huynh–Feldt correction of estimates of sphericity, the results did not show significant differences between the levels, F(1.39,18.12) = 3.56, p = .064. However, the results did show that there were significant differences in negative affect between the assessment stages, F(2,26) = 4.07, p = .029, η2 = .238. Figure 5 summarizes these data.

Results for the ESCQ showed differences between the levels of perception and understanding, V = .525, F(2,12) = 6.62, p = .012, η2 = .47. Applying the Huynh–Feldt correction of estimates of sphericity, the results showed significant differences between the levels, F(1.1,14.3) = 2.01, p < .01. Mauchly’s sphericity test indicated that the assumption of sphericity was violated, χ2(2) = 22.90, p < .01; therefore, multivariate analysis of variance findings is reported (ε = .540).

These results suggested differences between the levels of expression and emotional labeling, but Mauchly’s sphericity test indicated that the assumption of sphericity was violated, χ2(2) = 2.00, p < .01; therefore, multivariate analysis of variance findings is reported (ε = .56), V = .671, F(2,12) = 12.23, p = .001, η2 = .45. Moreover, applying the Huynh–Feldt correction of estimates of sphericity, the results showed significant differences between the levels, F(1.13,14.71) = 12.38, p < .01.

Similarly, concerning the emotional control and management scale results, Mauchly’s sphericity test indicated that the assumption of sphericity was violated, χ2(2) = 26.16, p < .01; therefore, multivariate analysis of variance findings is reported (ε = .53), V = .452, F(2,12) = 4.95, p = .027, η2 = .45. Applying the Huynh–Feldt correction of estimates of sphericity, the results showed significant differences between the levels, F(2.31,13.98) = 1.07, p = .027. Figure 6 presents all these data.

Fig. 6
figure 6

ESCQ (ranged 1–6) comparisons among pre, post, and 12-month follow-up

Full size image

Discussion

The purpose of the present study was to assess the impact of the APAC program among gifted children. Starting with reliabilities, we found some low values for internal consistency, presumably due to the very low number of items of one of the CDI’s subscales (e. g., ineffectiveness) and the homogeneity of the sample (all gifted children), with very low variance (in the AFQ-Y8, especially). In general, a small, homogenous sample represents a serious drawback for psychological measures (Thompson 2001). However, the alpha range of the measures used in our study extended from acceptable to high. With regard to the ineffectiveness CDI subscale and the very low alpha values obtained, it is note that this subscale includes just five items. Nevertheless, we inferred that there was some misunderstanding occurring when the participants attempted to handle the items, a misunderstanding that may have led to the measure’s low values.

With these reliability issues caveats in mind, the APAC program was broadly based on SG-MBIs (Van Gordon et al. 2015), but we also included activities for optimizing the gifted children’s EI abilities and especially their emotional understanding and emotion regulation (ER). Therefore, the program could be considered a mixed MBI, while some of measures used might also be interpreted as representing dispositional mindfulness (DM; Zhang et al. 2017).

According to our CAMM results, our gifted children did not improve significantly their DM (p = .23, not significantly) upon completing the APAC program. However, after a year, their DM had significantly decreased since the intervening posttest (p = .04) and was almost the same as it was during the initial assessment. As with many interventions, measures may return to initial values, even with MBIs (Langer et al. 2015). However, our participants had acceptable levels of DM (the means for the CAMM were successively 2.79, 2.96, and 2.71; scores ranged from 0 to 4). Future gifted children–MBI researchers using follow-up measures might consider implementing a transversal intervention along the academic year. We also concur with the recommendation of Tomlinson et al. (2018) that “future research may benefit the field [by] employing qualitative methods, which could shed more light on some of the existing findings” (2018, p. 39). In addition, it is worth restating that we used the eight-item Spanish version of the CAMM. This version removes the fifth and tenth items of the original (Turanzas 2013), because of the negative item–total correlation of both of these items, and the fact that they are loaded in a different factor (cf. the two-factor loaded situation recently identified by Ristallo et al. 2016). Therefore, the CAMM average scores might be higher than those obtained by other studies with Spanish samples (e.g., Calvete and Royuela-Colomer 2016).

Regarding avoidance and fusion (measured using the AFQ-Y8), our study’s participants reduced their levels of cognitive avoidance and inflexibility, even though our participants showed a low average of fusion and avoidance at pretest (M = 1.28). Renshaw (2016) carried out, in a sample of 219 high school students, an initial evaluation of the AFQ-Y8 and suggested that a cutoff score of ≥ 15 yielded optimal sensitivity (.86, .92) and specificity (.88, .87). If we had used direct scores instead of average scores with the AFQ-Y8, our gifted children sample would be under Renshaw’s cutoff (i.e., the three AFQ-Y8 assessment scores would be under 10). Renshaw also suggested that this instrument might have a screening purpose for identifying clinical-level depression and anxiety. In our case, the participants did not show signs of any clinical concern, as can be also observed from the CDI and STAIC scores (see Table 2).

Although the gifted children’s depression scores (CDI; Kovacs 1992) and anxiety scores (STAIC; Spielberger et al. 1982) were not at clinically significant levels, the APAC program reduced considerably the depression and anxiety scores in the posttest. The APAC program was shown to reduce levels of depression as much as other MBIs (see Kuyken et al. 2016; Pickard et al. 2016; Renshaw 2016) and levels of both state and trait anxiety (Lee et al. 2008). However, findings with CDI showed that, in fact, participants were depressed, but some differences appeared between the three-time measures, none of them reflecting depressive estate.

The introduction of activities for the optimization of EI abilities along with SG-MBI exercises increased the positive affect and reduced negative affect of our participants (as measured using the Spanish PANAS-C; Sandín 2004). According to the review by Tomlinson and associates (2018), many studies have reported strong evidence of the relationship between DM and both positive and negative affects. Therefore, any activity for optimizing DM and EI abilities may have also be expected to bring about a better well-being (Burke 2009; Schutte and Malouff 2011), to improve ER (Hinterman et al. 2012; Nyklíček 2011), to reduce the impact of negative emotions (Hill and Updegraff 2012), and to increase the positive affect (Kemeny et al. 2012).

The APAC program’s combination of mindfulness and EI ability training produced a better perception of three of the four capacities of the EI framework (perceiving, understanding, and regulating emotions in one’s self and others, but not using emotions; see Mestre et al. 2016). The average scores from the ESCQ (Faria et al. 2006) increased their values. At the 12-month follow-up assessment, average scores for perceiving, labeling, understanding, and managing emotions retained their increased values.

Consistent with our findings, three systematic reviews (Burke 2009; Harnett and Dawe 2012; Tomlinson et al. 2018) found evidence of the benefits of MBIs for school-aged children. As with our study, a large number of the reviewed articles contained a range of methodological issues such as not having control groups, a non-random allocation of participants, and small sample sizes; it is obvious that research into MBIs for children is still in the early stages. Our conclusions, in common with those of the other similar investigations, should be tentative (Tomlinson et al. 2018). The results that we have obtained are promising and feasible, as were those of many of studies reviewed by Burke (2009) or Harnett and Dawe (2012). However, as the most significant and enduring effects of the APAC program are likely to be those found during the follow-up assessments, we remain cautious and offer only prospective interpretations and inferences.

Now we address some issues relating to the study’s internal validity such as the absence of a control group. Given the nature of the gifted children’s association involved in this study and its ethical requirements, we were unable to find a group that agreed not to be trained, even with our commitment to carry out the intervention after the experimental group’s participation had ended. Another complication with regard to our study’s participants relates to the numerous tests and psychoeducational interventions to which gifted children are commonly exposed. However, our sample had no prior knowledge or experience with MBIs specifically. Indeed, it is only recently that scholars have begun to research and report on the potential effects on gifted children of MBIs (Haberlin 2015; Sharp et al. 2017).

A further issue concerns the study’s follow-up assessments. Instead of the recommended 6 months after the posttest, we had to wait 12 months before carrying out the follow-up evaluations, as the end of the MBI coincided with the beginning of the children’s final exams or subsequent vacations. Moreover, when the academic year started again, the children were involved in another brief experiment. In the prior literature, temporal repeated measures have been recommended for overcoming threats to internal validity when a control group is not available (Campbell and Stanley 1963/2010). It should be reiterated, though, that eight participants were not available for the study’s follow-up assessment stage, due to the various reasons.