Process Praise for Enhancing Young Children’s Self-esteem and Prosocial Behavior
유아들의 자아존중감과 친사회적 행동을 향상시키기 위한 과정중심 칭찬
- Publish: Journal of the Korean society of child welfare Volume , Issue47, p35~68, Sep 2014
The purpose of this study is to examine whether process praise is more effective than person praise in enhancing children’s prosocial behavior and self-esteem. For this study, teachers of the control group used person praise while teachers of the experimental group implemented a compliment activity program which was intended to construct a positive verbal environment by offering process praise. The study involved 40 5-year-old children in two different classes at a preschool located in Seoul, South Korea. This study used ANCOVA for the analysis of the results and the Johnson-Neyman procedure to provide additional evidence for the region of significance in the potential impact of process praise. There was a statistically significant difference between the control and experimental group, indicating that process praise was more effective than person praise in developing children’s prosocial behavior and self-esteem. Interestingly, for children who scored above certain points in both self-esteem and prosocial behavior, no significant differences between the experimental and control group were found.
본 연구의 목적은 유아들의 친사회적 행동과 자아존중감을 향상시키는데 있어서 과정중심 칭찬(process praise)이 일반적인(사람중심) 칭찬(person praise)보다 효과적인지 아닌지를 알아보는 것이다. 이 연구를 위해서, 통제집단의 교사는 유아들에게 일반적 칭찬 방식을 사용하고, 실험집단의 교사는 유아들에게 과정중심 칭찬을 함으로써 긍정적인 교실 환경을 구성하고자 칭찬활동 프로그램을 시행하였다. 본 연구는 서울에 위치한 유치원에서 임의로 두 학급을 선정하여 각 학급에 재원중인 총 40명의 유아들을 대상으로 하였다. 연구결과는 칭찬활동 프로그램에 참여한 아이들의 자아존중감과 친사회적 행동 점수는 향상되었고 실험집단과 통제집단 사이에 나타난 점수 차이는 통계적으로 유의미하였다. 이는 과정중 심의 구체적 칭찬이 일반적 칭찬 방식보다 아이들의 친사회적 행동과 자아존중감을 발달시키는데 더 효과적이라는 것을 시사한다. 그러나 자아존중감과 친사회적 행동에 관한 사전측정 검사에서 특정 점수 이상을 획득한 아이들의 경우, 프로그램 실시 후 측정검사 점수에서 실험집단과 통제집단 간 유의미한 차이가 발견되지 않았다. 본 연구는 ANCOVA를 사용하여 칭찬활동 프로그램 참여에 따른 집단 간 변화된 능력차이를 알아보았고, Johnson-Neyman 분석을 사용하여 과정중심 칭찬의 잠재적인 영향력이 가지는 의미 영역(region of significance)을 살펴 보았다.
process praise , self-esteem , prosocial behavior , preschool , ANCOVA , region of significance
This paper is intended to introduce the Compliment Activity Program (CAP) as a nonintrusive and effective means for seeking positive social and personality development and to see if the group that receives process praise has increased in scores of prosocial behavior and self-esteem. We conducted the program, which was designed to construct a positive verbal environment as it offers process praise, in a preschool classroom.
In the program, how children are praised and encouraged mattered. Praise has long been regarded as a naturalistic and nonintrusive management method or intervention strategy for facilitating desirable behavior (Polick, 2011). Although empirical studies have supported that praise contingent on children’s desired behavior reinforces their appropriate behavior and academic skills (Goetz, Holmberg, & LeBlanc, 1975; Sutherland & Wehby, 2001), praise can serve as either an external reward or a facilitator of intrinsic motivation, depending on the ways in which children and teachers interact verbally and nonverbally (Meece & Soderman, 2010). For example, in a negative verbal environment, where teachers’ praise and statements are characterized by comparison, negation, and insincerity, praise merely functions as an external reward (Meece & Soderman, 2010). Therefore, the focus of the compliment activity program we used lies in how teachers praise children rather than whether or not children are given a compliment.
What is important is the quality of praise, not its frequency. Partin, Robertson, Maggin, Oliver, and Wehby (2010) suggested the criteria of quality praise that is effective in increasing children’s positive behavior:
The assumption of the study is that as far as praise meets these given criteria, it would not bring detrimental effects.
The type of praise also affects its quality. There are two different types of praise: person and process.
Person praisejust makes a judgment about what a child has done (i.e., “Good job!”) and delivers no specific information about a child’s performance (Brophy, 1981). It “comments on children’s abilities, goodness, or worthiness after their performance of a task or ⋯ expresses the adult’s global evaluation of the child on the basis of the child’s performance (e.g., disappointment or pride in the child as a whole)” (Kamins & Dweck, 1999: 835). On the other hand, process praisefocuses on the effort or strategy that the child used on a task. Since effort and strategy can be modified when faced with difficulty, process praise keeps children adaptive to both success and failure situations (Corpus & Lepper, 2007). It refers to a praise statement that identifies to a child what a child has done (i.e., “I like the way you play with your friend by sharing the toys with him.”), recognizes the appropriate behavior, gives a child specific feedback, and inform realistic standards of desired behavior (Henderlong & Lepper, 2002; Reinke, Lewis-Palmer, & Martin, 2007; Wright, 2008). It has been found that process praise increases children’s on-task behavior (Sutherland et al., 2000; Wright, 2008), academic gains (Novak & Hammond, 2001), and academic self-concept (Chalk & Bizo, 2004).
Considering the above-mentioned criteria for quality praise and the effectiveness of process praise over person praise, we modified and supplemented the Compliment Activity Program (hereafter CAP) that was designed by Seong (2003) and adjusted by Kim (2004) and Oh (2006). However, CAP in this study is considerably different from the previously known compliment programs. Compliment activity programs used in the above studies and others did not emphasize the use of process praise. In their articles, any description of how to praise is not included. Rather, they focus on the number of praises done as a reward for children’s desirable behavior. In our study, teachers in the experimental group only were asked to use process praise and to take into account the criteria of quality praise when they gave a compliment to children for this project.
The study comes in the wake of long-lasting concerns about how to enhance children’s social and personality development, specifically children’s prosocial behavior and self-esteem. According to Hawkins et al. (1999), reinforcing children’s desired behaviors and enhancing good social skills had a long-term effect on improving both children’s attachment to school and their academic achievement and on reducing the number of violent incidents at school. Furthermore, when a child is victimized by his or her peers, “coping through acting prosocially rather than or in addition to relying on prosocial support from peers may be a more feasible coping strategy” (Griese, 2011: 54). Particularly for children at high levels of victimization, it is difficult for them to ask for social support from their peers. Prosocial behavior also serves a protective role in reducing the impact of relational victimization on children’s loneliness (Griese, 2011).
In addition, research on the relation between self-esteem and aggressive behavior has been controversial. Some have argued that having high self-esteem can prevent children from becoming involved in aggressive and bullying behavior (Anderson, 1994; O’Moore & Kirkhan, 2001; Renzetti, 1992). Low self-esteem is often associated with disruptive behavior (Kirschner, 1992; Wiehe, 1991). However, other studies have found that those who have high self-esteem are more likely to be aggressive as they receive any “aversive feedback that challenges their high self-perceptions” (Hoffman, 2003: 3). In this respect, aggressive behavior becomes a sort of strategy for children with high self-esteem to protect themselves from being exposed to ego threats (Baumeister, Smart, & Boden, 1996; De La Ronde & Swann, 1993). Ha and Edwards (2004) found that low self-esteem was indirectly related to adolescent aggression through the mediating effect of low prosocial behavior. The discussion of previous research on adolescents is included here as it would shed some light onto how to proceed in studying this topic in early childhood. Therefore, it is important to develop positive alternatives such as children’s prosocial behavior and attitude toward others as well as their self-esteem (Hyson & Taylor, 2011). Also, given the controversial research findings about children with high self-esteem, as mentioned above, it is necessary to see if children with different pre-intervention levels of self-esteem and prosocial behavior would respond differently to positive verbal stimuli.
It is also important to distinguish between explicit self-esteem and implicit self-esteem. Explicit self-esteem refers to conscious evaluations of the self. It is captured by self-reports affected by self-presentation bias (Paulhus, 2002). According to Park and John (2011), “newly formed attitudes are accessible at an explicit level, whereas the older, more habitual attitudes may exist in memory, more likely at an implicit level” (74). Unlike explicit self-esteem, implicit self-esteem presents a long-term attitude and is automatically activated evaluation of the self existing at an unconscious level (Park & John, 2011).
This study, however, used a test that measures children’s explicit self-esteem while implicit self-esteem could be measured by using the Self-esteem Implicit Association Test (Greenwald et al., 1998). Therefore, we were not able to show that process praise leads to increases in implicit self-esteem. Above all, Compliment Activity Program implemented in our study was not run for long enough to cultivate such a habitual attitude as implicit self-esteem. Hence, we planned our program in a way that it targets explicit self-esteem rather than implicit self-esteem. The assumption of our study is that explicit self-esteem and implicit self-esteem are different in that the former is consciously presented whereas the latter is outside one’s consciousness.
Most intervention methods introduced hitherto are based on the behaviorist approach. Having children monitor and report their peers’ prosocial behavior helped increasing children’s incidental prosocial behavior (Skinner, Cashwell, & Skinner, 2000) though it is uncertain that the intervention would bring about any long-term impact. Similarly, the findings of Jones, Young, and Friman’s study (2000) support that positive peer reporting on the cooperative behaviors and peer acceptance reinforces a socially rejected youth’s prosocial behavior.
Research on the intervention for self-esteem was limited to the discussion of clinical interventions in most cases. For example, Choi, Lee, and Lee (2008) conducted group music intervention in which children sing songs, make musical instrument, and play instrument. They found that children who participated in the intervention showed increase in self-esteem. Most of self-esteem building interventions target grade-level students since it is believed that students’ self-esteem often decreases during their school years (Wigfield & Eccles, 1994). The contents of the Dalgas-Pelish’ (2006) self-esteem enhancement program deliver the direct message of what influences self-esteem and how the program can improve children’s self-esteem.
Otherwise, some programs (e.g., Incredible Years, PATHS, etc.) that have been devised and used in order to enhance children’s self-esteem do not incorporate either praise or curriculum which could be used in educational settings for young children. For example, Incredible Years provides the guidelines for praise (Webster-Stratton, 1994), but it is not integrated with curriculum.
Although recent research has found that person praise has adverse effects on children with low self-esteem (Brummelman, Thomaes, de Castro, Overbeek, & Bushman, 2014; Brummelman, Thomaes, Overbeek, de Castro, van den Hout, & Bushman, 2014), most of the studies, which examined the effects of praise on children’s self-esteem and prosocial behavior (Jeon, 2011; Jeong, 2007; Kim, 2011; Yoon & Jung, 2010), do not specifically deal with differential effects of person and process praise on children and discuss young children but school aged ones (Jeong, 2007; Kim, 2011). As Brummelman, Thomaes, Overbeek, de Castro, van den Hout, and Bushman (2014) argue, person praise makes children feel attribute their failure to the self and thereby feel ashamed of it, which results in aggravating their emotional vulnerability. Although it has been reported that process praise benefits student motivation and school achievement (Corpus & Lepper, 2007; Pomerantz & Kempner, 2013), there is a paucity of research on how “process praise” affects young children’s self-esteem and prosocial behavior. Therefore, it is necessary to develop a process-praise-based educational program that is intended to improve both children’s prosocial behavior and self-esteem, which can be incorporated into children’s everyday experiences.
This sub-section provides a conceptual ground for this study. Given the far-reaching effects of children’s prosocial behavior and self-esteem on their learning and social development, praise can be an effective tool for facilitating children’s positive behavior and perception of the self and others in classroom settings, without intentionally inculcating moral and social values in the children. Although praise is generally believed to have such positive effects, it is simultaneously regarded as “ineffective and sometimes even dysfunctional” (Henderlong & Lepper, 2002: 774). According to Cannella (1986), children can be dispossessed of agency when praise functions as a controlling instrument. A negative view emphasizes that praise is merely a verbal reinforcement that has an evaluative nature (Henderlong & Lepper, 2002). Another detrimental effect of praise is that it creates pressure (Baumeister et al., 1990) and leads to social comparison (Kohn, 1986). Mosier (2009) argues that praise rather damage[s] a child’s self-esteem by making a child feel pressured into attaining arbitrary standards” (3). Therefore, when a value judgment and an evaluative nature are embedded in praise, it results in detrimental effects on children’s development.
Since there are contrasting views of praise, it is necessary to understand that how we praise children affects the type of motivation with which they will be engaged. For example, praise that brings children to positive affect and emotional gratification can draw children’s intrinsic motivation. Children’s curiosity and enjoyment encourages them to increase their prosocial behavior. On the other hand, when praise involves judgment and evaluation, it takes the form of a social reward by which children’s extrinsic motivation is fueled (Kohn, 1993). In this case, their motivation comes from the external environment.
For this reason, process praise should be used in order to foster children’s prosocial development and self-esteem without limiting their autonomy. Praise is different from simple feedback, recognition, and encouragement. According to Henderlong & Lepper (2002), sincere praise effectively enhances children’s intrinsic motivation and perseverance and has the following characteristics. First, praise does not focus on the outcome, but on the process. Second, it is endogenous. In other words, children’s behavior is not controlled by praise. Third, praise emphasizes children’s positive traits and competence rather than undermines them. Last, it is descriptive and guides children to understand high, yet realistic, expectations for them. Process praise that was used in our study was actually based on these features of the praise considered to promote children’s intrinsic motivation. According to Kanouse et al. (1981), when praise is general enough to be inconsistent with children’s perception of the self, they are likely to reject the praise. Therefore, praise should be specifically describing what they do. These are the features that comprise quality praise.
Quality praise enhances children’s prosocial behavior and self-esteem. First, prosocial behavior such as sharing, helping, comforting, and informing is “voluntary behavior intended to benefit another” (Eisenberg, Fabes, & Spinrad, 2006). It emerges in the early years (Eisenberg, Fabes, & Spinrad, 2006; Warneken & Tomasello, 2012). Hyson and Taylor’s (2011) extensive literature review highlights that educators can develop children’s prosocial behavior in the following ways: by building secure relationships with children, by creating a close-knit caring community in a classroom, by modeling kindness and consideration, by having prosocial expectations for children, and by supporting and assisting families through family and community outreach. Among these strategies, building secure relationships with children requires “responding sensitively to children’s everyday needs, interacting in emotionally supportive ways, [and] listening and conversing with sincere attention” (Hyson & Taylor, 2011: 77). Giving a compliment to children involves these components.
Secondly, a positive verbal environment in which children experience sincere, constructive, and encouraging praise (Gartrell, 2007) fosters children’s self-worth and self-esteem (Kostelnik et al. 2009; Meece & Mize, 2009; Meece & Soderman, 2010). Children build self-confidence when they are praised. This eventually strengthens their self-esteem (Lee, 2001; Seong, 2003). Acknowledging the positive reciprocal correlation between self-esteem and prosocial behavior (Eisenberg & Fabes, 1998), it is effective to enhance both personal traits in order to create a non-violent and supportive learning environment for children and ensure children’s optimal development.
Considering the contrary findings about the effects of praise, it is necessary to examine whether praise can be effective in enhancing children’s prosocial behavior and self-esteem when it meets the above-mentioned criteria of quality praise. Most of the hitherto known programs that used praise as an intervention method concern only how often and when you should praise children. They do not point out the quality of praise―ways in which you should praise children. However, in our study, CAP incorporates praise into our class activities and emphasizes using process praise that meets the criteria for quality praise noted at the beginning of the introduction section. If praise is descriptive, consistent, continuous, and focused on specific aspects of children’s behavior or work, we presume that praise can be beneficial to children’s self-esteem and prosocial behavior.
The research questions of this study are as follows.
1. Does the group that receives the Compliment Activity Program have increase in children’s self-esteem?
2. Does the group that receives the Compliment Activity Program have increase in children’s prosocial behavior?
3. Does the program have the equivalent effect across children at the different developmental levels of self-esteem and prosocial behavior?
For this study, we selected 40 5-year-old children in two different classes at a preschool located in Seoul, South Korea. 20 of them (7 boys and 13 girls) in one class were assigned to the experimental group whereas 20 (7 boys and 13 girls) in the other class were assigned to the control group (Table 1). The preschool used the integrated curriculum called the Nuri Curriculum. The numbers of boys and girls in both classes were identical. All of them were from middle socioeconomic families. The two classes were not different from each other in any significant way; they were not different from their counterparts in other preschools in the urban areas in South Korea.
] Sex characteristics of subjects
The four female teachers of the classes (one head teacher and one assistant teacher for each class) participated in the study. The head teachers for the two classes had similar educational backgrounds for teacher training. The teacher for the experimental class had taught for 10 years, and the teacher for the control class had taught for 10 years. It was at the beginning of the semester that the Compliment Activity Program was implemented. Children were randomly assigned to classes according to the preschool policy at this time of the year without considering any pre-determined criteria. The classrooms were randomly selected. Although two comparable classrooms were chosen, it is not completely sure that the groups are comparable. Since there was the potential nonequivalence between the groups due to the lack of randomization in sampling individual children, a nonequivalent-groups design, a frequently-used form of quasi-experimental designs, was used as in reality it is not always possible to randomly select samples (Trochim & Donnelly, 2008). Analysis of Covariance (ANCOVA) as a statistical procedure also helped reduce the effects of initial group differences by adjusting posttest scores for variability on the covariate (pretest scores) (Maxwell & Delaney, 2004).
This ex-post-facto study was designed to evaluate the effectiveness of process praise on children’s prosocial behavior and self-esteem. Therefore, the independent variable was process praise while the dependent variables were prosocial behavior and self-esteem scores. To best assess the effects of process praise, this causal-comparative study was intended to compare the prosocial behavior and self-esteem scores of the children in the experimental group with the prosocial behavior and self-esteem scores of the children in the control group. Before we started the program with the experimental group, we administered a pre-test for both groups in order to measure their prosocial behavior and self-esteem scores. Then, the program using process praise was implemented for the experimental group only while children in the control group participated in the similar activities based on the same curriculum. However, the only difference between experimental and control groups was the contents of activities during circle time and the forms of praise that teachers used (whether praise was person or process).
1) Prosocial Behavior Test
Since the development and prevalence of prosocial behavior varies across different cultures (Benenson, Markovits, Roy, & Denko, 2003), a test measuring Korean children’s prosocial behavior should be suitable for use in Korean culture. Although the measurement methods are culture specific in this study, the Compliment Activity Program is based on the idea that process praise is a universal tool for encouraging and promoting children’s prosocial development and self-esteem.
The prosocial behavior test was developed by Kim (2003). It consists of 42 items, including 7 sub-sections: leadership (8 items), helping (8 items), communication (7 items), autonomous regard (6 items), approaching (5 items), sharing (3 items), and empathy (4 items). Each item is on a 5-point Likert scale and could be rated on a 1-to-5 response scale. The higher the scores are, the more one is likely to show prosocial behavior. Cronbach’s alpha was used to assess the reliability of the Prosocial Behavior Test with respect to (a) how well the individual items of the scores fit together and (b) whether they assess the same construct (Table 2).
] tem compositions and Cronbach Alpha value of a Test for Children’s Prosocial Behavior
If the Cronbach’s alpha is above 0.7, a scale is regarded as having good internal consistency. As the Cronbach’s alpha values for all sub-categories are above or close to 0.7, the items show good internal consistency. Although the Cronbach’s alpha values calculated from the ‘communication’ and ‘empathy and emotional control’ measures were
α= .67 and α= .66, respectively, according to Nunnally (1978), a Cronbach’s alpha of 0.6 is sufficient to be an acceptable value for research purpose.
2) Self-esteem Test
For this study, we used a self-esteem test designed by Harter and Pike (1984) and modified by Kim (1997). The test includes the following categories: cognitive ability, peer acceptance, physical ability, mother acceptance, and self acceptance. Each category has 6 items, and the test has 30 items in total. Each item is on a 4-point Likert scale and could be rated on a 1-to-4 response scale. The higher the scores are, the higher a child’s self-esteem is. The Cronbach’s alpha for this test is
α= .80 (Table 3). Each category shows good internal consistency.
] tem compositions and Cronbach Alpha value of a Test for Children’s Self-Esteem
1) Teacher Training
The two female teacher participants in both groups have more than 10 years of teaching experience and majored in early childhood education. Before we started the program, the teachers were trained in understanding the role of a teacher and the purpose and method of the program for three days. The teacher in the experimental group was informed of the role of a teacher and the purpose of the study and trained in learning how to use process praise. For training, teachers were asked to talk about their understanding of praise. The researcher who was an expert in process praise observed and corrected the way the teachers praised children so that they could distinguish process praise from person one. For the teacher in the control group was informed of using person praise only, not process praise, while the teacher in the experimental group was asked to use process praise only.
In the experimental group, the focus of the teachers’ praise was on the effort or strategy that children used on their tasks. For example, children in this condition were told, “I am sure that you worked really hard on the task.” Teachers in the control group used person praise which “conveyed an evaluation of the child based on his or her performance” (Kamins & Dweck, 1999: 838). Children in the control group were given evaluative feedback that comments on children’s worthiness and abilities (e.g., “You are a good boy.”; “You are really good at this.”). By communicating with teachers (i.g. meetings and phone calls) on a weekly basis, researchers made sure whether teachers in the experimental and control groups actually offered different types of praise.
Prosocial behavior and self-esteem tests were administered to 40 children in experimental and control groups. As for the tests, the researchers showed pictures to an individual child, gave a direction, and recorded his or her response to each picture. All tests were administered for one day at the end of the first month of the school year.
3) Compliment Activity Program
The program was conducted for 9 sessions for 4 weeks a couple of weeks after the semester started. Therefore, all participating children of the classes were new to their teachers, curriculum, and classroom circumstances. The two classes were operating the exactly same curriculum based on the national ‘Nuri’ curriculum, and the activities offered in the classes were not different in any significant way. Each session of the program had different topics selected in accordance with the preschool curriculum and utilized different materials (Table 4). We used a modified version of the program initially designed by Sung (2003).
] Topics of the Sessions in the Compliment Activity Program
We set up goals for each session which aim to promote children’s self-esteem and prosocial behavior through building positive attitudes towards the self and others including peers, family, and teachers. Children’s engagement in self-understanding and self-appreciation activities was extended to developing their understanding of, respect for, and positive action towards others. During circle time, the children started singing songs related to the topic of the session. For example, in session 2, the topic of the session was the self. The goal was to appreciate the self and find the strength of the self. The children sang a song that praised their faces. The teacher read a book relevant to the topic, titled “A bear danced when you were born”. Then, the children were asked to observe their faces in a mirror, to talk about good things about themselves, and to draw a picture describing their strengths (e.g., the teacher said, “Let’s draw a picture on what we are good about.”). After completing the drawing, the teacher asked each child what he or she thought about their own pictures. Then, she praised the children about their pictures, pointing out specific aspects of what was nicely captured and expressed in the picture of each child. Finally, the children talked about their feelings about the activity. The brief summary of a session of the Compliment Activity Program is as follows.
• Topic of the activity: Self (Session 2)
• Age of children: Age 5
• Type of the activity: Story sharing
• Duration of time: 20-30 minutes
• Goal: Appreciate oneself and find one’s strengths.
• Relevance to the curriculum:
• Materials: Picture book (A bear danced when you were born), papers, crayons, mirrors, camera.
(1) Intro: ① Sing a song about their faces while sitting in a circle.
(2) Activity development: ① Teacher reads a book ‘A bear danced when you were born.’
(3) Wrap-up: ① Children talk about their feelings about the activity.
Each session lasted for about 30 minutes. During the 4-week period of time, the teachers tried to use process praise continuously and consistently throughout the entire daily activities. Activities given for children in the control group were designed based on the same national curriculum using almost the same activities that the experimental group did. Therefore, the contents of the activities done in both groups were not significantly different. Furthermore, the teacher in the control group used person praise as a form of verbal reward on a child’s task accomplishment whereas the teacher in the experimental group was continuously trained for how to use process praise which describes, recognizes, and appreciates what a child has done. The teachers in the two groups did not know that the praise received by the other class was different from that which they delivered.
In order to measure the effectiveness of process praise, we carried out the specific program rather than letting teachers praise children in natural settings. This allows for the possibility that the program could be replicated in an organized manner across different contexts in which children develop and learn. This systematic approach reduces the risk of substantial variation in praise patterns occurring in different forms of relationships.
After the program was terminated, the children were tested for their prosocial behavior and self-esteem scores. The same measurement tools were used for the post-test.
The pre- and post-test means and standard deviations in self-esteem scores of the experimental and control groups are presented in Table 5.
] Pre-test and Post-test Means in Self-Esteem
Table 6 presents the pre- and post-test means and standard deviations in the prosocial behavior scores of the groups.
] Pre-test and Post-test Means in Prosocial Behavior
These data reveal the lack of equality of the groups. Also, as can be seen from the data in Tables 5 and 6, both of the groups showed improvement in the post-test scores in prosocial behavior and self-esteem.
After administering the pre-test, we utilized an independent sample t-test to compare the mean scores of the two groups to see if there was any significant difference between the groups regarding their self-esteem and prosocial behavior pre-test scores. There was difference between the self-esteem pre-test scores of the control and the experimental groups(t(38)= -2.339, p=.025). However, regarding the prosocial behavior pre-test scores, there is no significant difference between the two groups (t (38) = .163, p=.872). For this reason, we conducted ANCOVA after computing the mean scores of the groups in the pre- and post-tests.
The classroom was used as the unit of analysis as a classroom within treatment effect was statistically significant. The adjusted mean and standard error for the two groups of children’s self-esteem scores after completing the program are shown in Table 5. According to the ANCOVA results, the difference between the post-test self-esteem scores in the control and experimental groups was statistically significant (F = 83.999, p < 0.001) as seen in Table 5. This result indicates that the group that received Compliment Activity Program had an increase in the total self-esteem scores (Table 5). The pre-test self-esteem score of the experimental group was M(SD)=90.70 (6.578) and its post-test score was M(SD)=106.50(3.72). The mean self-esteem score in the experimental group increased significantly, compared to the control group. In the five sub-factors of self-esteem, children in the experimental group showed significant improvements (cognitive ability (F=42.669, p < 0.001), peer acceptance (F=73.131, p < 0.001), physical ability (F=46.142, p < 0.001), mother acceptance (F=40.045, p < 0.001), and self-acceptance (F=5.747, p < 0.05)). The range of the post-test self-esteem scores of the control group was from 72 to 110 while that of the experimental group was from 101-114. This finding supports the research hypothesis that CAP would be beneficial in enhancing children’s self-esteem.
Likewise, the difference between the post-test prosocial behavior scores in the control and experimental groups was statistically significant (F = 173.390,p < 0.001) as seen in Table 6. This shows that the group that received CAP also had an increase in the total prosocial behavior scores (Table 6). The pre-test prosocial behavior score of the experimental group was M(SD)= 120.15(18.239) and its post-test score was 182.95(16.96). There was a significant increase in the seven sub-factors of prosocial behavior (leadership (F=107.970, p < 0.001), helping (F=136.812, p < 0.001), communication (F=133.674, p < 0.001), autonomous consideration (F=165.469, p < 0.001), approaching (F=79.327, p < 0.001), sharing (F=55.691, p < 0.001), and empathy and emotional control (F=54.429, p < 0.001)). The range of the post-test prosocial behavior scores of the control group was from 81 to 147 while that of the experimental group was from 145-201. This finding also supports the research hypothesis that the prosocial behavior of the children in the experimental group would be enhanced compared to the control group. For self-esteem, the statistical power of the experiment, which was measured by G-power ver. 3.1., is 0.98, which indicates that the CAP was effective (
α=.05, η2 ρ=.69). The statistical power of experiment for prosocial behavior is 0.99 ( α=.05, η2 ρ=.82). According to Cohen (1988), 0.9 is considered an adequate value for power.
The interactions between the covariates and the independent variables in Group * Pre-test Prosocial Behavior (F=7.467, p=.010) and Group * Pre-test Self-esteem (F=23.408, p <.001) were found to be significant. Therefore, it may not be possible to make a general statement about the difference between the groups, adjusted for the covariate. However, it is possible to say that there is a significant difference between the groups for some values of the covariate, but not for others.
Since we found the interactions between the covariate (group) and the independent variable (pre-test scores) in both the prosocial behavior and self-esteem scores, we used Johnson-Neyman techniques to identify the region of significance while still holding the results of the ANCOVA. According to D’Alonzo (2004), when the assumption of homogeneity of the regression slopes is violated with the ANCOVA, the Johnson-Neyman procedure works as an alternative approach. Hayes and Matthes (2009) also note that the J-N technique “provide[s] information about the range of values of the moderator where the focal predictor has a statistically significant effect and where it does not” (928). Here, the moderator refers to a group variable. The J-N procedure has been recommended when the regression slops are heterogeneous and the group effects are the main interest (Hayes & Matthes, 2009; Johnson & Neyman, 1936; Kowalski, Schneiderman, & Willis, 1994). The region of significance was calculated using the computational tools D’Alonzo (2004) provides.1)
For children whose self-esteem pre-test scores were below 102.67 the Compliment Activity Program was beneficial in enhancing their self-esteem, but for children who score above 115.86 in the pre-test for self-esteem, the program appears to make no difference. In this study, the maximum pre-test scores in self-esteem were below 115. For children scoring between 102.67 and 115.86 in their pre-test, there is insufficient evidence to conclude that the program made any difference in their self-esteem.
Regarding children’s prosocial behavior, the Johnson-Neyman region of nonsignificance was calculated to determine where the experimental and control groups differed.2) According to the J-N procedure, the region of nonsignificance was 169.82 to 501.94. For children whose prosocial behavior pre-test scores were less than 169.82, there was a significant difference between the control and experimental groups. That is, the Compliment Activity Program was beneficial in increasing the prosocial behavior for children who scored less than 169.82 on prosocial behavior. As the upper bound far exceeds the range of interest, only the lower bound (denoted by the XL vertical line in Figure 2) was included in the figure. Therefore, it can be concluded that for children scoring higher than 169.82 in their pre-test, there was no difference between the experimental and control groups’ post-test prosocial behavior scores. In other words, there is insufficient evidence to conclude that the program was beneficial in developing prosocial behavior for children with already high scores.
1)The region of significance calculated by the computational tools is as follows. See D’Alonzo (2004) for the computations for the region of significance. X: covariate Fα = F ratio from the F table N = Total sample size n1 , n0 = Number of subjects in groups Z = 1 (control group) and Z = 0 (experimental group) respectively SSres = Residual sum of squares for interaction model For children’s self-esteem scores, I computed the region of significance as follows: 2)The Johnson-Neyman region of nonsignificance was calculated using the following computations.
The results of the present study show that the Compliment Activity Program that used process praise with young children is beneficial in children’s self-esteem and prosocial behavior. Therefore, the study has supported the first and second research hypotheses that the Compliment Activity Program makes difference in children’s self-esteem and prosocial behavior. The curriculum and activities used in the experimental and control groups were not different in any significant way. Given that the program features teachers’ use of process praise for children’s desired behavior while children in the control group were given person praise, the finding of this study supports the idea derived from the findings of other studies that quality/process praise is beneficial in increasing children’s self-esteem and prosocial behavior (Eisenberg, Fabes, & Spinrad, 2006; Hyson & Taylor, 2011; Meece & Soderman, 2010).
However, interestingly, for children who scored higher than certain points in both self-esteem and prosocial behavior, there was no significant difference between the experimental and control groups. Accordingly, this study not only showed the group differences in children’s self-esteem and prosocial behavior when process praise exists but also provided additional evidence for the region of significance in the significant difference in the before-after treatment. That is, the regions of significance show that children who received process praise had an overall increase in their self-esteem and prosocial behavior.
Given the limitation of this study that only examined whether compliments enhance children’s self-esteem and prosocial behavior, it is uncertain as to why the program did not make any significant difference for children high in self-esteem. This could be a ceiling effect that occurs when subjects with high scores cannot show improvement on a measure. Otherwise, one possible explanation of these findings is that children high in self-esteem were affected relatively less by praise compared with those with low self-esteem. Seery, Weisbuch, and Brooke Vick (2004) found in their study that people with unstable high self-esteem exhibited a threat pattern similar to those with stable low self-esteem and showed the lowest task engagement, whereas those with stable low self-esteem exhibited the highest. The cause of this is unclear, yet they discuss that the underlying self-doubt that individuals with unstable high self-esteem possess makes them defensive in the event or to the feedback from others, thus leading to little change in the emotional and behavioral reactions of those with unstable high self-esteem. According to Trzesniewski, Donnellan, & Robins (2003), self-esteem stability is low during childhood. Therefore, the children in this study are also likely to have unstable self-esteem. For this reason, praise might not make any significant difference in the self-esteem and prosocial behavior of children with unstable high self-esteem who also tend to be at the higher level of prosocial development.
The findings of this study do not provide any sufficient evidence as to why the impact of the Compliment Activity Program were not significant for children with a certain level of self-esteem and prosocial behavior. However, it is recommended at least that practitioners working with children should not expect the same outcomes of their praise on all children. Particularly for children scoring less than certain points of self-esteem and prosocial behavior, creating positive verbal environments full of process praise helps children build positive self-perception and behavior towards others. Most of the children in this study scored within the region of significance, except for a child whose scores showed no treatment difference. This indicates that the Compliment Activity Program was beneficial overall in both domains though this study does not give any causal relations between process praise and targeted traits.
At least, praise was effective in strengthening the relation between self-esteem and prosocial behavior. For the control group, the correlations between pretest self-esteem and pre-test prosocial behavior and between pretest self-esteem and post-test prosocial behavior were not significant. Likewise, the correlation between pre-test self-esteem and pre-test prosocial behavior was insignificant for the experimental group, as there was no praise intervention in this case. However, the pre-test self-esteem scores in the experimental group were positively correlated with post-test prosocial behavior scores (r=.515, p<.01), which shows the influence of the program on the change in this relation. Although self-esteem is known to be positively correlated with prosocial behavior (Ha & Edwards, 2004), this study has found that the correlation was significant when there was praise intervention.
The implication of this study is that practitioners working with children in the field should use process praise in order to support children’s abilities. As humans are social beings, prosocial behavior is not only a key to one’s happiness as we all live with others, but also seeds resilience which enables one to handle everyday stressors (Duquo, Klevens, Ungar, & Lee, 2005). Given the positive effect of process praise on children, it is significant to provide practitioners, particularly those who work with children facing stressful situations at home and/or school, with guidelines for the compliment activity program with process praise as well as training opportunities for practices, enhancing the quality of education and care services given in children’s microsystems including child care, (after)school, and community settings. This would ultimately support children’s emotional and social well-being, promoting their inner strengths.
In addition, it is necessary to train professionals who can provide customized services suited to children’s individual and developmental needs while ensuring the universality of the target population. The current issues in the policy regarding child welfare services include the lack of universal services for all children as most of the child welfare services are geared towards the specific needs of children requiring protection. Therefore, the process praise approach can be included in the training education for practitioners working with all children as it enhances their professional capacities to effectively support children’s social skills and developmental needs, which would promote a level of professionalism in the services provided for children in turn.
The study has some limitations in that it did not account for possible effects of contextual variables such as the relationship between the evaluator and the children and whether the classroom climate in the control and experimental groups was cooperative or competitive. Henderlong and Lepper (2002) pointed out that it is important to take into account contextual factors in assessing the effects of praise on children, and the contextual factors can be considered in future research regarding the effect of praise on children’s self-esteem and prosocial behavior. Second, as the study participants were limited to 5 year olds, it would be beneficial to involve children of different ages and examine the effects of the program by children’s age. Third, given that there were less change in children who were already high in self-esteem and prosocial behavior at pre-tests, more subjects could help clarify these results. Due to the small sample size, it was not possible to look at gender differences which might be another important factor to be considered. Finally, the study conducted the Compliment Activity Program only for a period of 4 weeks. The duration of the program can be extended so that the program effects could be assessed more precisely.
[<Table 1>] Sex characteristics of subjects
[<Table 2>] tem compositions and Cronbach Alpha value of a Test for Children’s Prosocial Behavior
[<Table 3>] tem compositions and Cronbach Alpha value of a Test for Children’s Self-Esteem
[<Table 4>] Topics of the Sessions in the Compliment Activity Program
[<Table 5>] Pre-test and Post-test Means in Self-Esteem
[<Table 6>] Pre-test and Post-test Means in Prosocial Behavior
[<Figure 1>] The region of significance in the influence of Compliment Activity Program on children’s self-esteem
[<Figure 2>] The region of significance in the influence of Compliment Activity Program on children’s prosocial behavior