Student and Parent Attitudes Before and After the Gifted Identification Process Matthew C. Makel Duke University Talent Identification Program Questionnaires were given to students prior to testing for giftedness to assess how the gifted identification process changes attitudes toward gifted programs and parent-child relationships. after those identified as gifted had been participating in a pullout gifted program for one semester, another questionnaire was administered to all original par- ticipants. Comparable data were also collected from the students’ parents. results sug- gest that prior to identification, gifted and nongifted students and their parents shared similar attitudes toward giftedness. after identification, attitudes showed either no change or decreased. Compared to their gifted counterparts, nongifted students and their parents tended to report lower attitudes after identification. the impact of this attenuation and the use of preidentification data are also discussed. Previous research on gifted youth has typically been conducted by investigating students who have already been labeled and are already participating in gifted programs (e.g., Lubinski, Webb, Morelock, & Benbow, 2001). This occurs even when the identification process itself is one of the primary areas of interest (e.g., VanTassel-Baska, Feng, & Evans, 2007). This may unintentionally lead to an assumption of a tautology: If a student is in a gifted program, she is considered gifted; if she is not in a gifted program, she is not gifted. Relying solely on data gathered from students who are already participating in gifted programs fails to account for selection biases for membership into that group. Further, it could prevent an accurate analysis of differences— preexisting or otherwise—between gifted and nongifted youth. It is reasonable to assume that, prior to identification, the two groups have cognitive differences. However, if other differences (e.g., attitudes toward school or social relationships) exist, analyses restricted to post- identification of gifted youth should consider that restriction—but they typically do not. Matthew C. Makel is Gifted Education Research Specialist at the Duke University Talent Identification Program. Journal for the Education of the Gifted. Vol. 33, No. 1, 2009, pp. 126–143. Copyright ©2009 Prufrock Press Inc., http://www.prufrock.com 126 Attitudes and the Gifted Identification Process 127 This paper proposes that there are four distinct and relevant events related to the gifted identification process that could cause significant change in attitudes toward giftedness. These four events are nomina- tion, testing, labeling, and participation. Although I could not locate a study that addresses how each event influences students, it is predicted that any and all of these events can potentially be highly influential. One student could burst with pride upon being nominated for a gifted program, whereas others could feel immense pressure. Similarly, a student being tested for giftedness could leave the testing situation confident or unsure. The event that most explicitly divides students is that of labeling; some are told they are gifted, whereas others are told they are not. Finally, as eligible students begin participating in gifted programming, some may respond positively while others may struggle. As noted above, not all previous research has focused entirely on postidentification gifted youth. For example, in the Fullerton Longitudinal Study (FLS; Gottfried, Gottfried, Bathurst, & Guerin, 1994), the investigators did not label participants as gifted or non- gifted. The FLS consisted of an array of developmental, cognitive, and behavioral measures administered to 107 one-year-old infants and their parents. Parents were also asked to rate their children’s per- formance and abilities. At age 8, children were given the Wechsler Intelligence Scale–Revised (WISC–R; Wechsler, 1974) and deter- mined to be gifted (n = 20) or nongifted (n = 87). The investigators then retrospectively compared group differences across the measures they had been collecting in real time for the previous 7 years. At no point of the data collection were participants notified of the results of any measures or of gifted status. Thus, the FLS was a study that wholly sampled preidentified youth and eliminated potential effects possibly solely due to being labeled gifted. Gottfried et al. (1994) found cognitive performance differences as early as 18 months. Gifted participants generally performed higher than nongifted participants.1 Parents of gifted youth rated their child’s performance higher than did the parents of nongifted participants. Additionally, parents of gifted children provided more enriching and stimulating environments (e.g., more books in the home). However, as Gottfried et al. pointed out, the parent-child relationship is bidi- rectional; youth help shape their environment by way of requesting and reacting to parent actions. 128 Journal for the Education of the Gifted While Gottfried and colleagues (1994) studied gifted youth prior to identification, other researchers have investigated how student atti- tudes change as they begin to participate in gifted programs. For exam- ple, Gibbons, Benbow, and Gerrard (1994) compared the changes in student attitudes before, during, and after participation in a summer program for gifted youth. Although their initial measure occurred prior to participation in the gifted program, it was after the student had been accepted to the program. Thus, any changes that may have taken place during the identification process were not revealed. In general, as opposed to attitudes prior to participation in the summer program, they found that male students who were not performing well in the gifted program (relative to their peers) reported that academics were not as important to them; however, academic importance returned to initial levels during the third measurement. Hardly any change was reported for female students (Gibbons et al., 1994). In a similar fashion, Coleman and Fults (1985) contrasted the self-concept of gifted students with relatively low IQ scores (e.g., a student whose IQ was slightly above the cut-off score) who had begun participating in a gifted program with a comparable group who had not yet begun participating in a gifted program. The group who had begun participating reported slightly lower self-concept scores than the yet-to-participate group. Other researchers (e.g., Marsh & Hau, 2003; Marsh, Kong, & Hau, 2000) have developed numerous theories explaining how students’ self-concept changes as a result of participat- ing in gifted programs. However, numerous social, emotional, and cognitive factors need to be taken into account when investigating how attitudes change over time. Researchers (e.g., Jacobs, Lanza, Osgood, Eccles, & Wigfield, 2002) have investigated students’ competence beliefs (i.e., belief in possessing the ability to perform a task) and subjective task value (i.e., the degree to which a student values being able to perform a task). In multiple domains (e.g., math or language learning), both competence beliefs and subjective task value have been found to decline starting as early as first grade. However, much of the task value decline could be accounted for by controlling for competence beliefs. Additionally, students were more likely to value a task if they considered themselves competent at it. Attitudes and the Gifted Identification Process 129 Others (e.g., Fredericks, Blumenfeld, & Paris, 2004) have sug- gested that school engagement or commitment (in terms of behaviors, emotions, and cognition), may play a significant role in student atti- tudes and can be facilitated (or impeded) by specific learning environ- ments. It may be that the gifted identification process acts as a catalyst for (or cushion against) the decline in student engagement (and thus their attitudes as well) in school. If this is the case, participating in the gifted identification process could have extensive impact on attitudes toward school in general as well as toward gifted programs. To assess the impact of participating in the gifted identification process, the current study analyzed how students’ and parents’ atti- tudes toward gifted programs and school in general changed while par- ticipating in the gifted identification process. Two notable additions are made in the current study. First, this study expands on traditional methods of analysis by comparing attitudes prior to being tested for giftedness with attitudes after the labeling process. Second, students tested for, but not identified as, gifted are included in the analysis. This group is relevant because it is unknown whether merely partici- pating in the gifted identification process influences student attitudes toward, and performance in, school. In schools that have separate testing for giftedness, this unique group is officially told, “You are not gifted.” Such a statement could have numerous significant long-term effects. Thus, telling a third grader that she is not gifted could create a self-fulfilling prophecy wherein she performs worse than she would have had she not participated in the gifted identification process. The purpose of the current research is to investigate how the identification process is associated with the attitude change of (a) gifted youth, (b) youth who are deemed ineligible for participation in a gifted program, and (c) the parents of both groups of students. Method Participants Letters seeking participation were sent to all families in a suburban school district in the Midwestern United States who had nominated 130 Journal for the Education of the Gifted a child to be tested for giftedness. The gifted program serves third- through sixth-grade students; thus, second- through fifth-grade stu- dents were eligible to participate in the testing process. To test for giftedness, the school district used the Cognitive Abilities Test Form 6 (CogAt; Lohman & Hagen, 2001). The cutoff score used by the school district for participation in the gifted program was the 92nd percentile. Of the 142 families that were sent letters, 49 returned forms. Nine participants were deemed eligible to participate in the gifted program (n = 4 girls) and thus labeled gifted; 36 (n = 23 boys) were deemed ineligible. Henceforth, the two groups will be referred to as gifted and nongifted students respectively. Measures Questionnaires of student and parent attitudes and perceptions were developed specifically for this study in order to maximize the cor- respondence between items and the objectives of this investigation. These questionnaires consisted of items on a 5-point Likert scale rang- ing from 1 Never to 5 always, quantifying how each item described the participant. The questionnaires had five subscales. The Gifted Program–Academic subscale measured attitudes toward the academic aspect of participating in a gifted program (e.g., “Being in a gifted program will make me like school more than I do now”). The Gifted Program–Social subscale measured attitudes toward the social aspect of participating in a gifted program (e.g., “Being in a gifted program will help me make friends”). The General Academic subscale measured general perceptions of academic ability and school performance (e.g., “School is easy for me”). The Parent–Academic subscale measured perception of the child-parent relationship on academic issues (e.g., “My parents think I try hard in school”). The Parent–Social subscale measured perceptions of the child-parent relationship on social issues (e.g., “I get along with my parents”). See Tables 1 and 2 for a complete list of the subscales and items. The internal reliability of each subscale was calculated via Cronbach’s alpha level. In general, Cronbach’s alpha levels were low (see Tables 1 and 2 for Cronbach alpha levels). However, for the most part, they were not so low as to not merit further analysis (Feldhusen & Dai, 1997). One notable aberration was the Gifted Program–Social Attitudes and the Gifted Identification Process 131 Table 1 Student Subscales Subscale (a) Item Gifted Program–Academic (.76) Being in a gifted program will make me like school more than I do now. Being in a gifted program would be fun. School would be more fun if I was in a gifted program. General Academic (.46) I like school. I am smarter than my friends. I am smart. I get bored in class. (R) I try hard in school. School is easy for me. Parent–Academic (.62) My parents ask me about school. My parents tell me I am smart. My parents think I try hard in school. Parent–Social (.58) I have fun with my parents. My parents like my friends. My parents and I have arguments about school. (R) My parents and I agree on things. My parents let me have fun. I get along with my parents. If I have a problem, I can talk to my parents about it. Note. (R) reverse scored item. Cronbach alpha values are in parentheses. subscale. Reliability for this subscale was so low that its results do not merit further discussion in the Results section. Procedure As stated previously, all students who had been nominated by their par- ents to be tested for giftedness were sent a letter informing them that a research study was being conducted on student and parent attitudes toward school. Enclosed with the letter were consent forms as well as the first parent questionnaire. I then visited each of the district’s six elementary schools to administer the Time 1 student attitude ques- tionnaire. Two weeks later, students participated in the district’s testing 132 Journal for the Education of the Gifted Table 2 Parent Subscales Subscale (a) Item Gifted Program–Academic (.87) My child would prefer being in gifted classes. My child would enjoy participating in a gifted program. Participating in a gifted program would make school more fun for my child. General Academic (.60) My child is smart. My child is smarter than his/her friends. School is easy for my child. My child likes school. My child gets bored in class. (R) Parent–Academic (.54) My child tries hard in school. I tell my child he/she is smart. I speak with my child about what happens to her/him at school. Parent–Social (.61) My child and I agree on things My child and I get into arguments about school. (R) I like my child’s friends. My child and I get along. My child has fun with me. If my child has a problem, she/he feels comfortable talk- ing to me or another parent about it. Note. (R) reverse scored item. Cronbach alpha values are in parentheses. for giftedness. About 6 weeks afterward, the district notified those tested whether they had been identified as gifted. The following fall, eligible students began participating in the gifted program. One year after the first questionnaire, after one semester of potential participa- tion in a gifted program, the Time 2 questionnaires were administered using the same procedure as Time 1. Between-subjects and within-subjects analyses were conducted with respect to the subscales. The nonparametric Wilcoxon signed- ranks and Wilcoxon rank-sum tests were used for all analyses that contained a group with an n < 30 (i.e., any analysis involving a gifted group) because normality could not be assumed with such a small sample. Independent sample and paired-sample t-tests were conducted Attitudes and the Gifted Identification Process 133 for all other analyses. Bias corrected effect sizes ĝ (Hedges & Olkin, 1985) were calculated for all analyses. Because the sample size was small, some findings that were nonsignificant but had large effect sizes are also reported. Because this was an exploratory study seeking poten- tial differences, it was decided that limiting Type II errors was more important than limiting Type I errors. Consequently, no corrections for multiple tests were made when analyzing the data. Results Four parents returned Time 1 forms after the student Time 1 measures had been administered. Seven parents did not return Time 2 question- naires. All eligible students participated at Time 2. Between Time 1 and Time 2, four families moved out of the school district and thus their (Time 1 only) data were omitted from analyses. Data from 45 parents (38 mothers) and 45 children (28 boys) were analyzed. Student Results All student means and standard deviations are reported in Table 3. At Time 1, gifted students reported significantly higher General Academic subscale perceptions than did their nongifted peers (M = 4.31 and 3.93 respectively), z = -1.973, p < .05, ĝ = .86. There were no other statistically significant between-subjects student differences at Time 1. However, despite nonsignificant findings, the difference between gifted and nongifted students on scores at Time 1 for the Parent–Academic subscale had a large effect size (ĝ = .98). At Time 2, gifted students reported higher attitudes than nongifted students in the Gifted Program–Academic subscale (M = 4.00 and 3.33 respec- tively), z = -2.27, p < .05, ĝ = .99. There were no other statistically significant differences between gifted and nongifted students. At Time 2, gifted students reported statistically significant lower attitudes on the Parent–Academic subscale than they had at Time 1 (M = 4.81 and 4.33 for Times 1 and 2 respectively), z = -2.06, p < .05, ĝ = 1.93. Nongifted students reported lower attitudes on the Gifted Program–Academic subscale at Time 2 than they had at Time 1 (M = 4.25 and 3.33 for Time 1 and 2 respectively), t(33) = -6.14, p < .001, 134 Journal for the Education of the Gifted Table 3 Means (and Standard Deviations) for Student Attitude Subscales Subscale G Time 1 G Time 2 (n = 7) (n = 9) Gifted Program–Academic 4.48 (.60) 4.00 (.77)* General Academic 4.31 (.40)* 3.91 (.21)† Parent–Academic 4.81 (.18) 4.33 (.27)‡ Parent–Social 4.49 (.36) 4.12 (.31)† NG Time 1 NG Time 2 (n = 35) (n = 37) Gifted Program–Academic 4.25 (.80) 3.33 (.64)*** General Academic 3.93 (.44) 3.87 (.28) Parent–Academic 4.31 (.54) 4.40 (.55) Parent–Social 4.31 (.43) 4.20 (.59) Note: G = Gifted; NG = Nongifted. * p < .05 Between-subjects. † p < .10 Within-subjects change from Time 1 to Time 2. ‡ p < .05 Within-subjects change from Time 1 to Time 2. *** p < .001 Within-subjects change from Time 1 to Time 2. ĝ = 1.26. Also of potential interest, gifted students showed negative trends on the General Academic subscale; the effect size of the change was large (ĝ = 1.23). Parent Results All parent means and standard deviations are reported in Table 4. There were no statistically significant differences between the parent groups at Time 1. At Time 2, the parents of gifted students reported higher attitudes on the General Academic subscale than the parents of nongifted youth (M = 4.13 and 3.76 respectively), z = -2.27, p < .05, ĝ = 1.09. At Time 2, parents of nongifted students showed nega- tive trends on the Gifted Program–Academic (M = 3.83 and 3.53 for Times 1 and 2 respectively), t(29) = -2.0, p = .055, ĝ = .34, and General Academic subscales (M = 3.96 and 3.76 for Times 1 and 2 respectively), t(29) = -2.03, p = .052, ĝ = .57. Additionally, despite Attitudes and the Gifted Identification Process 135 Table 4 Means (and Standard Deviations) for Parent Attitude Subscales Subscale G Time 1 G Time 2 (n = 7) (n = 9) Gifted Program–Academic 3.39 (.73) 4.00 (.63) General Academic 4.00 (.30) 4.13 (.30)* Parent–Academic 4.46 (.35) 4.28 (.25) Parent–Social 4.08 (.24) 4.14 (.19) NG Time 1 NG Time 2 (n = 35) (n = 37) Gifted Program–Academic 3.83 (.88) 3.53 (.87)† General Academic 3.96 (.35) 3.76 (.34)† Parent–Academic 4.21 (.56) 4.20 (.57) Parent–Social 4.11 (.37) 3.97 (.36) Note: G = gifted; NG = Nongifted.* p < .05 Between-subjects. † p < .10 Within-subjects change from Time 1 to Time 2. being nonsignificant, the change reported by the parents of gifted students on the Gifted Program–Academic subscale had a large effect size (ĝ = .85). Discussion Previous research on the effects of giftedness typically focus solely on those participating in gifted programs. The current study sought to expand upon this by using a broader population (all those participating in the gifted identification process) and a broader time frame (prior to testing and after participation had begun). In general, this study found only cognitive performance differences between gifted and nongifted students prior to the gifted identification process. These findings sug- gest that attitudes of youth who will be labeled gifted are not neces- sarily separated from their nongifted peers by a gulf of (standard) deviations. However, participating in the gifted identification process