The Journal of Educational Research, 104:5470, 2011 Copyright C Taylor & Francis Group, LLC ISSN: 0022-0671 print

/ 1940-067 online DOI:10.1080/00220670903567372

The Journal of Educational Research

55

School Moves, Coping, and Achievement: Models of Possible Interactions

HELEN JOANNA BOON
James Cook University, Australia ABSTRACT. A sample of 1,050 regional Australian secondary students participated in a study investigating the relationship between mobility and academic achievement. Measures of mobility, academic achievement, suspensions, coping strategies, parental education, and family structure were used to test the hypothesis that academic coping strategies interact with mobility to negatively predict behavioral problems and academic achievement or, alternatively, to support academic achievement in mobile students. Analyses applied to a theoretical model indicated that positive coping strategies protect students from behavioral problems and lower academic achievement linked to mobility. The model accounted for 53% of the variance in achievement. Possible explanations for prior inconsistent ndings are suggested. Keywords: achievement, adolescents, behavior, coping, mobility, SEM, suspensions

obility and its effects on academic achievement have been the focus of many recent studies. In the present study I examined the links between school mobility, academic achievement, challenging behavior, and coping strategies in an Australian regional adolescent sample. More specically, I tested the hypothesis that the relationship between mobility and academic achievement is associated with coping strategies and mediated by suspensions. Australia and the United States have very mobile populations (Benson, Haycraft, Steyaert, & Weigel, 1979; Fields, 1997). Recent Australian research estimated 30% of families with children move at least once every three years (Department of Education Science and Training, & Department of Defence [DEST & DoD], 2002). Student mobility was dened in the DEST and DoD (2002) study as more than two moves in a period of three years (p. 1) and in other reports as a nonpromotional school change (Rumberger, 2003, p. 6), and is associated with a range of antecedents and outcomes. Students move schools for a number of reasons. In Australia, for example, children of Defence Force personnel move frequently, but this type of move is qualitatively different, in terms of support, compared with a move of relocation

for work or to seek work, and probably involves different family structures and attitudes (Benson et al., 1979). Divorce, family breakdown, redundancy, poverty, or lifestyle changes may also necessitate a school move (DEST & DoD, 2002). Similar mobility reasons have been noted in the United Kingdom (Dobson & Heathorne, 1999) and United States (Schafft, 2006; Ou & Reynolds, 2008), with mobility rates of 3050% cited in low-income urban U.S. families (Eckholm, 2008). In addition to socioeconomic and structural factors, school mobility is also linked to challenging behavior at school (Engec, 2006; Sorin & Iloste, 2006). Engec examined ofcial records for the mobility rates of three quarters of a million students in the United States and found that those students who were mobile were suspended two to three times as often as nonmobile students.1 Wood, Halfon, Scarlata, Newacheck, and Nessim (1993) reported that American children who moved frequently were 50100% more likely to experience a learning disorder, a delay in growth or development, or to have four or more behavioral problems than children who moved infrequently. Sorin and Iloste found one of the factors linked to mobility in Queensland is challenging behavior. Although it is still uncertain whether challenging behavior is a precursor or result of mobility, it is an indicator of adjustment to school (e.g., Jimerson, Egeland, Sroufe, & Carlson, 2000; McEvoy & Welker, 2000) and often precipitates suspensions (Jimerson et al., 2000). Mobility may be the result of behavior problems (Sorin & Iloste, 2006), with a change of school deemed to be a fresh start by parents and students. Alternatively, it may precede it and cause adjustment problems. A relationship between adversity in life circumstances and adolescent aggressive and antisocial behavior has empirical support from cross-sectional (Jackson & Warren, 2000; Rafnsson, Jonsson, & Windle, 2006) and longitudinal studies (Gest, Neeman, Hubbard, Masten, & Tellegen, 1993;
Address correspondence to Helen Joanna Boon, School of Education, James Cook University, Townsville, North Queensland, Australia, 4811. (E-mail: helen.boon@jcu.edu.au)

Hoffmann & Su, 1997; Kim, Conger, Elder, & Lorenz, 2003). Children of mobile families have been found to have more behavioral, emotional, and school problems than do other children. Simpson and Fowler (1994) analyzed behavioral and emotional variables and school functioning data for 10,362 American children in Grades 112. Mobile students were 2.3 times more likely to have emotional or behavioral problems and 2.2 times more likely to have received psychological help. The study did not, however, establish causality effects. Mobility is widely held to be disruptive to a students education either directly, by disrupting curriculum continuity or indirectly through social stress and engagement issues affecting the student (Ou & Reynolds, 2008). A negative link between academic achievement and mobility has been found in Australia both in literacy and numeracy domains (DEST & DoD, 2002) and in the United States, where mobility has been shown to negatively predict achievement and school completion (Ou & Reynolds, 2008). Establishing a causal relationship between mobility and educational attainment is difcult, however, due to co-occurring confounding variables. For example, mobile students are more likely to be poor, more likely to come from a single-parent home, and more likely to live in a household in which the householder is unemployed or failed to complete high school (Heinlein & Shinn, 2000). Nonetheless, studies have shown that even after controlling for prior attainment and demographic factors, mobility impacted achievement negatively in UK students (Strand & Demie, 2007; Temple & Reynolds, 1999). Yet, research examining mobility and academic outcomes has not consistently shown negative connections (Benson et al., 1979). When mobility is a choice rather than enforced, such as in military families or in cases of migration to another country, it is more closely linked with positive rather than negative outcomes (Buerkle & Christenson, 1999; Marchant & Medway, 1987). Children of mobile middle and high socioeconomic status (SES) families (Dobson & Henthorne, 1999), children with high IQs (Long, 1992), and children whose parents are university graduates (Long, 1975) rarely report negative effects. Buerkle and Christenson (1999) contended that it is not mobility per se that determines childrens outcomes, but rather the reasons for moving and the familys attitude to moving. Research on the impact of family relocation suggests that there are risks of impairment to childrens psychosocial adjustment mediated by the nature of the move (e.g., divorce, bereavement), negative parental attitudes to the move, poor pre-move adjustment, and the number of family moves (Humke & Schafer, 1995). Because the family provides the primary context in which children acquire ways of coping with stress (Compas, Worsham, & Ey, 1992), it seems likely that family appraisal of a move may inuence a childs coping response to the move. It is suggested that one of the ways school mobility is translated into achievement

outcomes is via student coping resources. If correct, this hypothesis would go some way to explain why moving schools is sometimes connected with negative school outcomes and sometimes not. The link between adaptive coping strategies and supportive parenting behavior has been demonstrated (Hardy, Power, & Jaedicke, 1993; Kliewer, Fearnow, & Miller, 1996; Oliva, Jim nez, & Parra, 2009). Researchers have also e demonstrated that active coping strategies, such as problem solving and seeking social support, are associated with more positive child outcomes than are avoidant strategies for example, wishful thinking, denial of the problem (Ayers, Sandler, West, & Roosa, 1996; Herman-Stahl, Stemmler, & Petersen, 1995; Lengua & Sandler, 1996). Coping resources appear to enhance functioning during and recovery from the experience of adversity and to support resilience against stressors (Fredrickson, 2001). Recent research has also linked adaptive coping to classroom engagement in adolescents (Reschly, Huebner, Appleton, & Antaramian, 2008). Coping Coping comprises a diverse number of cognitive, affective, and behavioral characteristics (Endler, Parker, & Summerfeldt, 1993; Oakland & Ostell, 1996) responding to contextual demands (Pearlin & Schooler, 1978; Schwartz, Neale, Marco, Shiffman, & Stone, 1999), and including cognitive appraisal (Folkman & Lazarus, 1980, 1985; Lazarus, 1999; McCrae, 1984). Folkman and Lazarus (1985) dened coping as a persons constantly changing cognitive and behavioral efforts to manage specic external and/or internal demands that are appraised as taxing or exceeding the resources of the person (p. 141). Lazarus and Folkman (1984) also posited that stress occurs when an individual perceives that he or she cannot manage the demands of a situation. It is therefore the appraisal of a potential stressor that determines the level of distress an individual experiences. They described two types of coping: problem-focused coping, aimed at actively attempting to alter the source of the stress, and emotion-focused coping, used to reduce and manage emotional distress associated with a situation. These two types of coping are usually deployed together to varying degrees (Lazarus, 2000). Coping strategies are used to alleviate negative emotions arising from, for example, failing an assessment, not completing school tasks, and social isolation (Folkman & Lazarus, 1988; Kamins & Dweck, 1999; Lazarus, 1993). A large Australian study showed adolescents who failed to cope were using inappropriate coping strategies (Lewis & Frydenberg, 2002). Frydenberg and Lewis (1994) found evidence that an adolescents choice of coping strategies is fairly consistent regardless of the nature of the concern, indicating perhaps familial or learned trends. A longitudinal study (Oliva et al., 2009) showed that high levels of family cohesion, communication, affection, and monitoring protected adolescents against externalizing behaviors subsequent to stressful

56

The Journal of Educational Research

The Journal of Educational Research

57

life events. Family processes then may provide a buffer for stressful events, such as moving school, perhaps by modeling adaptive coping. Conversely, low levels of adaptive coping lead to externalizing, challenging behavior (Rafnsson et al., 2006). Because mobility can be associated with a failure to meet school demands (DEST& DoD, 2002), adaptive coping strategies are important to help maintain a students engagement with school tasks. There appears to be a gap in the literature as far as mobility and academic coping is concerned. Tero and Connell (1984) developed the Academic Coping Inventory for use with school-aged students. They and others (Mantzicopoulos, 1997) showed that positive coping strategies were linked to higher achievement whereas projective, noncoping, and denial strategies correlated negatively with achievement. Persistence with demanding tasks and academic self-regulation, such as correcting mistakes and trying to understand academic material, is considered to be positive or adaptive problem-solving coping, as it leads to increased learning and better outcomes. By contrast, denying academic difculties, blaming others, including the teacher, or downplaying the importance of school assessment is characteristic of nonadaptive coping (Kaplan & Midgley, 1999). Positive coping strategies mediate positive classroom affect (Kaplan & Midgley) and are implicated in fostering resilience (Boon, 2008b; Howard & Johnson, 2000). Projective coping is linked to disruptive behavior (Friedel, Marachi, & Midgley, 2002). Study Aims The study aimed to shed light on the relationship between mobility, suspensions, coping strategies, and academic achievement in an Australian adolescent sample using structural equation modeling (SEM) techniques, in particular, to examine hitherto unexplored links between academic coping strategies, mobility and suspensions, while controlling for family structure and parental education. Family structure and parental education impact upon achievement and mobility, but have been inadequately controlled for by previous studies (DEST& DoD, 2002; Engac, 2006). Evidence also shows academic failure is linked to suspensions (Boon, 2008a; McEvoy & Welker, 2000). The extent to which suspensions are also signicantly linked with mobility in Australian students has not been fully explored. I hypothesized that one of the negative impacts of mobility on achievement is related to the students academic coping characteristics and mediated by behavioral characteristics, indexed by suspensions. Sudden or unexpected moves due to redundancy, bereavement, or divorce can be especially disruptive because of broken social ties and discontinuous school experiences, compounding the effects of a stressful home environment (Butler et al., 1990; Schafft, 2006). The ability to cope successfully with such stresses is likely to ameliorate their detrimental effects. However, coping strategies may be compromised because of longstanding difculties and SES disadvantage (DuBois, Felner, Meares, & Krier, 1994).

In other words, a history of difculties in an individuals family may diminish the capacity of an individuals adaptive coping or the parents capacity to demonstrate adaptive coping to their children, who in turn might model similar attitudes and behaviors. Buerkle and Christenson (1999) found a higher percentage of mobile families compared to nonmobile families cited a number of stressful life events. Similar claims have been found elsewhere (Mao, Whitsett, & Mellor 1998). Moreover, psychosocial adjustment measures linked to forced mobility have been shown to be long lasting (Braver, Ellman, & Fabricius, 2003). Linking these ndings is evidence that coping strategies employed by individuals are consistent regardless of the nature of the concern, indicating perhaps familial patterns (Frydenberg & Lewis, 1994). In other words, family type might impact upon adolescent coping styles. In all, I hypothesized that mobility exerts its negative effects on achievement directly because of curriculum and social disruption from one school to the next. Students who have been subject to long-standing stresses might succumb to this additional stress and display challenging behaviors resulting in suspensions, which then impact on academic achievement. This will constitute an indirect inuence of mobility on achievement. Other students who are moving for reasons deemed to be advantageous to the family circumstances, or students who are resilient, might employ positive academic coping to meet school demands. Because achievement is highly predicted by school engagement, coping strategies directed at academic problem solving should decrease the negative effects between achievement and mobility. School engagement, one of the strongest predictors of academic achievement and a protective factor against school dropout (Janosz, Archambault, Morizot, & Pagani, 2008) is characterized by behavioral (e.g., compliance, participation in school activities and extracurricular activities), affective (e.g., socioemotional interest in school), and cognitive (e.g., learning motivation and uses of self-regulatory coping strategies) dimensions (Fredricks, Blumenfeld, & Paris 2004), including problem-based coping strategies (Reschly et al., 2008). Adaptation to a new school involves dealing with new social and academic expectations requiring behavioral and cognitive responses. Depending on student internalizing processes and coping strategies these demands might not be adequately met (Compas, 1987). Student internalizing processes lead to adaptive or nonadaptive behavioral and academic coping, partly as a result of family (Kim & Kim, 2008) or school correlates (Marchand & Skinner, 2007; Newman, 2000). Academic coping, then, is an important and largely unexplored dimension that may link mobility to achievement through enhanced classroom engagement brought about by a proactive problem-solving response to classroom demands. The use of structural equation modeling (SEM) extends the understanding of how independent variables isolated in previous studies using multiple regression techniques

contribute to explain the variance of the dependent variable. Moreover, SEM models several multiple regression equations simultaneously, enabling the use of moderators and mediators as necessary (Byrne, 2001; Kline, 1998). When using a dichotomous variable as a mediator, maximum likelihood (ML) estimations cannot be used, so in this study polychoric correlations were employed to enable ML estimation to be used. The questions investigated in the study were the following: Research Question 1: Is mobility linked to suspension level in Australian students? Research Question 2: Is there a direct association between mobility and achievement or are mobility effects mediated by behavior problems (indexed by suspensions)? Research Question 3: Are coping strategies employed by mobile students whose academic achievement is satisfactory or better different from those employed by mobile students who are failing? Research Question 4: Do positive coping strategies act as a protective factor against suspension, promoting higher achievement? Research Question 5: Do mobility and coping strategies exert a separate measurable effect upon achievement and suspensions when sociodemographic variables are controlled? Finally, a theoretical model (Figure 1) based on prior research, tested the following hypotheses: Hypothesis 1 (H1 ): Mobility has direct and indirect effects on achievement outcomes and therefore mobility is included as a factor in the model to capture the indirect effects and H2 : Coping strategies employed by students impact upon achievement directly in the case of positive coping and indirectly, in the case of projective and noncoping via their links with challenging behavior (indexed by suspensions) after controlling for sociodemographic factors. The basis of the model stems from the hypothesis that mobility effects would be ameliorated if positive coping were employed. This would protect against challenging behavior and suspensions because stress would be resolved because engagement with school tasks would be higher. If the coping strategies employed are mainly noncoping and projective coping, stressors are less satisfactorily dealt with in context, leading to inappropriate behavior, less engagement time with scholastic tasks, possible withdrawal, and truancy. This would further exacerbate existing gaps in knowledge due to mobility, the net result being lower academic achievement and possibly higher suspension rates. Denial coping, characterized by tendencies for students to shrug off a negative event or say they do not care very much about it, is not thought to play a signicant role in predicting achievement or challenging behavior because the student reports indifference to the stress situation.

Method Sample The sample consisted of 1,127 Year 810 students, aged 1215 years, from three state high schools in a regional city in North Queensland, Australia. Participating schools were selected because they were located in separate geographical areas of the city serving as catchments for a range of families of different SES. Parents were sent letters informing them of the study and requesting parental permission for student participation. The resulting sample represented 81% of the students enrolled in the schools present on the days data were collected. Self-report questionnaires were completed during class period between 2 and 4 weeks after midyear report cards had been issued to students. In total, 1,050 complete questionnaires were obtained; the remaining 77 surveys were randomly missing various parts and were not included. I randomly checked 15% of the student responses for accuracy of grades, suspensions, and mobility with the participating schools as students supplied their names on the questionnaire. Measures Academic achievement. English and mathematics midyear grades were recorded as grades following the approach used by Paulson, Marchant, and Rothlisberg (1998). The grades were coded as the following: E = 0, D = 1, C = 2, B = 3, and A = 4, representing marks of up to 25% (E), 2549% (D), 5065% (C), 6680% (B), and over 80% (A), respectively. Grade C is the cutoff for a passing grade in the subject. An English or mathematics grade of less than 2 indicates that the student is failing the subject. Suspensions (Challenging behavior). Students reported the number of times they were suspended. These data were coded as the following: never suspended (0), suspended once (1), and suspended many times (2). For use in the SEM analyses, the suspension categories were collapsed into two: never suspended (0) and suspended (1). Academic Coping Inventor (ACI). The ACI, developed by Tero and Connell (1984), measures students self-reported coping strategies. This inventory includes four scales assessing the positive ( = .75), projective ( = .72), denial ( = .72), and noncoping ( = .77) strategies. All items begin with the stem, when something bad happens to me in school, and include the examples, such as not doing well on a test, or not being able to answer a question in class. Items on scales measuring coping strategies were responded to on a 4-point Likert-type scale ranging from 1 (not at all true) to 4 (very true) to conform to the procedure used by the developers of the scales. The ACI (12 items; see Appendix) was subjected to a conrmatory factor analysis (CFA) using

58

The Journal of Educational Research

The Journal of Educational Research

59

TABLE 1. Descriptive Statistics for Mobile and Nonmobile Students Nonmobile students Mobile students (n = 925; %) (n = 125; %) 2.2 7.8 42.6 33.7 13.7 4.1 13.3 30.9 24.6 27.0 71.7 28.3 76.6 23.4 36.9 63.1 84.8 15.2 12.0 26.4 31.2 24.8 5.6 10.4 29.6 36.8 14.4 8.8 80.8 19.2 84.8 15.2 58.4 41.6 62.4 37.6

Measure English grade E D Pass (C) B A Mathematics grade E D Pass (C) B A Fathers education Nongraduate Graduate Mothers education Nongraduate Graduate Family structure Nonintact family Biological intact family Overall suspension Never suspended Suspended

who passed (Table 2). A multivariate analysis of variance (MANOVA) was computed resulting in signicant main effects (based on Wilkss lambda criterion), F(4, 120) = 3.9, p <.005, p 2 = .12, indicated a large effect size. The differences between low and mediumhigh achieving students were due to differences in positive coping, F(1, 123) = 11.4, p <.001, p 2 = .09, a large effect size, and projective coping, F(1, 123) = 4.4, p < .05, p 2 = .05, a medium effect size, conrming initial hypotheses that coping factors are likely to be involved in the way mobility affects students achievement.

Model Testing SEM helps to overcome the problems associated with the effects of measurement error and correlated measurement error on the outcome variable (Baron & Kenny, 1986) which attenuate the estimation of relationships between observed variables (Kline, 1998, 2005; Maruyama, 1998). Moreover, SEM techniques developed over the last two decades permit the use of dichotomous categorical predictor variables in a model without the need to employ more complicated logistic regression models (J reskog & S rbom, 1984; Muth n, o o e 1984). Assumptions of multivariate normality need not apply to the exogenous (measured) variables (Bollen, 1989). However, the lack of multivariate normality usually inates the chi-square t statistic such that the possibility of the model being rejected due to Type I error is increased (Kline, 1998). AMOS 16 utilizes categorical data successfully when ML estimation is used and the sample size is large, both conditions being robust to violations of normality if this is not too extreme (Schermelleh-Engel, Moosbrugger, & M ller, u 2003). Raw data were checked for univariate normality and multivariate normality assumptions and were found to be satisfactory (kurtosis values < 2, critical ratios < 3; Kline, 2005). The theoretical model has ve dichotomous variables (suspensions, mobility, mothers education, family structure, and fathers education) and four latent variables (noncoping [3 indicator items], positive coping [3 indicator items],

FIGURE 1. Theoretical model (predictor variable covariances and error terms have been omitted for clarity).

the Amos 16.0 (Arbuckle, 2006) computer software program; it yielded good t indices after the CFA: root mean square error of approximation (RMSEA) = .049 (90% CI [0.041, 0.057]; p of close t (PCLOSE) = 0.564); 2/df = 3.52, p < .001; goodness of t index (GFI) = .976; adjusted goodness of t index (AGFI) = .959; comparative t index (CFI) = .964. Denition of mobility. The denition of mobility is problematic since different research studies have adopted different working denitions (DEST& DoD, 2002). Based on Dunn, Kadane, and Garrows (2003) research, which demonstrated a quantiable effect on achievement from a single school move, a mobile student in this study was a student who reported attending at least one different school since the beginning of the school year. Thus, respondents reporting that they had been in the school for less than one semester (Grade 8), less than 1 year (Grade 9) and less than 2 years (Grade 10) formed the mobile student category. This was a more stringent denition of mobility than that which many other studies have adopted (DEST & DoD, 2002). For use in the SEM analyses, mobility categories were coded into two: nonmobile (0) and mobile (1). Sociodemographic variables. The sociodemographic used were parental educational attainment (graduate or nongraduate status was coded as 1 for graduate and 0 for nongraduate for both mother and father) and family structure (biological

parents or intact family structure was coded as 1 and blended or other family structure was coded as 0). Results The SPSS 16 program was used to perform all statistical analyses, whereas SEM analyses were developed using the AMOS 16 program (Arbuckle, 2006). Table 1 presents summary statistics for mobile and nonmobile students, for academic grades, suspension level, parental education, and family structure. Results show that a larger proportion of mobile compared to nonmobile students achieved at a lower level, had more suspensions (37.6% compared to 15.2%), had parents who were not graduates (80.8% nongraduate fathers compared to 71.7%, 84.8% nongraduate mothers compared to 76.6%), and did not live with both biological parents (58.4% compared to 36.9%). An anticipated result was that mathematics is more severely affected by mobility than English achievement, probably due to the sequential nature of the curriculum, whereby discontinuity is likely to impose a more severe disadvantage on mathematics than, perhaps, on English. To assess whether mobility is linked to suspensions a 2 2 contingency table analysis was performed for suspensions and mobility. A statistically signicant relationship emerged, with a 2(1, N = 1050) = 37.5, p < .001. The proportion of variance in suspension associated with being mobile was 19%, showing a signicant link between mobility and higher levels of suspension, conrming other researchers claims (e.g., Sorin & Iloste, 2006). To investigate the hypothesis that mobile students who succeed academically adopt different coping strategies to those who do not, the mobile sample was split into two groups, those who achieved a passing grade (C) or better in both English and mathematics and those who did not. The coping strategies adopted by mobile students who failed both were signicantly different from those of students

TABLE 2. Means and Standard Deviations of Coping Strategies for Low- and MediumHigh-Achieving Mobile Students (N = 125) Coping Positive coping Projective coping Denial coping Noncoping Student group Mediumhigh achiever Low achiever Mediumhigh achiever Low achiever Mediumhigh achiever Low achiever Mediumhigh achiever Low achiever M 2.73 2.30 1.73 1.99 2.05 2.08 2.11 2.12 SD 0.74 0.66 0.72 0.68 0.69 0.70 0.87 0.74 n 68 57 68 57 68 57 68 57

60

The Journal of Educational Research

The Journal of Educational Research

61

1. The number of degrees of freedom of a model, as exacttting models are often highly parameterized (few degrees of freedom) with few opportunities for disconrmation, and thus scientically of little use (Markland, 2007). 2. Sole reliance on any one t test statistic as a means of assessing a model says nothing of the models ability to predict outcome variables (Gofn, 2007) because tests tend to support acceptance of a model based on the extent to which the estimated covariance matrix is consistent with the actual covariance matrix. 3. A golden rule of exclusively adopting the exact t statistic over a number of close t tests is unreasonable because it implies that the model would t perfectly in the population, an unreasonable assumption given the complexity of psychological processes (Bentler, 2007; MacCallum, 2003). 4. Model modication aids in SEM programs can inuence researchers to add freed parameters so that the model passes the exact t test, thus capitalizing on samplespecic variations (Gofn, 2007). As the debate continues, Breakwell, Hammond, FifeSchaw, and Smith (2006) proposed that judgment should be made in regards to all t statistics, close and exact, to accept or reject a model, taking care to interpret model t

FIGURE 2. Fitted theoretical Model 1 representing the relationship between coping strategies mobility, suspensions and academic achievement (N = 1,050) and showing total variances (R2). All parameter estimates shown are standardized.

denial coping [3 indicator items], projective coping [3 indicator items]), and achievement, comprising English grade (5 indicator categories) and mathematics grade (5 indicator categories). Consistent with standard SEM modeling conventions, unobserved or latent variables are depicted as ellipses in all models and show parceled factors. Moreover, due to the dichotomous nature of the mediating variable suspensions, polychoric correlations were used with maximum likelihood estimation (Kline, 2005). In determining whether the proposed model was a good representation of the data, or how well the model ts the data, information regarding the similarity between the estimated and actual variancecovariance matrices was calculated by the software. A range of t indices or test statistics is available for this estimation and more are being developed (Kline, 2005). In recent years a debate has arisen around the issue of the most appropriate discriminator of model t represented by the t indices. This centers on whether the exact t statistic (chi square [ 2]), which estimates the degree to which the data and the model-implied data are exactly the same, is the best arbiter of model t, or the close t statistics (e.g., CFI, GFI, RMSEA), which ask whether the t is close enough to suggest the model ts the data well. Passing the exact t statistic is more difcult than passing the close t statistics because of various issues (chi square analysis is very sensitive to many sources of error in model specication, to failures of distributional assumptions as well as to sample size issues), and therefore many theorists and practitioners have argued a model can still be accepted if it has passed the close t tests. Reasons proposed against sole reliance on a single t statistic (including the chi square) include the following:

holistically. They and others (Chen, Curran, Bollen, Kirby, & Paxton, 2008; Kline, 2005) have advised researchers to report the chi-square statistic or 2/df statistic so that those favoring the exact test statistic know if the model has passed that test. Because the chi-square t index is highly sensitive to sample size, overall model t was assessed by examining the CFI (Bentler, 1990) and the GFI, for which values of 0.90 or greater are considered acceptable, whereas values of 0.951.00 are considered good (Byrne, 2001; SchermellehEngel et al., 2003). The RMSEA was also examined, for which values of approximately .08 indicate acceptable model t and .05 or less indicate good model t in relation to the degrees of freedom (Schermelleh-Engel et al., 2003). Model 1 was tested using the full sample of 1,050 students (Figure 2). The tted model yielded the following: 2/df = 3.4, GFI = .964, CFI = .952, and RMSEA = .048 (90% CI [0.043, 0.053], PCLOSE = 0.746), all of which are strong indicators of good model t.3 Examination of all standardized regressions revealed the one between fathers education and suspensions and denial coping and suspensions was not signicant at p < .05. The low significance of the pathway between denial coping and suspensions conrmed hypotheses. In order to test if the t of the model would be signicantly improved by the removal of the pathway between fathers education and suspensions, a 2 test was performed. Results of trimming this pathway yielded the following t indices: 2/df = 3.37, GFI = .964, CFI = .953 and RMSEA = .048 (90% CI [0.042, 0.053], PCLOSE = 0.775), and the 2 test was not signicant, 2 = 2(115, N = 1050)(114, N = 1050) = 387.538387.530 = 0.008, p < ns. The pathway was not trimmed because this respecication was neither empirically or theoretically driven (Kline, 2005) and in other samples this pathway may be statistically signicant. Figure 2 shows that the model accounts for 53% of the variance (R2) in achievement and all standardized regression weights () were signicant at p < .001 except for fathers education to suspensions (p < .98) and denial to suspensions (p < .18). All covariances were signicant at p < . 005, except for the covariance between mothers education and noncoping, p < .90, fathers education and noncoping, p < .86, and mobility and denial coping, p < .74. Hence, students who could not cope had parents who were not graduates, and denial coping was not implicated with mobility. Total standardized effects from mobility to achievement were .29, comprising .07 standardized indirect effects through suspensions and .22 standardized direct effects. Thus, mobility was linked to achievement directly and indirectly via suspensions. It is important to note that suspensions are the mediator of both non- and projective coping to negatively predict achievement. The model is consistent with empirical data, showing mobility acts independently and in addition to demographic factors to negatively predict achievement. Positive coping positively predicted achievement directly and indirectly,

62

The Journal of Educational Research

The Journal of Educational Research

63

by suppressing suspensions. Both noncoping and projective coping predicted suspensions, consistent with them being considered nonadaptive strategies, whereas denial coping, as predicted, had no signicant associations with either suspensions or achievement. An intact family structure and mothers higher education negatively predicted suspensions and positively predicted achievement. Fathers education was positively linked to achievement, as would be expected because fathers education was coded 1 for graduate status. An alternative model, Model 1a, was also tested (Figure 3). In this model, denial coping and noncoping were linked directly to achievement, following the hypothesis that noncoping with academic tasks would result in poorer academic outcomes directly and denial coping would not motivate a student to attend to academic demands, also leading to poorer academic results. Standardized regression weights for denial to achievement ( = .04) and noncoping to achievement ( = .04) were not statistically signicant (p < .21 and p < .28, respectively). This model showed marginally worse t indices 2/df = 3.43, GFI = .963, CFI = .952, RMSEA = .048 (90% CI [0.043, 0.053], PCLOSE = 0.710). A rival model, Model 2 (Figure 4) was tested to investigate the hypothesis that mobility and particular demographic characteristics are mediated by coping strategies to achievement. This model had signicantly less adequate t indices, 2/df = 9.5, GFI = .90, CFI = .82, RMSEA = .10. Adding direct pathways from structural family factors and mobility to achievement and suspensions did not improve model t sufciently 2/df = 6.7, GFI = .928, CFI = .881, RMSEA = .07. It suggests coping strategies, developed over time, depend on more factors than mobility, SES, and family structure. For example, parenting style and maternal attri-

bution style, factors not measured in the present study, are likely to play a part in the development of adolescent coping strategies. Therefore, Model 1 is more consistent with hypotheses, which suggested that structural factors, mobility, and coping repertoires would predict achievement directly and indirectly through suspensions. In addition, a model equivalent to Model 1 was tested, Model 3 (Figure 5). This model had exactly the same t indices as the theoretical Model 1, 2/df = 3.4, GFI = .964, CFI = .952, RMSEA = .048 (90% CI [0.043, 0.053], PCLOSE = 0.742) and accounted for 50% of the variance in achievement, 32% of the variance in suspensions, and 25% of the variance in mobility. This model tested the idea that ineffective coping predicts mobility directly, or leads to suspensions that predict moving schools in search of a fresh start, a direct inuence, or indirectly via poor academic achievement. Standardized regression beta () weights were signicant at p < .005 except for the following: positive coping to mobility, p < .90, fathers education to mobility, p < .18, fathers education to suspensions, p < .96, mothers education to mobility, p < .70, projective coping to mobility, p < .40, and noncoping to mobility p <. 07. The model supports the idea that students who adopt projective and noncoping strategies move school because they either get suspended (total standardized effects suspension to mobility = 0.27, indirect standardized effects = 0.14) or move schools because of poor achievement outcomes (total direct effects, = .35). Denial coping predictably acts as a protective factor against mobility and suspensions in this model. The results from the model imply that those students who move schools do not live with both biological parents and that their parents are not university graduates. However, longitudinal data need to be obtained for this model of interactions to be conrmed, unlike Model 1, because the mobility factor, which is an outcome, was measured concurrently with all other measures. By contrast, academic achievement data, the outcome variable of Model 1, was measured just after students were sent their report cards from school, following the conceptual temporal structure of the model. The accounted variance of mobility in Model 3 was 25%, leaving 75% of the reasons for mobility unexplained. Given that moving school usually follows a family decision to relocate rather than an adolescents dissatisfaction with school, this model was less consistent with data measuring family relocation. Nonetheless, the model supported census data conrming those most likely to move are nongraduates (Australian Bureau of Statistics [ABS], 2008). Figure 6 shows the tted Model 3.

FIGURE 4. Alternative Model 2.

mobile students in order to clarify the puzzle of inconsistent prior results attendant on student mobility. Hypotheses of this study were that, after controlling for family structure and family SES, higher achieving mobile students would be using different academic coping strategies than lower achieving mobile students, and unproductive coping strategies would predispose students under the stress of moving schools to ex-

hibit behavioral problems, which in turn would negatively predict academic achievement. Structural equation models were used to test these hypotheses. Results show that mobile students have signicant differences in achievement and suspensions compared to nonmobile students and within the mobile group higher achieving students use more positive and less projective coping.

Discussion
FIGURE 3. Alternative Model 1a.

In the present study I compared the academic results, suspensions, and coping strategies of mobile and nonmobile students, and examined the within-group differences of

FIGURE 5. Equivalent Model 3.

64

The Journal of Educational Research

The Journal of Educational Research

65

Importantly, the ndings provide an explanation as to why some previous researchers have found negative links between mobility and achievement while others have not. The classication of students as mobile in this study has been very stringent, using a single nonpromotional move to categorize a student as mobile. Despite this rigorous selection, results show signicant differences in achievement between mobile and nonmobile students, demonstrating that mobility acts independently of demographic factors to negatively predict achievement. Where multiple school moves are investigated it is likely that these associations are even stronger. It was argued that a school move may result in different outcomes based on whether it was due to positive or negative family relocation reasons. By examining mobility correlates even without certainty as to the reason for a move an important task is performed because, whatever the reason for moving, schools need to provide academic and social support to these students to bridge academic gaps and assist with the social isolation students experience, for at least a period of time, following a move. A question addressed by the research was whether mobility is linked to suspensions. Results showed a strong association between mobility and suspensions, (standardized regression weights ( = .21) consistent with previous U.S. research (Engec, 2006). The direct positive link to suspensions (Model 1) might signal both academic and nonacademic adjustment issues faced by mobile students. For example, peers, teacher, or home stressors (DEST & DoD, 2002) could impact behavior at school and, ultimately, as another corollary, achievement. Model 3 offered an alternative mechanism connecting suspensions to mobility. Student behavior rather than family relocation was set as the prime cause of mobility. Behavior predicted mobility directly or indirectly through poor academic outcomes linked with suspensions. This fresh start model concedes the possibility that some within-student factors, learning disabilities, temperament variables, social isolation and so on, predict mobility. Assuming this temporal ordering of factors, it showed mobile students had nongraduate parents and nonintact families, conrming evidence found elsewhere (DEST & DoD, 2002). If correct, it shows one quarter of all school moves reported by this sample of 125 mobile students were caused by student variables. Doubt is cast on this model, however, by the very data that give it credence. According to labor statistics in Australia, those who are most likely to move are laborers and sales personnel, usually nongraduates (ABS, 2008). Therefore, it seems more likely that students move school as a result of family relocation. Alternatively, within-student factors may be coupled with, or resultant from, frequent family mobility to give rise to the ordering found in this fresh start model. Reciprocal relations between achievement and suspensions, and hence mobility, are also likely (Boon, 2008a; McEvoy & Welker, 2000), but they are not effectively estimated by nonrecursive structural equation models due to

technical analysis problems exacerbated by cross-sectional data (Kline, 2005). Disentangling the direction of these relations requires longitudinal data and multiple school hubs as aggregating units. Effects contingent on the context of schools and neighborhoods may support Model 3 over Model 1 because it is possible that the studentschool t is poor, does not support resilience, and gives rise to mobility. In other words, Model 3 gives preference to the notion that childrens achievement and coping strategies have a contextually situated nature rather than having a trait-like quality, as modeled by Model 1 and suggested by other researchers (Frydenberg & Lewis, 1994). It is of note that irrespective of the temporal ordering of variables, the positive role of adaptive coping in protecting against suspensions and promoting achievement is validated by both models. The second question addressed by the research contends whether mobility is responsible for a negative association with achievement directly or whether the effects of mobility were mediated by behavioral patterns, indexed by suspensions. Model 1 showed mobility exerts a smaller overall effect (regression weights, = .29) on achievement than suspensions (regression weights, = .33) although the suspension regression weight () included indirect mobility and coping strategy effects. Results support initial hypotheses that mobility effects on achievement are partly mediated by student behavioral factors and are consistent with Buerkle and Christensons (1999) claims that mobility is likely to be entangled with other long-standing problematic familial, socioeconomic, and adjustment issues, giving rise to nonadaptive coping. Where mobility is viewed as positive and is voluntary it is more likely that gaps in learning inevitably associated with changing schools or school systems (DEST& DoD, 2002) would be addressed by the student and supported by family or the organization instigating the move, for example, the Defence Force. Indeed, recent empirical data suggest coping style is inuenced by parents and teachers (Friedel, Cortina, Turner, & Midgley, 2007). Friedel et al. (2007) used the Academic Coping Inventory and found parents and teachers inuenced the childs adopted coping strategies, leading to achievement outcomes. For example, higher achievement is enhanced by engagement and perseverance with tasks, characteristic of positive coping (Reschly et al., 2008). Such might be the response of students who have moved with military parents or for job promotion, or parents who are graduates and active in supporting their childrens academic ambitions. Alternatively, adolescents who display nonadaptive behaviors may persist with nonadaptive coping because their coping strategies, established previously, have been consolidated through interactions with stressed, disadvantaged parents or similar-minded peers (Hartup & Van Lieshout, 1995). This may explain the responses of mobile students whose behaviors result in suspensions. Prior reports of inconsistent links between mobility and achievement are thus partly explained by the role played by adaptive coping strategies.

FIGURE 6. Fitted equivalent Model 3 representing a fresh start model of the relationship between coping strategies, suspensions, mobility and academic achievement (N = 1,050) and showing total variances (R2). All parameter estimates shown are standardized.

66

The Journal of Educational Research

The Journal of Educational Research

67

The hypothesized model, Model 1, shows a probable mechanism of interaction between coping strategies, suspensions, and academic achievement, accounting for a signicant 53% of the variance in achievement for the whole sample (N = 1,050). It highlights the importance of academic coping strategies for both mobile and nonmobile students. The model assumed adolescents employ a consistent repertoire of coping strategies developed over time prior to a school move as a result of modeling family interactions, reecting other researchers claims (Boekaerts, 2002; Frydenberg & Lewis, 1994; Hartup & Van Lieshout, 1995). The direct negative link between achievement and mobility most likely corresponds to the gaps or lags in learning associated with moving schools or school systems. The study results also suggest that mobility exerts a measurable negative effect on achievement even when parental education and family structure variables are controlled. Maternal education and family structure positively predicted achievement and also protected against suspensions. Findings illustrate adolescent psychosocial adjustment (characterized by positive adaptation to school rules and expectations and indexed by suspensions) is strongly inuenced by maternal input and a cohesive family unit. Models 1 and 3 support these patterns. A central question of the study concerned the role of positive coping as a protective factor. As was hypothesized, positive coping predicted higher achievement and lower suspensions, acting as protective mechanisms in mobile students. This may explain how some mobile students avoid the negative impact of mobility on achievement and behavior. A positive view of the school move, or coaching by teachers and parents to respond to academic challenges by adopting problem-solving strategies may be examples of how mobility effects can be counteracted by positive coping, reecting prior research (Friedel et al., 2007). A quick adjustment to the new school environment, indicating resilience, may also manifest in a positive coping focus. Students who view the move positively and have sufcient family support and strong academic motivational goals are also more likely to adopt positive coping strategies (Boekaerts, 2002) and adapt quickly to the new environment, minimizing the negative impact of mobility. When mobility is perceived as undesirable, or is frequent, projective and noncoping is thought to be reinforced predicting suspensions and lower achievement. Results show mobile students who were failing employed projective coping to a greater degree than positive coping strategies. It seems whatever the cause of mobility, if positive coping strategies are encouraged and adopted, challenging behavior and suspensions are not be as likely and higher achievement is promoted. A coping strategies index may be an important indicator of academic engagement for mobile adolescents. Directions for Future Researchers Further research into the correlates of mobility needs to include a number of additional factors to establish the mech-

anism with which mobility interacts with coping to affect achievement. These include clarication of the reason for the school move, whether it is for family relocation or a fresh start for the student; the number of moves a student has had; whether the move was within the same education district or from further aeld; whether the student receives economic support, where such policies are available, for example, a free lunch. Such data should enable the models proposed in the present study to be applied across groups of students differentiated by category of reason for moving. This should rene results reported here and conrm whether all students whose behavioral attributes predispose them to moving schools share coping strategy patterns. Additionally, it may help counselors identify students who may have particular needs. For example, internalizing students often go unnoticed because they are quiet and do not disturb the classroom environment. Internalizers who could be from disadvantaged or advantaged backgrounds may be managing to get by, but may do much better with support. Conversely, they may be failing and in need of support, but because they are not exhibiting challenging behaviors they do not attract the teachers attention or get effective support from home. Differences between boys and girls and different ethnic groups can also be examined by testing for invariance across groups using the proposed models. Longitudinal data dening suspensions as prior to or after a school move would assist schools and parents to apply interventions to support students appropriately. Hierarchical data using large school hubs to control for contextual variables would also help rene results. An important question to investigate is whether parents coping styles complement or correspond to their childrens coping styles. In some cases it may be ineffective coping due to personality problems that leads to parental mobility rather than structural or SES factors. Results from previous research suggest that parents coping is likely to be modeled by their children. If this is the case, then, it is critical to provide a student whose academic coping shows signs of ineffectiveness with alternative teacher and peer models to enable the student to improve his or her approach to academic tasks. If research highlights parentoffspring similarities, this approach needs to be applied as early in a childs school career as possible. Research in this area may illuminate how resilience development can be better understood and fostered. Although inuences upon coping strategies were not investigated here, it is known that adaptive coping strategies can be developed (Frydenberg & Lewis, 1993). Stress caused by a new environment varies depending on contextual factors, such as the goal structure of the school or classroom or the apparent support from parents and teachers. This being the case, interventions can be put into place at school to assist students to develop appropriate coping strategies which have been linked to better educational outcomes (Resnick et al., 1997). Parental support should be offered and sought upon moving to a new school in order to foster positive links with

the school and connectedness to the new environment. These are both known to foster better educational outcomes (Resnick et al., 1997). This is particularly important for single-parent families, disadvantaged families, and ethnic minority families (Schuler, 1990). Such interventions may go some way toward preventing a student from developing challenging behaviors or withdrawal in the form of absenteeism or truancy perhaps because of social isolation. Allan and Bardsley (1983) stated that many mobile children showed signs of unresolved stress in the form of bragging, aggression, rumination, or xations. By way of supporting these children they advised that they should be allowed to express their feelings and share their concerns in ways that relieve their stress but are not confronting to others. Although they were referring to primary-aged schoolchildren, similar support may benet secondary students, particularly in cases of an unplanned move that may give rise to depression, a prevalent mood disorder in adolescents (Barlow & Durand, 2005). Limitations The study ndings make an important contribution to the research relating to the effects of mobility, childrens achievement, and coping strategies. Nonetheless, there are some limitations to the study. Most important is the crosssectional nature of the research design, which precludes causality conrmation. Because of the reciprocal and contextual nature of human behavior, a causal ordering between constructs can only be supported by longitudinal evidence. The limited SES variables obtained, the unspecied reason for each school move and information on possible learning disabilities that students may have also limit the scope of the study. However, the results support the theoretical Model 1, accounting for 53% of the variance in achievement for 1,050 students, and demonstrate a plausible reason for inconsistencies reported in the past in connection with the academic achievement of mobile students. Further, Model 1 accounts for the trajectories of resilient mobile students who perhaps, despite family adversity, adopt positive academic coping and succeed academically. Conversely, Model 3 demonstrates the correlates and a mechanism whereby mobility may be deemed the only answer for students whose patterns of coping do not serve them to meet contextual school demands. Studies of longitudinal design are needed to establish the inuences that exist among these factors and the degree of reciprocity that may be present between adolescent characteristics and other factors connected to mobility. Mobility data delineates the number of moves each student has had and the timing of those moves, the reason for each move, and the students gender and cultural background. This is particularly important in places such as Australia, the United Kingdom, and the United States, where there are large numbers of immigrants. The coping strategies of mobile students whose moves are due to different causes (e.g., suspension and

a fresh start as opposed to a parental job promotion) would highlight how best to support and understand the needs of these vulnerable students.
NOTES
1. Comparable gures for Australian students are not widely available. When I tried to obtain suspension rates of secondary students from schools, the State Education Department forbade schools from releasing those records. 2. The CFA was performed using polychoric correlations following the advice of J reskog and S rbom (1996). o o 3. The polychoric covariance matrix on which this analysis is based is not reported in this article but may be obtained by contacting the author. REFERENCES Allan, J., & Bardsley, P. (1983). Transient children in the elementary school: A group counselling approach. Elementary School Guidance and Counselling, 17, 162169. Arbuckle, J. L. (2006). Amos 7.0 users guide. Chicago, IL: SPSS. Australian Bureau of Statistics. (2008, February). Labour mobility, Australia. Retrieved from http://www.abs.gov.au/AUSSTATS/ abs@.nsf/DetailsPage/6209.0Feb%202008?OpenDocument Ayers, T. S., Sandler, I. N., West, S. G., & Roosa, M. W. (1996). A dispositional and situational assessment of childrens coping: Testing alternative models of coping. Journal of Personality, 64, 923958. Barlow, D. H., & Durand, V. M. (2005). Abnormal psychology: An integrative approach (4th ed.). Belmont, CA: Thomson Wadsworth. Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 11731182. Benson, G. P., Haycraft, J. L., Steyaert, J. P., & Weigel, D. J. (1979). Mobility in sixth graders as related to achievement, adjustment, and socioeconomic status. Psychology in the Schools, 16, 444448. Bentler, P. M. (1990). Comparative t indices in structural models. Psychological Bulletin, 107, 238246. Bentler, P. M. (2007). On tests and indices for evaluating structural models. Personality and Individual Differences, 42, 825829. Boekaerts, M. (2002). Intensity of emotions, emotional regulation, and goal framing: how are they related to adolescents choice of coping strategies? Anxiety, Stress & Coping, 15, 401412. Bollen, K. A. (1989). Structural equations with latent variables. New York, NY: Wiley. Boon, H. J. (2008a). Risk or resilience? What makes a difference? Australian Educational Researcher, 31(1), 81102. Boon, H. J. (2008b). Students at risk: A bioecological investigation. Saarbrucken, Germany: VDM Verlag. Braver, S. L., Ellman, I. M., & Fabricius, W. V. (2003). Relocation of children after divorce and childrens best interests: New evidence and legal considerations. Journal of Family Psychology, 17, 206219. Breakwell, G. M., Hammond, S., Fife-Schaw, C., & Smith, J. A. (2006). Research methods in psychology (3rd ed.). London, UK: Sage. Buerkle, K., & Christenson, S. (1999). A family view of mobility among low-income children. CURA Reporter, 29, 712. Butler, D., Glasgow, D., Exelby, G., Fielding, G., Pattimore, L., & Caspani, P. (1990). Meeting the needs of mobile students in our community: An introduction. Townsville, Australia: Australian Education Council, Townsville & District Education Centre. Byrne, B. (2001). Structural equation modelling with AMOS. Mahwah, NJ: Erlbaum. Chen, F., Curran, P. J., Bollen, K., Kirby, J., & Paxton, P. (2008). An empirical evaluation of the use of xed cutoff points in RMSEA test statistic in structural equation models. Sociological Methods & Research, 36, 462494. Compas, B. E. (1987). Coping with stress during childhood and adolescence. Clinical Psychology Review, 7, 275302. Compas, B. E., Worsham, N. L., & Ey, S. (1992). Conceptual and developmental issues in childrens coping with stress. In A. M. LaGreca, L. J. Siegel, J. L. Wallander, & C. E. Walker (Eds.), Stress and coping in child health (pp. 724). New York: Guilford.

68
Department of Education Science and Training & Department of Defence. (2002). Changing schools: Its impact on student learning. Canberra, Australia: Commonwealth of Australia. Dobson, J., & Heathorne, K. (1999). Pupil mobility in schools: Interim report (DfEE Research Report 168). London: HMSO. DuBois, D. L, Felner, R. D., Meares, H., & Krier, K. (1994). Prospective investigation of the effects of socioeconomic disadvantage, life stress, and social support on early adolescent adjustment. Journal of Abnormal Psychology, 103, 511522. Dunn, M. C., Kadane, J. B., & Garrow, J. R. (2003). Comparing harm done by mobility and class absence: missing students and missing students and missing data. Journal of Educational and Behavioral Statistics, 28, 269 288. Eckholm, E. (2008, June 24). To avoid student turnover, parents get rent help. The New York Times. Retrieved from http://www.nytimes.com/ 2008/06/24/us/24move.html? r=1&ex=1214971200&en=fa9030b1e1b 1b140&ei=5070&emc=eta1&oref=slogin Endler, N. S., Parker, J. D. A., & Summerfeldt, L. J. (1993). Coping with health problems: Conceptual and methodological issues. Canadian Journal of Behavioral Science, 25, 384396. Engec, N. (2006). Relationship between mobility and student performance and behavior. The Journal of Educational Research, 99, 167178. Fields, B. (1997). The social and educational effects of student mobility: implications for teachers and guidance ofcers. Australian Journal of Guidance & Counselling, 7, 4556. Folkman, S., & Lararus. R. S. (1980). An analysis of coping in a middle-aged community sample. Journal of Health and Social Behavior, 21, 219239. Folkman, S., & Lazarus, R. S. (1985). If it changes it must be a process: A study of emotion and coping during three stages of a college examination. Journal of Personality and Social Psychology, 48, 150170. Folkman, S., & Lazarus, R. S. (1988). Coping as a mediator of emotion. Journal of Personality and Social Psychology, 54, 466475. Fredricks, J. A., Blumenfeld, P. C., & Paris, A. H. (2004). School engagement: Potential of the concept, state of the evidence. Review of Educational Research, 74, 59109. Fredrickson, B. L. (2001). The role of positive emotions in positive psychology: The broaden-and-build theory of positive emotions. American Psychologist, 58, 218226. Friedel, J. M., Cortina, K. S., Turner, J. C., & Midgley, C. (2007). Achievement goals, efcacy beliefs and coping strategies in mathematics: The roles of perceived parent and teacher goal emphases. Contemporary Educational Psychology, 32, 434458. Friedel, J. M., Marachi, R., & Midgley, C. (2002, April). Stop embarrassing me!: Relations among student perceptions of teachers, classroom goals, and maladaptive behaviors. Paper presented at the annual meeting of the American Educational Research Association, New Orleans, LA. Frydenberg, E., & Lewis, R. (1993). Adolescent Coping Scale. Melbourne, Australia: ACER. Frydenberg, E., & Lewis, R. (1994). Coping with different concerns: Consistency and variation in coping strategies used by adolescents. Australian Psychologist, 29, 4548. Gest, S. D., Neeman, J., Hubbard, J., Masten, A. S., & Tellegen, A. (1993). Parenting quality, adversity and conduct problems in adolescence: Testing process-oriented models of resilience. Development and Psychopathology, 5, 663682. Gofn, G. D. (2007). Assessing the adequacy of structural equation models: Golden rules and editorial policies. Personality and Individual Differences, 42, 831839. Hardy, D. F., Power, T. G., & Jaedicke, S. (1993). Examining the relation of parenting to childrens coping with everyday stress. Child Development, 64, 18291841. Hartup, W. W., & van Lieshout, C. F. M. (1995). Personality-development in social context. Annual Review of Psychology, 46, 655687. Heinlein, L. M., & Shinn, M. (2000). School mobility and student achievement in an urban setting. Psychology in the Schools, 37, 349357. Herman-Stahl, E. M., Stemmler, M., & Petersen, A. C. (1995). Approach and avoidant coping: Implications for adolescent mental health. Journal of Youth and Adolescence, 24, 649665. Hoffmann, J., & Su, S. (1997). The conditional effects of stress on delinquency and drug use: A strain theory assessment of sex differences. Journal of Research in Crime and Delinquency, 34, 4678. Howard, S., & Johnson, B. (2000). Young Adolescents displaying resilient and non-resilient behavior: Insights from a qualitative studyCan schools make a difference ? Retrieved from http://www.aare.edu.au/00pap/how00387.htm

The Journal of Educational Research

Humke, C., & Schafer, C. (1995). Relocation: A review of the effect of residential mobility on children and adolescents. Psychology, 32(1), 1624. Jackson, Y., & Warren, J. (2000). Appraisal, social support, and life events: Predicting outcome behavior in school-age children. Child Development, 71, 14411457. Janosz, M., Archambault, I, Morizot, J., & Pagani, L. S.(2008). School engagement trajectories and their differential predictive relations to dropout. Journal of Social Issues, 64, 2140. Jimerson, S., Egeland, B., Sroufe, L. A., & Carlson, B. (2000). A prospective longitudinal study of high school dropouts examining multiple predictors across development. Journal of School Psychology, 38, 525549. J reskog, K. G., & S rbom, D. (1984). LISREL VI: Analysis of linear struco o tural relationships by the method of maximum likelihood (3rd ed.). Mooresvile, IN: Scientic Software. J reskog, K. G., & S rbom, D. (1996). PRELIS 2: Users reference guide. o o Chicago, IL: SSI. Kamins, M., & Dweck, C. S. (1999). Person vs. process praise and criticism: Implications for contingent self-worth and coping. Developmental Psychology, 35, 835848. Kaplan, A., & Midgley, C. (1999). The relationship between perceptions of the classroom goal structure and early adolescents affect in school: The mediating role of coping strategies. Learning and Individual Differences, 11, 187212. Kim, K. J., Conger, R. D., Elder, G. H. Jr., & Lorenz, F. O. (2003). Reciprocal inuences between stressful life events and adolescent internalizing and externalizing problems. Child Development, 74, 127143. Kim, H. S., & Kim, H. S. (2008). Risk factors for suicide attempts among Korean adolescents. Child Psychiatry Human Development, 39, 221235. Kliewer, W., Fearnow, M. D., & Miller, P. A. (1996). Coping socialization in middle childhood: Tests of maternal inuence and paternal inuences. Child Development, 67, 23392357. Kline, R. B. (1998). Principles and practice of structural equation modelling. New York, NY: Guilford. Kline, R. B. (2005). Principles and practice of structural equation modelling (2nd ed.). New York, NY: Guilford Lazarus, R. S. (1993). From psychological stress to the emotions: a history of changing outlooks. Annual Review of Psychology, 44, 121. Lazarus. R. S. (1999). Stress and emotion: A new synthesis. New York, NY: Springer. Lazarus, R. S. (2000). Toward better research on stress and coping. American Psychologist, 55, 665673. Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York, NY: Springer. Lengua, L. J., & Sandler, I. N. (1996). Self-regulation as a moderator of the relation between coping and symptomatology in children of divorce. Journal of Abnormal Child Psychology, 24, 681701. Lewis, R., & Frydenberg, E. (2002). Concomitants of failure to cope: What we should teach adolescents about coping. British Journal of Educational Psychology, 72, 419431. Long, L. H. (1975). Does migration interfere with childrens progress in school? Sociology of Education, 48, 369381. Long, L. H. (1992). International perspectives on the residential mobility of Americas children. Journal of Marriage and the Family, 54, 861869. MacCallum, R. C. (2003). Working with imperfect models. Multivariate Behavioral Research, 38, 113139. Mantzicopoulos, P. (1997). How do children cope with school failure? A study of social/emotional factors related to childrens coping strategies. Psychology in the Schools, 34, 229237. Mao, M. X., Whitsett, M. D., & Mellor L. T. (1998). Student mobility, academic performance, and school accountability. ERS Spectrum, 16(1), 315. Marchand, G., & Skinner, E. A. (2007). Motivational dynamics of childrens academic help-seeking and concealment. Journal of Educational Psychology 99, 6582. Marchant, K. H., & Medway, F. J. (1987). Adjustment and achievement associated with mobility in military families. Psychology in the Schools, 24, 289294. Markland, D. (2007). The golden rule is that there are no golden rules: A commentary on Paul Barretts recommendations for reporting model t in structural equation modelling. Personality and Individual Differences, 42, 851858. Maruyama, G. M. (1998). Basics of structural equation modeling. Thousand Oaks, CA: Sage.

The Journal of Educational Research

McCrae, R. (1984). Situational determinants of coping responses: Loss, threat and challenge. Journal of Personality and Social Psychology, 46, 919928. McEvoy, A., & Welker, R. (2000). Antisocial behavior, academic failure, and school climate: a critical review. Journal of Emotional and Behavioral Disorders, 8, 130140. Muth n, B. O. (1984). A general structural equation model with dichotoe mous, ordered categorical and continuous latent indicators. Psychometrika, 49, 115132. Newman, R. S. (2000). Social inuences on the development of childrens adaptive help seeking: The role of parents, teachers, and peers. Developmental Review, 20, 350404. Oakland, S., & Ostell, A. (1996). Measuring coping: A review and critique. Human Relations, 49, 133155. Oliva, A., Jim nez, J. M., & Parra, A. (2009). Protective effect of supportive e family relationships and the inuence of stressful life events on adolescent adjustment. Anxiety, Stress & Coping, 22, 137152. Ou, R.-S., & Reynolds, A. J. (2008). Predictors of educational attainment in the Chicago longitudinal study. School Psychology Quarterly, 23, 199229. Paulson, S. E., Marchant, G. J., & Rothlisberg, B. A. (1998). Early adolescents perceptions of patterns of parenting, teaching and school atmosphere: Implications for achievement. Journal of Early Adolescence, 18, 526. Pearlin, L. I., & Schooler, C. (1978). The structure of coping. Journal of Health and Social Behavior, 19, 221. Rafnsson, F. D., Jonsson, F. H., & Windle, M. (2006). Coping strategies, stressful life events, problem behaviors, and depressed affect among Icelandic adolescents: A cross-cultural replication study. Anxiety, Stress and Coping, 19, 241257. Reschly, A. M, Huebner, E. S., Appleton, J. J., & Antaramian, S. (2008). Engagement as ourishing: the contribution of positive emotions and coping to adolescents engagement at school and with learning. Psychology in the Schools, 45, 419431. Resnick, M. D., Bearman, P. S., Blum, R. W., Bauman, K. E., Harris, K. M., Jones, J., . . . Udry, R. (1997). Protecting adolescents from harm: Findings from the national longitudinal study on adolescent health. Journal of the American Medical Association, 278, 823832. Rumberger, R. (2003). The causes and consequences of student mobility. Journal of Negro Education, 72(1), 621.

69
Schafft, K. A. (2006). Poverty, residential mobility, and student transiency within a rural New York. Rural Sociology, 71, 212231. Schermelleh-Engell, K, Moosbrugger, H., & M ller, H. (2003). Evaluating u the t of structural equation models: Tests of signicance and descriptive goodness-of-t measures. Methods of Psychological Research Online, 8(2), 2374. Schuler, D. B. (1990). Effects of family mobility on student achievement. ERS Spectrum, 8, 724. Schwarz, J. E., Neale. J., Marco, C., Shiffman, S. S., & Stone, A. A. (1999). Does trait coping exist? A momentary assessment approach to the evaluation of traits. Journal of Personality and Social Psychology, 77, 360369. Simpson, G. A., & Fowler, M. G. (1994). Geographic mobility and childrens emotional/ behavioral adjustment and school functioning. Paediatrics, 93, 303310. Sorin, R., & Iloste, R. (2006). Moving schools: Antecedents, impact on students and interventions. Australian Journal of Education, 50, 227241. Strand, S., & Demie, F. (2007). Pupil mobility, attainment and progress in secondary school, Educational Studies, 33, 313331. Temple, J. A., & Reynolds, A. J. (1999). School mobility and achievement: longitudinal ndings from an urban cohort. Journal of School Psychology, 37, 355377. Tero, P. F., & Connell, J. P. (1984). When children think theyve failed: An academic coping inventory. Unpublished manuscript. Wood, D., Halfon, N., Scarlata, D., Newacheck, P., & Nessim, S. (1993). Impact of family relocation on childrens growth, development, school function and behavior. Journal of the American Medical Association, 270, 13341338.

AUTHOR NOTE Helen Joanna Boon teaches educational psychology and research methodology at James Cook University in Australia. Her current research interests include adolescent and community resilience, as well as issues relevant to science education, teacher training, and ethics.

70

The Journal of Educational Research

Copyright of Journal of Educational Research is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.

APPENDIX The Academic Coping Inventory 1. When something bad happens to me in school, I try to gure out what I did wrong so that it wont happen again. 2. When something bad happens to me in school, I say that the teacher didnt cover the things on the test. 3. When something bad happens to me in school, I say it was the teachers fault. 4. When something bad happens to me in school, I tell myself it didnt matter. 5. When something bad happens to me in school, I get angry at the teacher. 6. When something bad happens to me in school, I can tell myself Ill do better next time. 7. When something bad happens to me in school, I worry that other students will think that Im dumb. 8. When something bad happens to me in school, I say I didnt care about it 9. When something bad happens to me in school, I get really mad at myself. 10. When something bad happens to me in school, I feel really stupid. 11. When something bad happens to me in school, I try to see what I did wrong. 12. When something bad happens to me in school, I say it wasnt important