Sei sulla pagina 1di 16

Journal of Educational Psychology 2010, Vol. 102, No.

3, 741756

2010 American Psychological Association 0022-0663/10/$12.00 DOI: 10.1037/a0019237

Effects on Teachers Self-Efficacy and Job Satisfaction: Teacher Gender, Years of Experience, and Job Stress
Robert M. Klassen
University of Alberta State University of New York at Buffalo

Ming Ming Chiu

The authors of this study sought to examine the relationships among teachers years of experience, teacher characteristics (gender and teaching level), three domains of self-efficacy (instructional strategies, classroom management, and student engagement), two types of job stress (workload and classroom stress), and job satisfaction with a sample of 1,430 practicing teachers using factor analysis, item response modeling, systems of equations, and a structural equation model. Teachers years of experience showed nonlinear relationships with all three self-efficacy factors, increasing from early career to mid-career and then falling afterwards. Female teachers had greater workload stress, greater classroom stress from student behaviors, and lower classroom management self-efficacy. Teachers with greater workload stress had greater classroom management self-efficacy, whereas teachers with greater classroom stress had lower self-efficacy and lower job satisfaction. Those teaching young children (in elementary grades and kindergarten) had higher levels of self-efficacy for classroom management and student engagement. Lastly, teachers with greater classroom management self-efficacy or greater instructional strategies self-efficacy had greater job satisfaction. Keywords: self-efficacy, teachers, job satisfaction, motivation

An emerging body of research shows that teachers selfefficacythe beliefs teachers hold about their capability to influence student learningis associated with student factors, like achievement and motivation (Caprara, Barbaranelli, Steca, & Malone, 2006), as well as teacher factors, like job commitment and job satisfaction (e.g., Caprara, Barbaranelli, Borgogni, & Steca, 2003). In spite of the evident association between teachers selfefficacy and student and teacher outcomes, little is known about how self-efficacy and job stress are related to teachers job satisfaction or how teachers self-efficacy is related to years of experience. Teachers self-efficacy is believed to be most malleable in the challenging early stage of a teachers career and then to increase and become more firmly established as teachers gain experience (Tschannen-Moran & Woolfolk Hoy, 2007; Wolters & Daugherty, 2007). However, middle and late career stages bring their own challenges that can influence motivation and job satisfaction (e.g., Kooij, de Lange, Jansen, & Dikkers, 2008; Spickard, Gabbe, & Christensen, 2002). In this study, we use advanced modeling techniques (factor analyses, item response models, systems of equations, and structural equation models) to extend our understanding of teachers motivation beliefs. First, we hypothe-

size a model of the relationships among self-efficacy, overall perceived job stress, stress from classroom and workload factors, teacher characteristics (gender and teaching level), and job satisfaction. Second, we examine patterns of self-efficacy beliefs of 1,430 teachers with varying years of experience.

Teachers Self-Efficacy
Self-efficacy refers to individuals beliefs about their capabilities to carry out a particular course of action successfully (Bandura, 1997). Extensive research supports the claim that selfefficacy is an important influence on human achievement in a variety of settings, including education, health, sports, and business (Bandura, 1997). In educational research, the self-efficacy beliefs of students have been shown to play an important role in influencing achievement and behavior. Furthermore, researchers are finding that teachers self-efficacy influences their teaching behaviors and their students motivation and achievement (Skaalvik & Skaalvik, 2007; Tschannen-Moran & Woolfolk Hoy, 2001). Teachers with low self-efficacy experience greater difficulties in teaching, higher levels of job-related stress (Betoret, 2006), and lower levels of job satisfaction (Klassen et al., 2009). Self-efficacy researchers agree that teachers self-efficacy should be operationalized to reflect beliefs about capability and therefore should be phrased in terms of can do rather than will do. Can is a judgment of capability; will is a statement of intention (Bandura, 2006, p. 308, italics in original; also see Bong, 2006). In addition, self-efficacy measures should reflect a particular context or domain of functioning, rather than global functioning (Bandura, 1997). A global measure of teachers self-efficacy might ask, How confident are you in your teaching ability? whereas a domain-focused measure would inquire about teachers confidence to accomplish particular tasks. Tschannen-Moran and Hoy (2001)
741

Robert M. Klassen, Department of Educational Psychology, University of Alberta, Edmonton, Alberta, Canada; Ming Ming Chiu, Department of Learning and Instruction, Graduate School of Education, State University of New York at Buffalo. The authors gratefully acknowledge funding support to Robert M. Klassen from the Social Sciences and Humanities Research Council of Canada. Correspondence concerning this article should be addressed to Robert Klassen, Department of Educational Psychology, University of Alberta, Edmonton, Alberta T6G 2G5, Canada. E-mail: robert.klassen@ualberta.ca

742

KLASSEN AND CHIU

created a teachers self-efficacy measure with item stems of How much can you do . . .? that explored teachers beliefs about their capabilities in three key classroom domains: implementing instructional strategies, managing student behaviors, and engaging students in the learning process. By including items from these three critical areas, and by situating the three areas in teachers classrooms, the authors balanced the demands for specificity (i.e., self-efficacy assessments that reflect particular tasks) and practical usefulness (i.e., multifaceted measurement that is not microscopically operationalized [Pajares, 1996, p. 562]) in a meaningful context (i.e., teachers classrooms). Although earlier teachers self-efficacy measures were marred by faulty conceptualization, including a focus on ability, not capability, and a focus on external influences, not internal beliefs (see Tschannen-Moran, Woolfolk Hoy, & Hoy, 1998), more recent measures such as TschannenMoran and Woolfolk Hoys (2001) Teachers Self-Efficacy Scale adhere more closely to the theoretical guidelines proposed by Bandura (1997, 2006), specifically in the focus on forward-looking capabilities (e.g., I can craft good questions for students) and not global ability (e.g., I am a good teacher). Although Bandura (1997) hypothesized that self-efficacy beliefs remain relatively stable once established, researchers have noted that little evidence exists about how (teachers) efficacy beliefs change or solidify across stages of a career (Tschannen-Moran et al., 1998, p. 238). A few studies have been conducted on the relationship between teaching experience and teachers selfefficacy, yielding varied results. Ross, Cousins, and Gadalla (1996) found mixed support for the influence of experience on teachers self-efficacy, and Ghaith and Yaghi (1997) found negative correlations between years of experience and teacher selfefficacy, with both studies using modest-sized samples (52 and 25, respectively). Woolfolk Hoy and Burke Spero (2005) conducted a longitudinal study in which they collected data from teachers at two points during their teacher-training program and at the end of their first year of teaching. Results showed a significant rise in teachers self-efficacy during teacher training, followed by a decline at the end of their first teaching year, but once again, the research was hampered by a modest sample of 29 teachers. A recent study by Wolters and Daugherty (2007) used a large online sample of teachers (N 1,024) from the United States to examine the influence of teaching experience on teachers selfefficacy and goal structures. Teachers were divided into four experience groups: 1 year, 15 years, 6 10 years, and 11 years of experience. Then, they completed the three-factor TschannenMoran and Woolfolk Hoy (2001) Teachers Self-Efficacy Scale as well as a measure of goal structures. Results showed modest effects of experience on self-efficacy for instructional strategies (2 .04) and self-efficacy for classroom management (2 .02), but no effect of experience on self-efficacy for student engagement. Although the researchers have made an important contribution by linking experience with teachers self-efficacy, their findings paint an incomplete picture, with two potential limitations. One problem is that the relationship between teachers self-efficacy and experience may not be linear. For example, Woolfolk Hoy and Burke Spero (2005) found that teachers selfefficacy initially rose and then fell over three data collection points at the beginning of teachers careers. Another problem is that teachers with more than 10 years of experience were treated as a single group. The authors acknowledged that the lack of differ-

entiation among the most-experienced teachers may mask changes in teachers self-efficacy that may occur toward the end of their careers (p. 189). In fact, most teachers have more than 10 years of experience: recent statistics show that American teachers have an average of about 14 years of experience, and 60% of teachers have 10 or more years of experience (U.S. Department of Education, 2009). Thus, additional research on how experience affects teachers self-efficacy across the career span is needed. The developmental course of occupational self-efficacy is not uniform from early to late adulthood, and teachers self-efficacy may ebb and flow over the course of a career as it is influenced by life and career events and challenges. Bandura (1997) suggested that some workers at mid-to-late career stages may restructure or scale down overambitious goals due to waning self-efficacy, although this experience is not universal. Workplace environments influence self-efficacy beliefs, with supervisors verbal persuasion and modeling serving as important spurs to workers self-efficacy development (Bandura, 1997). Kooij et al. (2008) suggested that age-related physical and psychological factors can influence work motivation, but workplace factors can mediate how age-related concerns are interpreted. For teachers, the combination of successful past experience; verbal support from principals, students, peers, and parents; and opportunities for observation of successful peers builds self-efficacy for teaching (Tschannen-Moran et al., 1998). The influence of the sources of self-efficacy, however, may change over time, with verbal persuasion and contextual factors playing a more important role for novice teachers than for veteran teachers (Tschannen-Moran & Woolfolk Hoy, 2007). Self-efficacy beliefs in the workplace are not static and reflect a lifelong process of development that ebb and flow according to personal attributes and interpretation of environmental circumstances.

Job Satisfaction and Job Stress


Despite reports of high levels of teachers job stress (Chaplain, 2008; Schwarzer & Hallum, 2008), many teachers find personal satisfaction in their work. Job satisfactionperceptions of fulfillment derived from day-to-day work activitiesis associated with higher levels of job performance (Judge, Thoresen, Bono, & Patton, 2001). Caprara et al. (2003) considered job satisfaction a decisive element (p. 823) influencing teachers attitudes and performance and found self-efficacy to be an important contributor to teachers job satisfaction. Teachers report that job satisfaction is gained from the nature of day-to-day classroom activities, such as working with children, seeing students make progress, working with supportive colleagues, and overall school climate (Cockburn & Haydn, 2004). Teachers who are dissatisfied with their work display lower commitment and are at greater risk for leaving the profession (Evans, 2001; Ingersoll, 2001). Liu and Ramsey (2008) found that stress from poor work conditions had the strongest influence on teachers job satisfaction and noted that inadequate time for planning and preparation and a heavy teaching workload reduced satisfaction from teaching. Teaching may bring personal satisfaction, but it also brings stress, with demands from administrators, colleagues, students, and parents compounded by work overload, student misbehavior, and a lack of recognition for accomplishments (Greenglass & Burke, 2003). Teachers with greater teacher stress defined as the experience of negative emotions resulting from a teachers work

EFFECTS ON TEACHERS SELF-EFFICACY

743

(Kyriacou, 2001) have lower self-efficacy (Betoret, 2006; Schwarzer & Hallum, 2008; Skaalvik & Skaalvik, 2007), poorer teacherpupil rapport, and lower levels of effectiveness (Abel & Sewell, 1999; Kokkinos, 2007). Teachers with high levels of job stress may gain satisfaction from work, but the level of satisfaction may be muted by stress from role ambiguity, low autonomy, or frequency or level of conflict with students and colleagues (Greenglass & Burke, 2003). Teaching has been listed among the highstress professions, with as many as one-quarter of teachers reporting that teaching is a very stressful job (Kyriacou, 2001). Whereas previous studies have conceptualized teachers job stress as a unidimensional construct (e.g. Schwarzer & Hallum, 2008), other studies have shown that workload and student misbehavior (i.e., classroom factors) contribute separately to teachers overall stress (Boyle, Borg, Falzon, & Baglioni, 1995). Teachers with high levels of stress from these two sources show higher negative health and vocational outcomes, including burnout (emotional exhaustion, depersonalization, and reduced personal accomplishment), absenteeism, and exit from the teaching profession (Betoret, 2006; Jepson & Forrest, 2006; Kyriacou, 2001).

level, teaching experience, gender, and demographic factors like teachers cultural or national background.

Current Study
Although researchers have begun to examine teacher motivation by studying self-efficacy, job stress, and job satisfaction, few have proposed explanatory models that take into account teacher characteristics such as years of experience, teaching level, and gender. In the current study, we propose and test a model that accounts for these contextual factors and also includes control variables for teachers ethnic heritage and grades taught within schools (see Figure 1). Two research questions are addressed. First, how is teachers self-efficacy related to years of experience? We predicted that teachers self-efficacy would increase in early to midcareer (e.g., Wolters & Daugherty, 2007) but show declines in the late career stage, as has been found in previous research conducted outside educational settings (Kooij et al., 2008). Second, what are the relationships among teachers self-efficacy, job stress (overall stress and sources of stress), job satisfaction, and contextual factors (teacher characteristics and school level)? We hypothesized that teachers self-efficacy would be influenced by teachers stress (Skaalvik & Skaalvik, 2007) and that teachers job satisfaction would be influenced by teacher characteristics, teacher stress and its two sources, and three domains of teachers self-efficacy (e.g., Caprara et al., 2003; Liu & Ramsey, 2008).

Teacher Characteristics
Teaching level and teacher gender are related to teachers jobrelated beliefs. Elementary school teachers report higher levels of self-efficacy for student engagement than teachers in middle or high schools (Wolters & Daugherty, 2007). Liu and Ramsey (2008) found that women experience less job satisfaction than men, especially satisfaction from work conditions, and a number of researchers have noted that female teachers report higher stress than male teachers (e.g., Antoniou, Polychroni, & Vlachakis, 2006; Chaplain, 2008), possibly due to higher levels of overall workload (Greenglass & Burke, 2003). Whereas Klassen et al. (2009) found similar relationships between self-efficacy and job satisfaction for teachers from five North American and Asian countries, results from other studies suggest that teachers nationality and associated cultural beliefs can influence the relationships among job stress, job satisfaction, and teachers efficacy (Klassen, Usher, & Bong, in press; Liu & Ramsey, 2008). Models explaining teacher motivation must account not only for individual beliefs and motivation but also for teacher characteristics such as teaching

Method Participants
The participants were part of a convenience sample of 1,430 practicing teachers (69% women, 31% men) from western Canada. Teachers reported working in elementary schools (20%; usually Grades K 6), junior high schools (6%; usually Grades 79), high schools (9%; usually Grades 10 12), elementaryjunior high schools (13%; usually Grades K9), and junior highsenior high schools (12%; usually Grades 712), as well as other combinations in a mix of urban (38%), suburban (11%), rural (28%), and other or not reported (23%) settings. Teachers reported their ethnic heritage as AngloEuropean Canadian (92%), Asian Canadian (2%), First

Demographics School Range of grade levels Teacher Gender Country of birth Nationality Teacher Experience in School Years of experience Years in current school Range of grade levels taught by the teacher in this school Teacher Self-efficacy Classroom management Instructional Strategies Student Engagement Teacher Stress Classroom stress Workload stress Overall stress

Teacher Job Satisfaction

Figure 1.

Model of hypothetical relationships.

744

KLASSEN AND CHIU

Nations or Aboriginal Canadians (1%), African (1%), or South American (1%). Teachers had a mean age of 40.00 years (SD 10.79) and an average of 13.21 years of teaching experience (SD 13.97). The age and experience of teachers in this sample are consistent with provincial government data showing the median age of teachers at 40 44 years and median years of experience of 10 14 years (Alberta Education, 2009).

Procedure
Participants were attendees at one of several annual, compulsory, multidistrict teacher conferences, the total attendance of which was approximately 8,000 teachers from about 350 schools. Teachers were approached by one of a team of researchers in an exhibit hall and asked to complete a brief questionnaire titled What Motivates Teachers? Approximately 2,000 teachers were approached, and approximately 75% of the teachers completed the survey. Participants were asked to read the instructions and participate only if they were currently teaching in schools (i.e., not serving as administrators or counselors). Conference organizers did not permit researchers to request school identities from participants. (See previous studies for details regarding the procedure and measures [e.g., Klassen et al., 2009]).

Variables
The survey included (a) a front sheet describing the project and contact information for the lead researcher and the university ethics board, (b) a demographics section, and (c) four measures: a 12item teachers self-efficacy scale, a two-item job satisfaction scale, one item measuring overall job stress, and seven items measuring sources of job stress (see Appendix Table A8 for the survey items). Teachers self-efficacy. Tschannen-Moran and Woolfolk Hoy (2001) created and validated the Teachers Self-Efficacy Scale (TSES). Because it closely aligns with self-efficacy theory, TSES is superior to previous measures of teacher efficacy (Woolfolk Hoy & Burke Spero, 2005, p. 354). Researchers have investigated the TSES short- and long-form measures in a variety of settings and have found adequate reliability and validity for the whole scales and their three subscales: self-efficacy for classroom management, instructional strategies, and student engagement. For example, Klassen et al. (2009) found reliabilities ranged from .71 to .94 for TSES short-form subscales in five countries and significant relationships between the TSES subscales and job satisfaction in all settings. Wolters and Daugherty (2007) reported Cronbachs alpha coefficients above .80 for the TSES. The TSES long and short forms are hypothesized to consist of three factors that measure a teachers confidence to manage student behavior in the classroom (e.g., How much can you do to control disruptive behavior in the classroom?), to use effective instructional strategies (e.g., How much can you do to craft good questions for students?), and to engage all students in learning (e.g., How much can you do to motivate students who show low interest in school work?). These items show fidelity with self-efficacy theory because they measure teachers beliefs in their capabilities to carry out particular tasks (e.g., provide an alternative explanation when students are confused) in a particular context (i.e., the classroom). Participants in our study responded to the 12-item

TSES short form with a 9-point response scale, anchored by 1 (nothing) and 9 (a great deal). Job satisfaction and job stress. Job satisfaction was measured with two items from Caprara et al. (2003) on a 9-point scale. Items consisted of (a) I am satisfied with what I achieve at work, and (b) I feel good at work. The measure showed adequate reliability and validity in Caparara et al.s 2003 study and has been shown to be related to self-efficacy in previous studies (e.g., Klassen et al., 2009). Job stress was measured in two ways. First, following the approach used in recent studies of teacher stress (e.g., Boyle et al., 1995; Chaplain, 2008; Manthei, Gilmore, Tuck, & Adair, 1996), we measured overall job stress with a single item (I find teaching to be very stressful). Next, we used six items from Boyle et al.s (1995) Teacher Stress Inventory plus an additional item, class size, suggested from recent teacher stress research (Gates, 2007), to assess two major contributors of teaching stressworkload stress and classroom stress from student behavior. Boyle et al. found acceptable levels of reliability and validity in their 1995 study of teachers in the United Kingdom, and Klassen (in press) found the that workload stress and stress from student behavior were inversely related to job satisfaction. The job stress items were presented with the stem, As a teacher, how great a source of stress are these factors to you? with responses ranging from 1 (no stress) to 9 (extreme stress). Items representing sources of workload stress included too much work to do, having extra duties/responsibilities because of absent teachers, large class size, and responsibility for student achievement. Items representing classroom stress from student behavior included maintaining class discipline, impolite behavior and rudeness, and noisy students.

Analysis
We tested the internal validity of the questionnaire items for each teacher characteristic with factor analyses and minimized their measurement errors with item response models. To account for heteroskedasticity and contemporaneous correlation in the errors across equations with multiple outcome variables, we modeled teachers self-efficacy with a system of equations. Then, we estimated the association of job satisfaction with teacher self-efficacy and other variables with an ordinary least squares (OLS) regression. Lastly, we captured all these relationships simultaneously with a structural equation model (SEM). Factor analyses and item response models. We used factor analysis with varimax rotation to test the internal structure of participant responses to sets of questions regarding teachers selfefficacy, specifically whether they reflected (a) a single factor, (b) separate factors, (c) hierarchical factors, (d) nested factors, or (e) no factorsno valid construct(s) (Chow, Chiu, & Wong, in press; Gustafsson & Balke, 1993). Using Monte Carlo simulation studies, Hu and Bentler (1999) showed that a combination of the standardized root-mean-square residual (SRMR) and one of the following indices tends to minimize Type I and Type II errors under many conditions for both factor analyses and SEMs: TuckerLewis Index (TLI), incremental fit index (IFI), and root-mean-square error of approximation (RMSEA). We used the following threshold values to separate good, moderate, and poor fits for each measure: for SRMR, between .08 and .10 (good fit if less than .08; moderate fit if between .08 and .10; poor fit if greater than

EFFECTS ON TEACHERS SELF-EFFICACY

745

.10); for RMSEA, between .06 and .10; for TLI, between .96 and .90; and for IFI, between .96 and .90. For each construct, we reduced measurement error by modeling each questionnaire items characteristics by using item response (IR) models (Baker & Kim, 2004). Some questionnaire items capture higher levels of teacher self-efficacy more precisely, whereas others capture lower levels of teacher self-efficacy more precisely (item difficulty). Likewise, there is variation in the precision of each question for distinguishing among teachers with higher versus lower self-efficacy (discrimination). Each teacher response ranged from 1 to 9. Hence, we modeled all of these characteristics with a generalized partial credit response test model (GPCM-IR, Baker & Kim, 2004).

A nested hypothesis test (log likelihood chi-square) indicated whether each set of explanatory variables was significant (Kennedy, 2004). Nonsignificant variables were removed. Then, we entered a vector of z teacher stress variables: classroom stress, workload stress, and overall teaching stress (Z). Yiy 0 y e iy xyXiy zyZiy (4)

Pi r 1

r j 1

ai bij

mi 1

k 1

(1)
ai bij

j 1

Pi(r) is the probability that a teacher with underlying value for a specific characteristic will give a rating r for question i, accounting for the discrimination strength (ai) and difficulty (bij) of the questionnaire item. A simpler partial credit model (PCM-IR) might fit the data better if discrimination (ai) is identical across items. Thus, we tested GPCM-IR and PCM-IR models and identified the best-fitting model with Bayesian expected a posteriori (EAP) estimation and log-likelihood difference chi-square tests (Bock & Mislevy, 1982; Kennedy, 2004; Mislevy & Bock, 1990). We computed each teachers self-efficacy using the best models EAP estimation, which is more precise than classical statistics methods (Baker & Kim, 2004). We repeated this procedure for teachers sources of stress and job satisfaction. These analyses yielded three teacher self-efficacy indices (classroom management, instructional strategies, and student engagement), two sources of stress indices (workload stress and classroom stress), and one job satisfaction index. Explanatory model. After computing the index values for each teacher, we estimated their relationships with systems of equations, specifically sequential sets of seemingly unrelated regressions (SUR; Kennedy, 2004) to account for heteroskedasticity and contemporaneous correlation in the errors across equations via Eviews software (Lilien, Startz, Ellsworth, Noh, & Engle, 1995). We entered the variables according to time constraints, expected causal relationships, and likely importance. Y iy 0 y e iy (2)

Next, all of these explanatory variables were entered into an OLS regression with teacher job satisfaction as the outcome variable. The teacher self-efficacy variables were added last. We used the Sobel (1982) test to identify mediation effects by testing the hypotheses that the explanatory variables direct and total effects on the outcome variable do not differ in the presence of potential mediators. We found that a 10% increase in each continuous variable above its mean was linked to the outcome variable (result b 10%; for simple dummy variables: result b 34%, 1 SD 34%; for contrast-coded dummy variables: result b 2 34%). As percentage of increase is not linearly related to standard deviation, scaling is not warranted. An alpha level of .05 was used. Testing many hypotheses increases the likelihood that at least one test will incorrectly reject a null hypothesis (a false positive result). To control for the false discovery rate, we used the two-stage linear step-up procedure, which outperformed 13 other methods in computer simulations (Benjamini, Krieger, & Yekutieli, 2006). We used an SEM to test these results simultaneously (Jreskog & Srbom, 2004). As the linear and quadratic terms of years of experience were highly correlated (r .99), the two variables were combined into one variable (0.02133 years of experience2 years of experience) in the SEM to prevent a near-singular matrix error. Nonsignificant variables were removed to yield the final SEM.

Results Test and Summary Statistics


The factor analyses yielded three teacher self-efficacy indices (classroom management, instructional strategies, and student engagement), two sources of stress indices (classroom stress and workload stress), and one job satisfaction index. Table 1 presents means, standard deviations, and reliability coefficients for the study variables. Factor analysis results generally confirmed the expected three-factor pattern of the TSES although one item How much can you do to assist families in helping their children do well in school? did not load as expected with the efficacy for student engagement factor and was deleted. The content of this item stands alone in the TSES as a measure of teachers efficacy to influence events outside the classroom and does not appear to measure the same content as other items in the student engagement factor. Appendix Tables A1 and A2 present results for eigenvalues and factor analysis results. The large, dominant first eigenvalue and explained variance indicate single factors for each set of test questions. The GPCM-IR model fit the data for each of these teacher characteristics better than did the PCM-IR model, showing that the discrimination strength of the questionnaire items differed (see Appendix Tables A3 and A4 for summary statistics of variables from item response models; see Appendix Table A4 for

0y are the grand mean intercepts of Yiy, a vector of y outcome variables (classroom management self-efficacy, instructional strategies self-efficacy, and student engagement self-efficacy) for each teacher i. The residuals are eiy. First, we entered a vector of x teacher and school background variables: gender, country of birth, nationality, years of experience, the squared term of years of experience, range of grade levels in current school, years in current school, and range of grade levels taught by the teacher in the current school (X). Yiy 0 y e iy xyXiy (3)

746
Table 1 Means and Standard Deviations for Study Variables
Variables Classroom management self-efficacy (four items) Instructional strategies self-efficacy (four items) Student engagement self-efficacy (three items) Job satisfaction (two items) Workload stress (four items) Classroom stress (three items) Overall stress (one item) Note.

KLASSEN AND CHIU

Item mean 7.56 7.55 6.87 7.30 5.82 5.40 6.81

Scale range 1036 1136 727 218 436 327 19

Scale mean 30.25 30.21 20.61 14.60 23.26 15.12 6.81

Scale SD 3.94 4.32 3.44 2.68 6.50 5.46 2.05

.85 .76 .82 .84 .68 .84

N 1,430. Student engagement self-efficacy was three items.

standard errors of the GPCM-IR teacher characteristics and Appendix Table A5 for the correlationvariance covariance matrix).

Explanatory Model
Preliminary analysis using SUR, OLS, and mediation tests yielded a candidate model that was fit successfully via an SEM (see Figure 2; for SEM details, see Appendix Tables A6 A10, which show a good fit between the SEM and the data). Detailed results of the factor analyses, IR models, SUR, and OLS are available from the authors. Classroom management self-efficacy. Teachers gender, years of experience, school type, teaching grade, and sources of stress were linked to their classroom management self-efficacy (Figure 2). Compared with female teachers, male teachers averaged 5% better classroom management self-efficacy, 5%

0.16 34% (see previous analysis section on percentage increase). As shown in Figure 3, years of experience has nonmonotonic relationships with all teacher self-efficacies; on average, classroom management self-efficacy increases from 0 years of experience to about 23 years of experience and falls afterwards. At the peak, teachers with 23 years of experience averaged 76% greater classroom management self-efficacy than that of new teachers, 76% 0.19 ( 0.02133 232 23) 34% (see Appendix Table A6 and the discussion of SEM in the previous analysis section). Teachers working in elementary schools averaged 7% better classroom management self-efficacy than those in schools in which one or more sets of grade levels (elementary, junior high school, and senior high school) were combined, 7% 0.10 2 34%. Furthermore, teachers who taught kindergarten students

Years of + (non-monotonic) experience

School combinations (vs. Elementary)

0.10 Teaching 1st or 2nd grade (vs. Kindergarten) 0.05 0.16 Classroom management Self-Efficacy +0.26

Female Teacher +0.37 Workload stress +0.56

+0.16

Overall 0.21 Teaching stress 0.52 +0.24 Classroom stress 0.25 Instructional Strategies Self-Efficacy +0.29

Job Satisfaction

0.24

Other Schools (vs. Elementary) + (non-monotonic) + (non-monotonic)

Teaching other grades (vs. +0.16 Kindergarten)

0.25 0.09 0.12

Student Engagement Self-Efficacy

Figure 2. Structural equation model for teachers self-efficacies and job satisfaction with their explanatory variables. Teacher demographics and school characteristics are all exogenous variables, and other variables are endogenous variables. Ovals indicate latent variables, and rectangles indicate single variables. Questionnaire items for each latent variable are not shown. For the quadratic relationship between years of experience and the three teacher self-efficacy variables, see Figure 3.

EFFECTS ON TEACHERS SELF-EFFICACY

747

Figure 3. Relationship of years of experience with three teacher self-efficacy variables (teaching strategies, classroom management, and student engagement) based on the structural equation model results.

averaged 3% better classroom management self-efficacy than those who taught Grade 1 or 2 students, 3% 0.05 2 34%. The links of workload stress and classroom stress to classroom management self-efficacy differed substantially. Teachers whose workload stress exceeded the mean by 10% averaged 2% better classroom management self-efficacy, 2% 0.16 10%. In contrast, teachers whose classroom stress exceeded the mean by 10% averaged 5% worse classroom management self-efficacy, 5% 0.52 10%. These variables accounted for 25% of the variance in teachers classroom management self-efficacy (see Appendix Table A6, squared multiple correlations). Instructional strategies self-efficacy. Teachers years of experience and classroom stress were linked to their instructional strategies self-efficacy, showing a nonmonotonic relationship with an increase in instructional strategies self-efficacy up to about 23 years of experience and then falling. The instructional strategies self-efficacy of teachers with 23 years of experience averaged 88% greater than that of new teachers, 88% 0.22 ( 0.02133 232 23) 34% (see Table A6 and discussion of SEM in the preceding analysis section). When their classroom stress exceeded the mean by 10%, teachers averaged 3% less instructional strategies self-efficacy, 3% 0.25 10%. These variables accounted for 11% of the variance in teachers instructional strategies selfefficacy. Student engagement self-efficacy. Teachers years of experience, school type, teaching grade, and classroom stress were linked to student engagement self-efficacy, which again showed a nonmonotonic relationship with increasing self-efficacy up to midcareer and then falling in late career. The student engagement self-efficacy of teachers with 23 years of experience averaged 68% greater than that of new teachers, 68% 0.17 ( 0.02133 232 23) 34% (see Table A6 and discussion of SEM in the preceding analysis section). Teachers working in elementary schools averaged 8% more student engagement self-efficacy than those working in other types of schools, 8% 0.12 34%. Furthermore, teachers who taught kindergarten students averaged 6% more student engagement self-efficacy than those who taught students in higher grades, 6% 0.09 34%. When their classroom stress exceeded the mean by 10%, teachers averaged 3% less student engagement self-efficacy, 3% 0.25 10%. These variables accounted for 12% of the variance in teachers student engagement self-efficacy.

Job satisfaction. Teachers overall teaching stress and selfefficacies were linked to job satisfaction. Teachers with 10% greater overall teaching stress averaged 2% less job satisfaction, 2% 0.21 10%. Meanwhile, teachers with 10% more classroom management self-efficacy or 10% more instructional strategies self-efficacy averaged 3% more job satisfaction, 3% 0.26 10%; 3% 0.29 10%. These variables accounted for 31% of the variance in teachers job satisfaction. Stress and gender. Teachers with 10% more workload stress had 6% more overall teaching stress, 6% 0.56 10%. Workload stress accounted for 31% of the variance in teachers overall teaching stress. Teachers gender was linked to sources of stress. Female teachers averaged 13% more workload stress and 8% more classroom stress than male teachers, 13% 0.37 .34; 8% 0.24 .34. Gender accounted for 3% and 1% of the variances in workload stress and classroom stress, respectively.

Discussion
In this study, we modeled the relationships among teacher characteristics, years of experience, three forms of teachers selfefficacy, job stress, and job satisfaction. The factor analysis confirmed the expected factor pattern, and the SEM yielded by the SUR, OLS, and mediation tests fit the data well, with teachers self-efficacy for instructional strategies and classroom management positively influencing job satisfaction, whereas overall job stress lowered job satisfaction. The results show how self-efficacy varies with years of teachers experience. Furthermore, the results show how teachers self-efficacies mediate the links between two types of stress on job satisfaction. Female teachers had higher levels of both classroom and workload stress. Similar to previous findings, our results show that years of experience and job-related stress were related to teachers self-efficacy, which in turn influenced job satisfaction. The key new finding in the study was that teachers self-efficacy was influenced by years of experience in a nonlinear relationship, with the three factors of teacher efficacy increasing with experience for early and mid-career stage teachers and declining for teachers in the late career stages.

748 Teachers Self-Efficacy

KLASSEN AND CHIU

Teachers self-efficacy showed a nonlinear relationship with years of teaching experience; self-efficacy increased from 0 to about 23 years of experience and then declined as years of experience increased. Our results show that this relationship held true for each of the three factors of teachers self-efficacy, reflecting a relationship that has not been noted in previous research on teachers self-efficacy. Teachers confidence in engaging students, managing student behavior, and using effective instructional strategies showed the same pattern of growth and gradual decline. Whereas previous researchers have noted that self-efficacy increases with teachers experience (e.g., Wolters & Daugherty, 2007), the relationship between teachers experience and self-efficacy may be more complex than previously believed. Bandura (1997) proposed that self-efficacy beliefs remain relatively stable once established, and although this stability may be true within a specific career stage, the results from our study suggest that teachers gain confidence in their teaching skills through their early years and into the mid-career years but that these levels of confidence may decline as teachers enter the later stages of their careers. The career stages outlined by Huberman (1989) in his study of the professional life cycle of teachers map well on to the patterns of teachers self-efficacy found in the current study. According to Huberman, teachers undergo a process of survival and discovery in the early career years, during which the gulf between professional ideals and daily classroom life is exposed, and self-doubts and initial enthusiasm are entwined. About 4 6 years into their careers, teachers enter a period of stabilization, marked by a definitive commitment to the profession (or the choice to leave the profession). The mid-career years (718) are marked by periods of experimentation and activism or by a period of reassessment, during which teachers take stock of their careers and question their career choices. Huberman suggested two phases during the latercareer years. During Years 19 30, teachers experience a period of serenity, during which a gradual loss in energy and enthusiasm is compensated for by a greater sense of confidence and selfacceptance (p. 35, italics ours). Finally, teachers in the late-career stage (Years 31 40) move into a period of disengagement, marked either by serenity or disappointment and bitterness. Our finding of teachers self-efficacy peaking at about 23 years of experience and then declining in the later-career years corresponds with the motivation pattern suggested by Huberman. More recent studies have built on Hubermans work, with Day and Gu (2007) finding that most teachers in mid-career (i.e., Years 8 23) experience increases in motivation and commitment, whereas increased proportions of teachers in the later stage of their career stage (24 years of experience) report declining motivation. Authors of previous studies outside education have noted this decline in work motivation in the late-career stages. A recent study by Kooij et al. (2008) examined research on work motivation and aging and found that many age-related factors (i.e., chronological age, physical health, self-perception, social perception, skill obsolescence, and life stage) had a negative impact on the motivation beliefs of older workers. In addition, Kooij et al. found that work motivation was influenced by an interaction of age-related factors. For instance, declining health may be related to a deterioration of self-concept or changes in weighting of work- and leisure-related values, but stereotyped perceptions of peers also influence work-

ers motivation and result in reduced skills, motivation, and opportunities for promotion. The lower levels of older teachers self-efficacy beliefs may be influenced not only by biological and psychological changes related to chronological age but also by student and peer perceptions of declining competence influenced by stereotyped beliefs about aging. In sum, age-related changes in motivation beliefs, like self-efficacy, are influenced not only by chronological age but by the psychosocial context of the work environment. The contexts in which the teachers worked were also linked with their self-efficacy. Teaching in elementary schools and teaching kindergarten were linked with higher levels of self-efficacy for classroom management and student engagement. There has been surprisingly little research on how teaching context influences teachers selfefficacy. Wolters and Daugherty (2007) found that teachers in higher grade levels reported lower self-efficacy than teachers in lower grade levels and that the inverse relationship between teaching level and self-efficacy was especially marked for teachers of elementaryschool-aged students in comparison to teachers of middle- and highschool-aged students. Our study also found that teachers in higher grade levels had lower self-efficacy, but the pattern of grade-leveldependent self-efficacy was also found within teaching levels, at least in the early elementary school grades. Teachers of the youngest students had higher levels of self-efficacy than teachers of older students within elementary schools, and this result was observed for teachers self-efficacy for classroom management and student engagement, although not for instructional strategies. Together, these combined results suggest that variation of teachers selfefficacy associated with teaching level can also occur within a school.

Teachers Stress and Job Satisfaction


Teachers with higher overall teaching stress had lower job satisfaction, whereas classroom stress was indirectly linked to job satisfaction through self-efficacy for classroom management and instructional strategies. We predicted workload and classroom stress to be negatively linked with self-efficacy. As expected, teachers with greater classroom stress had less self-efficacy in all three factors, especially classroom management self-efficacy. Teachers who perceived higher levels of classroom stress from student misbehavior reported lower levels of self-efficacy for classroom management. Likewise, teachers reporting greater workload stress had greater overall stress. However, teachers reporting more workload stress had greater classroom management self-efficacy (with no significant differences in the other two self-efficacy factors). It may be that teachers who experience higher levels of classroom stress from student misbehavior (i.e., from noisy and impolite behavior) have lower confidence to manage that behavior due to a history of unsuccessful experiences. In contrast, it may be that teachers who perceive greater stress from responsibility for student achievement and heavy workloads exert more effort during lesson planning and are better prepared to manage student behaviors during class. Female teachers had higher levels of workload and classroom stress. A growing number of researchers have noted the link between gender and work-related stress. For example, Antoniou et al. (2006) found that female teachers experienced higher levels of work-related stress compared with male teachers, particularly for

EFFECTS ON TEACHERS SELF-EFFICACY

749

classroom and workload factors. Our results are consistent with those of previous studies showing modest but persistent gender differences in job stress among teachers (e.g., Antoniou et al., 2006; Chaplain, 2008). Greenglass and Burke (2003) proposed that the elevated work stress of females might stem from gender differences in nonwork domains, with higher total workload (school tasks plus domestic tasks) and higher role conflict between work and family roles. These previous findings do not explain female teachers higher levels of stress from student behavior. Hopf and Hatzichristou (1999) found female teachers to be more sensitive to externalizing behavior problems, especially from adolescent male students, and also found male teachers assessed childrens interpersonal behaviors as less problematic than did female teachers. Findings of gender differences in teacher stress bear further research. Results from the current study reinforce previous findings that teacher self-efficacy is linked with job satisfaction. Teachers with high levels of self-efficacy for classroom management and instructional strategies reported higher levels of job satisfaction, whereas teachers with high levels of overall stress reported lowered job satisfaction. Caprara et al. (2006) found that Italian teachers self-efficacy was linked to their job satisfaction, although their conceptualization of self-efficacy was less specific than the Tschannen-Moran and Woolfolk Hoy (2001) measure used in this study. In particular, Bandura (1997) and others (e.g., Pajares, 1996) have argued that more specific judgments of self-efficacy provide more information about how the construct influences beliefs and related behaviors. In the current study, we found teachers self-efficacies for classroom management and instructional strategies were directly related to job satisfaction, whereas self-efficacy for student engagement did not play a direct role. It appears that not all facets of teachers self-efficacy are linked to job satisfaction in the same way.

our findings about the apparent changes in teachers self-efficacy in the late-career stages lead us to propose that future studies should be focused on middle and late-career stage teachers motivation beliefs, an area that has been neglected despite the large number of teachers who are past the first decade of their teaching careers. Several data limitations hamper the generalizability of our results. Additional indicators of teachers success and functioning not included in our model may influence job satisfaction. Also, the measure of overall job stress consisted of only a single item, and job satisfaction was measured by two items. However, results of recent studies have supported the inclusion of single-item measures of job-related beliefs (e.g., Dolbier, Webster, McCalister, Mallon, & Steinhardt, 2005; Nagy, 2002) because of high levels of face validity and convenience for data collection in busy workplace settings, and investigators in many previous studies have measured job stress using one item (e.g., Boyle et al., 1995; Chaplain, 2008; Manthei et al., 1996). We did not measure the longitudinal development of teachers self-efficacy, and readers should not infer from our results that the pattern of rise and fall of self-efficacy holds true for individual teachers over the career span. The results from the current research are prone to the limitations emerging from our reliance on a common method to assess each variable. Lastly, the sample was restricted to Canadian teachers in one province, and although teachers came from a wide variety of schools, the sample was not randomized, and participants in this study may not be representative of other groups of teachers in different settings.

Practical Implications and Conclusion


Considerable research has examined the development of teacher motivation beliefs at the beginning stages of teachers careers, but the teaching workforce in many settings is decidedly graying, with more teachers at the mid- or late-career stages than at the beginning-career stage (U.S. Department of Education, 2009). Building an understanding of the motivation profiles of teachers across the career span makes sense because of the number of midand late-career teachers and because teachers motivation profiles and willingness to engage in new practices varies according to career stage (Drake, 2002). A teachers skills, knowledge, and effectiveness may change over time without a continuous and focused effort to build those skills and knowledge on the part of the teacher, school district officials, and school administrators (Drake, 2002). One-size-fits-all professional development that aims to build the skills and knowledge of new, mid-career, and the most experienced teachers may not be optimally effective. For example, Greller (2006) suggested that older workers professional development needs shift from learning general skills to learning specific skills. Older workers seek professional development opportunities that offer greater autonomy in content, learning pace, and learning environment (Greller, 2006). Professional development programs that are tailored to teachers career stages may enhance skills and knowledge but also boost the confidence that teachers at a later career stage have in their capabilities to teach effectively. Using professional development opportunities to boost skills and teachers self-efficacy may lower job stress and enhance satisfaction from teaching.

Limitations and Future Research


Future researchers can replicate the model of teacher motivation presented in this article, with the addition of factors such as students socioeconomic status and teachers collective efficacy to help account for more job satisfaction variance. The effect of whole-school motivation, or collective efficacy, has been shown to influence the individual job satisfaction experienced by teachers (e.g., Caprara et al., 2003; Klassen et al., in press). The role of teachers self-efficacy in relation to job stress and job satisfaction may vary as a function of cultural context, and additional research examining the relationships among the study variables should be conducted in contrasting cultural settings. Our research provides new insight into the pattern of change in teachers self-efficacy beliefs, but stronger claims about the development of teachers self-efficacy could be made through longitudinal studies. As noted earlier, some studies of teachers selfefficacy during the early career years (e.g., Woolfolk Hoy & Burke Spero, 2005) have been performed, but to our knowledge, no one has conducted a longitudinal study using the most recent, conceptually sound measures of teachers self-efficacy. Conducting longitudinal research of teachers motivation beliefs presents a host of practical challenges not found in cross-sectional research, but findings from longitudinal studies can inform our understanding of how motivation beliefs develop over the career. In related fashion,

750

KLASSEN AND CHIU gence, social problem solving skills, and psychological distress: A study of Chinese undergraduate students. Journal of Applied Social Psychology. Cockburn, A. D., & Haydn, T. (2004). Recruiting and retaining teachers: Understanding why teachers teach. London, England: RoutledgeFalmer. Day, C., & Gu, Q. (2007). Variations in the conditions for teachers professional learning and development: Sustaining commitment and effectiveness over a career. Oxford Review of Education, 33, 423 443. Dolbier, C. L., Webster, J. A., McCalister, K. T., Mallon, M. W., & Steinhardt, M. A. (2005). Reliability and validity of a single-item measure of job satisfaction. American Journal of Health Promotion, 19, 194 198. Drake, C. (2002). Experience counts: Career stage and teachers responses to mathematics education reform. Educational Policy, 16, 311337. Evans, L. (2001). Delving deeper into morale, job satisfaction, and motivation among education professionals. Educational Management and Administration, 29, 291306. Gates, G. S. (2007). Emerging thought and research on student, teacher, and administrative stress and coping. Greenwich, CT: Information Age. Ghaith, G., & Yaghi, H. (1997). Relationships among experience, teacher efficacy, and attitudes toward the implementation of instructional innovation. Teaching and Teacher Education, 13, 451 458. Greenglass, E. R., & Burke, R. J. (2003). Teacher stress. In M. F. Dollard, A. H. Winefield, & H. R. Winefield (Eds.), Occupational stress in the service professions (pp. 213236). New York, NY: Taylor and Francis. Greller, M. M. (2006). Hours invested in professional development during late career as a function of career motivation and satisfaction. Career Development International, 6, 544 559. Gustafsson, J. E., & Balke, G. (1993). General and specific abilities as predictors of school achievement. Multivariate Behavioral Research, 28, 407 434. Hopf, D., & Hatzichristou, C. (1999). Teacher gender-related influences in Greek schools. British Journal of Educational Psychology, 69, 118. Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indices in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 155. Huberman, M. (1989). The professional life cycle of teachers. Teachers College Record, 91, 3157. Ingersoll, R. (2001). Teacher turnover and teacher shortage. American Educational Research Journal, 38, 499 534. Jreskog, K., & Srbom, D. (2004). LISREL 8.7 for Windows [Computer software]. New York, NY: Scientific Software International. Jepson, E., & Forrest, S. (2006). Individual contributory factors in teacher stress: The role of achievement striving and occupational commitment. British Journal of Educational Psychology, 76, 183197. Judge, T. A., Thoresen, C. J., Bono, J. E., & Patton, G. K. (2001). The job satisfactionjob performance relationship: A qualitative and quantitative review. Psychological Bulletin, 127, 376 407. Kennedy, P. (2004). A guide to econometrics. Cambridge, MA: MIT Press. Klassen, R. M. (in press). Teachers stress: The mediating role of collective efficacy beliefs. Journal of Educational Research. Klassen, R. M., Bong, M., Usher, E. L., Chong, W. H., Huan, V. S., Wong, I. Y., & Georgiou, T. (2009). Exploring the validity of the Teachers Self-Efficacy Scale in five countries. Contemporary Educational Psychology, 34, 6776. Klassen, R. M., Usher, E. L., & Bong, M. (in press). Teachers collective efficacy, job satisfaction, and job stress in cross-cultural context. Journal of Experimental Education. Kokkinos, C. M. (2007). Job stressors, personality, and burnout in primary school teachers. British Journal of Educational Psychology, 77, 229 243. Kooij, D., de Lange, A., Jansen, P., & Dikkers, J. (2008). Older workers

Our study extends teacher motivation research by showing how teachers years of experience, gender, and three domains of selfefficacy (student engagement, instructional strategies, and classroom management) were related to their job stress (workload and classroom stress) and job satisfaction. Building on previous research showing that teacher self-efficacy often increases in the early stages of teachers careers, we found that early- to mid-career teachers reported progressively greater self-efficacies in these three areas, while late-career teachers reported less self-efficacy in each area. Female teachers had greater stress (from both workload and student behaviors during class) and lower self-efficacy for classroom management. Teaching younger children (in elementary grades and kindergarten) was linked with higher levels of selfefficacy for classroom management and student engagement. Finally, the impact of classroom and workload stress on job satisfaction was mediated by teachers self-efficacy.

References
Abel, M. H., & Sewell, J. (1999). Stress and burnout in rural and urban secondary school teachers. The Journal of Educational Research, 92, 287293. Alberta Education. (2009). Alberta teaching certification standards. Retrieved February 14, 2009, from http://education.alberta.ca/department/ stats/certification.aspx Antoniou, A.-S., Polychroni, F., & Vlachakis, A.-N. (2006). Gender and age differences in occupational stress and professional burnout between primary and high-school teachers in Greece. Journal of Managerial Psychology, 21, 682 690. Baker, F. B., & Kim, S.-H. (2004). Item response theory. Boca Raton, FL: CRC. Bandura, A. (1997). Self-efficacy: The exercise of control. New York, NY: Freeman. Bandura, A. (2006). Guide for constructing self-efficacy scales. In F. Pajares & T. Urdan (Eds.), Adolescence and education: Vol. 5. Selfefficacy and adolescence (pp. 307337). Greenwich, CT: Information Age. Benjamini, Y., Krieger, A. M., & Yekutieli, D. (2006). Adaptive linear step-up procedures that control the false discovery rate. Biometrika, 93, 491507. Betoret, F. D. (2006). Stressors, self-efficacy, coping resources, and burnout among secondary school teachers in Spain. Educational Psychology, 26, 519 539. Bock, R. D., & Mislevy, R. J. (1982). Adaptive EAP estimation of ability in a microcomputer environment. Applied Psychological Measurement, 6, 431 444. Bong, M. (2006). Asking the right question: How confident are you could successfully perform these tasks? In F. Pajares & T. Urdan (Eds.), Self-efficacy beliefs of adolescents (pp. 287305). Greenwich, CT: Information Age. Boyle, G. J., Borg, M. G., Falzon, J. M., & Baglioni, A. J., Jr. (1995). A structural model of the dimensions of teacher stress. British Journal of Educational Psychology, 65, 49 67. Caprara, G. V., Barbaranelli, C., Borgogni, L., & Steca, P. (2003). Efficacy beliefs as determinants of teachers job satisfaction. Journal of Educational Psychology, 95, 821 832. Caprara, G. V., Barbaranelli, C., Steca, P., & Malone, P. S. (2006). Teachers self-efficacy beliefs as determinants of job satisfaction and students academic achievement: A study at the school level. Journal of School Psychology, 44, 473 490. Chaplain, R. P. (2008). Stress and psychological distress among trainee secondary teachers in England. Educational Psychology, 28, 195209. Chow, B. W.-Y., Chiu, M. M., & Wong, S. (in press). Emotional intelli-

EFFECTS ON TEACHERS SELF-EFFICACY motivation to continue to work: Five meanings of age. Journal of Managerial Psychology, 23, 364 394. Kyriacou, C. (2001). Teacher stress: Directions for future research. Educational Review, 53, 2735. Lilien, D. M., Startz, R., Ellsworth, S., Noh, J., & Engle, R. (1995). Eview [Computer software]. Irvine, CA: Quantitative Micro Software. Liu, X. S., & Ramsey, J. (2008). Teachers job satisfaction: Analyses of the Teacher Follow-Up Survey in the United States for 2000 2001. Teaching and Teacher Education, 24, 11731184. Manthei, R., Gilmore, A., Tuck, B., & Adair, V. (1996). Teacher stress in intermediate schools. Educational Research, 38, 319. Mislevy, R. J., & Bock, R. D. (1990). BILOG 3 (2nd ed.) [Computer software]. Mooresville, IN: Scientific Software International. Nagy, M. S. (2002). Using a single-item approach to measure facet job satisfaction. Journal of Occupational and Organizational Psychology, 75, 77 86. Pajares, F. (1996). Self-efficacy beliefs in academic settings. Review of Educational Research, 66, 543578. Ross, J. A., Cousins, J. B., & Gadalla, T. (1996). Within-teacher predictors of teacher efficacy. Teaching and Teacher Education, 12, 385 400. Schwarzer, R., & Hallum, S. (2008). Perceived teacher self-efficacy as a predictor of job stress and burnout: Mediation analysis. Applied Psychology: An International Review, 57, 152171. Skaalvik, E. M., & Skaalvik, S. (2007). Dimensions of teacher self-efficacy and relations with strain factors, perceived collective teacher efficacy, and teacher burnout. Journal of Educational Psychology, 99, 611 625.

751

Sobel, M. E. (1982). Asymptotic intervals for indirect effects in structural equations models. In S. Leinhart (Ed.), Sociological methodology 1982 (pp. 290 312). San Francisco, CA: JosseyBass. Spickard, A., Gabbe, S. G., & Christensen, J. F. (2002). Mid-career burnout in generalist and specialist physicians. Journal of the American Medical Association, 288, 14471450. Tschannen-Moran, M., & Woolfolk Hoy, A. (2001). Teacher efficacy: Capturing an elusive construct. Teaching and Teacher Education, 17, 783 805. Tschannen-Moran, M., & Woolfolk Hoy, A. (2007). The differential antecedents of self-efficacy beliefs of novice and experienced teachers. Teaching and Teacher Education, 23, 944 956. Tschannen-Moran, M., Woolfolk Hoy, A., & Hoy, W. K. (1998). Teacher efficacy: Its meaning and measure. Review of Educational Research, 68, 202248. U.S. Department of Education. (2009). Special analysis 2005: Mobility in the teacher workforce. Retrieved February 12, 2009, from http:// nces.ed.gov/programs/coe/2005/analysis/sa01.asp Wolters, C. A., & Daugherty, S. G. (2007). Goal structures and teachers sense of efficacy: Their relation and association to teaching experience and academic level. Journal of Educational Psychology, 99, 181193. Woolfolk Hoy, A., & Burke Spero, R. (2005). Changes in teacher efficacy during the early years of teaching: A comparison of four measures. Teaching and Teacher Education, 21, 343356.

Appendix Ancillary Tables and Results


Table A1 Eigenvalues Showing Single Dominant Factors in Each Set of Test Questions
Eigenvalues Factor Classroom management self-efficacy Instructional strategies self-efficacy Student engagement self-efficacy Workload stress Classroom stress Note. 1st 2.76 2.39 2.22 2.24 2.06 2nd 0.45 0.62 0.46 0.45 0.75 3rd 0.41 0.54 0.33 0.31 0.67 1st/2nd 6.13 3.85 4.86 4.93 2.74 2nd/3rd 1.10 1.14 1.39 1.46 1.13 % of variance explained by 1st eigenvalue 69 60 74 75 52 2nd eigenvalue 11 16 15 15 19

The factor of job satisfaction only two variables; polychoric correlation 0.729.

(Appendix continues)

752

KLASSEN AND CHIU

Table A2 Three-Factor and Two-Factor Structures Best Fit Responses to Teacher Self-Efficacy and Teacher Stress Questions as Shown by Goodness-of-Fit Measures (Varimax Rotation)
Factor structure Goodness-of-fit measure Teacher self-efficacy Standardized root mean residual Comparative fit index Incremental fit index TuckerLewis index Root-mean-squared error of approximation Chi-square test Degrees of freedom p Adjusted goodness-of-fit index Relative fit index Teacher stress Standardized root mean residual Comparative fit index Incremental fit index TuckerLewis index Root-mean-squared error of approximation Chi-square test Degrees of freedom p Adjusted goodness-of-fit index Relative fit index Three 0.041 0.957 0.957 0.933 0.053 208 41 0.000 0.946 0.924 0.015 0.990 0.992 0.976 0.046 41 13 0.000 0.971 0.971 Two Hierarchical 0.066 0.901 0.901 0.871 0.074 409 41 0.000 0.903 0.863 0.068 0.881 0.882 0.825 0.138 322 12 0.000 0.808 0.821 Nested 0.114 0.924 0.924 0.901 0.091 674 33 0.000 0.639 0.892 0.075 0.814 0.815 0.740 0.163 405 7 0.000 0.745 0.736 Single 0.137 0.540 0.541 0.438 0.175 1,334 44 0.000 0.626 0.433 0.068 0.749 0.751 0.610 0.053 961 14 0.000 0.797 0.607

Table A3a Summary Statistics of Variables and of Item Response Models


GPCM vs. PCM IRT Variable Classroom management self-efficacy Instructional strategies self-efficacy Student engagement self-efficacy Job satisfaction Female Years of experience School grade range School combinations (vs. elementary) Other (vs. elementary) Teaching Grade 1 or 2 (vs. K) Teaching other grades (vs. K) Workload stress Classroom stress Overall stress Mean 0.08 0.03 0.00 0.03 0.70 13.21 SD 0.93 0.88 0.93 0.91 9.97 Min 3.39 2.99 3.03 2.87 0 0 1 1 1 1 2.31 2.22 1 Median 0.13 0.03 0.00 0.08 1 10 1 1 0 1 0.02 0.01 7 Max 1.75 1.69 1.90 1.50 1 43 1 1 1 1 1.93 2.25 9 df 4 4 3 3 LL 2 57.06 460.08 121.12 61.61

0.01 0.00 6.81

0.82 0.92 2.06

9 3

1865.20 233.77

Note. Values created from responses to sets of questions with item response model comparison tests, showing that the generalized partial credit models (GPCM) fit the data better than the partial credit models (PCM). IRT item response tests; LL 2 log-likelihood difference chi-square test; K kindergarten. p .001.

(Appendix continues)

EFFECTS ON TEACHERS SELF-EFFICACY

753

Table A3b Percentage of Teachers by School Level


Variable School grade range Elementary 20 High school 20 Junior high 10 Elementary/junior high 26 Junior high/high school 15 Elementary/junior high/high school 9

Table A3c Percentage of Teachers by Grade Level


Variable Teaching grade K 3 12 7 34 7 56 13 79 19 1012 20 Head Start 0.3 Multiple grades 29

Table A3d Percentage of Responses for Job Stress Scale


Teaching grade Variable Overall teaching stress Note. 1 2 2 2 3 5 4 6 5 8 6 11 7 23 8 17 9 26

Scale ranges from 1 (no stress) to 9 (extreme stress).

Table A4 Standard Errors of Each Teacher Property at 2nd, 16th, 50th, 84th, and 98th Percentiles, Showing Greater Standard Errors at Higher Percentiles
Standard errors at each percentile Teacher property Job satisfaction Classroom management Instructional strategies Student engagement Workload stress Classroom stress 2nd 0.07 0.12 0.19 0.09 0.21 0.10 16th 0.10 0.15 0.26 0.12 0.23 0.10 50th 0.14 0.25 0.41 0.12 0.42 0.19 84th 0.45 0.56 0.77 0.42 0.72 0.44 98th 1.16 0.93 1.11 0.96 1.00 0.98

(Appendix continues)

754

KLASSEN AND CHIU

Table A5 Correlations (Lower Left Triangle), Variances (Diagonal), and Covariances (Upper Right Triangle)
Variable 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. Classroom management self-efficacy Instructional strategies self-efficacy Student engagement self-efficacy Female gender Years of experience Years of experience squared School grade range: School combinations (vs. elementary) School grade range: Other (vs. elementary) Teaching Grade 1 or 2 (vs. K) Teaching other grades (vs. K) Workload stress Classroom stress Overall stress Job satisfaction K kindergarten. 1 2 3 0.44 0.44 0.87 0.05 0.15 0.12 4 5 6 7 0.07 0.02 0.07 0.06 0.42 12.76 0.61 8 0.05 0.02 0.10 0.06 0.65 17.19 9 0.02 0.02 0.02 0.01 0.09 2.66 10 0.01 0.00 0.04 0.02 0.00 0.08 11 0.07 0.05 0.08 0.05 0.13 1.81 12 0.32 0.18 0.17 0.04 0.08 1.67 0.00 13 14

0.86 0.41 0.50 0.78 0.51 0.36 0.07 0.21 0.16 0.19 0.12 0.03

0.03 1.47 36.66 0.01 1.82 54.85 0.02 1.39 37.50 0.21 0.27 10.24 0.06 99.43 3206 0.07 0.96 112,464 0.05

0.25 0.41 0.07 0.32 0.23 0.40 0.12 0.05 0.65 0.14 13.37 0.14 0.08 0.08 0.07 0.00 0.01 0.21 0.26 0.26 0.82

00.10 0.03 0.10 0.18 0.05 00.07 00.06 00.02 00.10 00.38 00.13 0.34 0.05 0.01 0.07 0.21 0.04 0.25 0.13 0.15 0.08 0.05 0.06 0.03 0.11 0.10 0.00 0.10 0.13 0.02 0.20 0.10 0.01 0.12 0.13 0.03 0.34 0.02 1.24

0.52 0.01 0.64 0.01 0.04 0.11 0.15 0.60 0.03 0.03 0.02 0.02 0.06 0.06 0.05 0.00

0.04 0.05

0.06 0.83 0.02 0.05 0.00 0.15 0.01 0.08 0.01 0.00 0.02 0.05 42.01 0.06

0.05 0.02 0.02 0.10 0.07 0.01 0.01 0.04 0.14 0.01 0.00 0.01 0.02 0.68 0.34 0.73 0.00 0.45 0.85 0.58 0.02 0.43 0.31 4.22 0.00 0.16 0.22 0.48

Note.

Table A6 Total Effect and Indirect Effect of Each Independent Variable on Each Dependent Variable
Dependent variable Stress Independent variable Female gender Total effect Indirect effect Years of experience quadratic term Total effect Indirect effect School grade range School combinations vs. elementary Total effect Indirect effect Other vs. elementary Total effect Indirect effect Teaching Grade 1 or 2 vs. K Total effect Indirect effect Teaching other grades vs. K (total effect) Workload stress Total effect Indirect effect Classroom stress Total effect Indirect effect Overall stress (total effect) Classroom management self-efficacy Instructional strategies self-efficacy Squared multiple correlations Note. Teaching other grades (vs. K) 0.24 Workload 0.37 Classroom 0.24 Overall 0.21 0.21 Classroom management 0.10 0.06 0.19 Self-efficacy Instructional strategies 0.06 0.06 0.22 Student engagement 0.02 0.02 0.17 Job satisfaction 0.10 0.10 0.11 0.11 0.03 0.03 0.11 0.01 0.05 0.09 0.56 0.16 0.52 0.25 0.25 0.01 0.01 0.08 0.08 0.21 0.21 0.21 0.26 0.29 0.31

0.1 0.16

0.04

0.03

0.01

0.31

0.25

0.11

0.12

Blanks indicate no total or no indirect effects. K kindergarten.

(Appendix continues)

EFFECTS ON TEACHERS SELF-EFFICACY

755

Table A7 Completely Standardized Solution of Structural Equation Model: Psi Matrix


Variable 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 1

10

11

Job satisfaction 0.69 Classroom management self-efficacy Instructional strategies self-efficacy Student engagement self-efficacy School grade range: School combinations (vs. elementary) School grade range: Other (vs. elementary) Teaching Grade 1 or 2 (vs. K) Teaching other grades (vs. K) Workload stress Classroom stress Overall stress

0.75 0.48 0.89 0.45 0.46 0.88 1.00 0.83 1.00 1.00 0.61 0.96 0.97 0.59 0.99 0.69

Note. K kindergarten. p .001.

Table A8 Completely Standardized Solution of Structural Equation Model: Lambda-Y Matrix


Factor Self-efficacy Survey item I am satisfied with what I achieve at work. I feel good at work. How much can you do to control disruptive behavior in the classroom? How much can you do to get children to follow classroom rules? How much can you do to calm a student who is disruptive or noisy? How much can you do to establish a classroom management system with each group of students? How much can you do to craft good questions for students? How much can you do to implement a variety of assessment strategies? How much can you do to provide an alternative explanation when students are confused? How much can you do to implement alternative strategies in your classroom? How much can you do to motivate students who show low interest in school work? How much can you do to get students to believe they can do well in school work? How much can you do to help students to value learning? How great a source of stress is having too much work to do? How great a source of stress is having extra duties/responsibilities because of absent teachers? How great a source of stress is having a large class size? How great a source of stress is being responsible for students achievement? How great a source of stress is having noisy students? How great a source of stress is maintaining class discipline? How great a source of stress is dealing with students impolite behavior or rudeness?
a

Stress Instructional strategies Workload Classroom

Job satisfaction 0.85 0.85


a

Classroom management

Student engagement

0.75a 0.78 0.75 0.80 0.60a 0.70 0.66 0.73 .78a .84 .74

0.66a 0.55 0.54 0.60 0.70a 0.90 0.80

The first lambda-Y estimates of each latent variable is fixed by default. p .001.

(Appendix continues)

756

KLASSEN AND CHIU

Table A9 Structural Equation Model Showing a Good Fit With the Data
Measure Standardized root mean residual (SRMR) Root-mean-squared error of approximation (RMSEA) TuckerLewis index (TLI) Incremental fit index (IFI) Comparative fit index (CFI) 2(305), p .000 Adjusted goodness of fit index Relative fit index Result .039 .041 .967 .971 .971 1068.00 .937 .954

Note. Threshold values separate good, moderate, and poor fits for each measure: for SRMR, good fit .08 moderate fit .10 poor fit; for RMSEA, between .06 and .10; for TLI, between .96 and .90; for IFI, between .96 and .90. Other measures are also included for reader interest.

Table A10 Sobel Mediation Tests for the Outcome Variable Job Satisfaction
Initial variable3Mediator Years of experience quadratic term3Class management self-efficacy School combinations (vs. elementary)3Class management self-efficacy Teaching Grade 1 or 2 (vs. K)3Class management self-efficacy Workload stress3Class management self-efficacy Classroom stress3Class management self-efficacy Workload stress3Overall stress Years of experience quadratic term3Instructional strategies self-efficacy Classroom stress3Instructional strategies self-efficacy

% change 20 19 13 20 13 10 13 21

z 2.039 2.638 2.225 3.285 4.212 2.678 1.986 3.795

p .05.

p .01.

p .001.

Received March 11, 2009 Revision received January 26, 2010 Accepted January 27, 2010

Potrebbero piacerti anche