Teachers’ Sense of Efficacy for the Formation of Students’ Character

The construct of teacher efficacy has been used widely in an attempt to capture teachers’ beliefs about their abilities to accomplish desired outcomes. Teacher efficacy is derived from a combination of Rotter’s (1966) social learning theory and Bandura’s (1977) social cognitive theory. During the 1970s, researchers for the Rand Corporation (Armor et al., 1976) applied Rotter’s (1966) notion of locus of control to teaching situations. This research examined whether teachers perceived their control of outcomes as internally or externally influenced. The internal-external locus of control distinction was measured by two items: 1) “When it comes right down to it, a teacher really can’t do much because most of a student’s motivation and performance depends on his or her home environment” and 2) “If I try really hard, I can get through to even the most difficult or unmotivated students.” Teachers who agree more with the first statement exhibit an external locus of control indicating that they believe their efforts are trumped by forces external to them, such as the influences of home, family, community, peers, media, and the like. Teachers in agreement with the second item, however, exhibit an internal locus of control that suggests confidence in their abilities to overcome these external influences.

While the locus of control measures formed the basis for the teacher efficacy construct, much of the subsequent research was derived from Bandura’s (1977) social cognitive theory. Bandura identified both self-efficacy and outcome expectancy as two crucial components of motivation and predictors of future behavior. Self-efficacy refers to an individual’s belief in his or her ability to act in a manner that will produce desired outcomes. Put another way, it involves one’s sense of competence in given situations. Outcome expectancies refer to an individual’s prediction of the likely consequences of his or her actions. Gibson and Dembo (1984) attempted to build upon the locus of control measures of teacher efficacy while applying Bandura’s theory to this construct. They developed a 30-item Teacher Efficacy Scale (TES) that they argued revealed two components of teacher efficacy: personal teacher efficacy (PTE) and general teacher efficacy (GTE). Gibson and Dembo (1984) asserted that PTE is related to Bandura’s notion of self-efficacy in that it refers to a teacher’s belief about his or her abilities as a teacher, and that GTE is a form of Bandura’s outcome expectancy because it refers to a teacher’s belief about “the degree to which the environment can be controlled, that is, the extent to which students can be taught given such factors as family background, IQ, and school conditions” (Gibson & Dembo, 1984, p. 570). Thus, a teacher’s efficacy beliefs were posited to be a combination of perceptions of personal influence and perceptions of the influence of factors external to the classroom.

Researchers have suggested through numerous studies that a teacher’s sense of efficacy has an effect on teachers’ enthusiasm, commitment, effort, goals, planning, organization, openness to new ideas, willingness to experiment with new methods, persistence in the face of obstacles, resilience with a student who is struggling, and, perhaps most importantly, the likelihood that a teacher will stay in teaching (Tschannen-Moran, Hoy, & Hoy, 1998; Tschannen-Moran & Hoy, 2001). Furthermore, teacher efficacy has been found to effect student achievement, student motivation, and student’s own sense of efficacy. The measurement of teacher efficacy, though, has been problematic and the subject of much debate. Specifically, researchers have noted problems with the TES such as the two-factor structure, the internal-external orientation of items, the confounding effects of the negative or positive phrasing of items, and the lack of specificity of teaching tasks described by TES items. Guskey and Passaro (1994) investigated the effect of the internal versus external orientation of items on the TES. The internally oriented items are those that used the first-person referent I and were designed to measure PTE, whereas the externally oriented items used the third-person referent teachers and were designed to measure GTE. They concluded that, “Contrary to these earlier studies, we could find no evidence to indicate the distinction between these two dimensions relates to personal efficacy versus teaching efficacy. Instead, our results indicate the difference to be an internal versus external distinction… The teachers we surveyed … did not distinguish between their personal ability to affect students and the potential of teachers in general” (Guskey & Passaro, 1994, pp. 636-637). Deemer and Minke (1999), however, argued that Guskey and Passaro’s study neglected to examine whether the positive and negative orientation of the items biased the factor structure of the TES. They conducted a study in which they revised the items on the original TES to reflect both positive and negative orientations across internal and external influences on teaching. Based on their findings, Deemer and Minke (1999) concluded that, “teacher efficacy, as measured by the TES, is actually unidimensional. That is, in the absence of wording confounds, those items appear to tap primarily an internal dimension similar to the personal teaching efficacy factor” (p. 8).

Another factor of the teacher efficacy construct that has received attention is the level of specificity at which instruments that attempt to measure the construct aim. Bandura (1997) asserted that teacher efficacy is specific to the teaching task and subject matter. For example, a teacher might feel efficacious when teaching grammar, but not feel efficacious when teaching mathematics. Moreover, a mathematics teacher may feel confident teaching algebra, but not calculus. Deemer and Minke (1999) criticized the TES as being a global instrument that tends to “decontextualize efficacy judgments and lack close relationships to specific teaching tasks” (p. 9). They went on to claim that, “Instruments that separately assess teachers’ perceptions in specific domains of teaching can be expected to tap the variations in efficacy judgments and increase the predictive power of efficacy perceptions” (p. 9). Several modified versions of the TES have been used for context-specific purposes such as science teaching (Riggs & Enochs, 1990; Ritter, Boone, & Rubba, 2001), special education (Coladarci & Breton, 1997; Brownell & Pajares, 1999), and computer education (Milbrath & Kinzie, 2000; Ross, Hogaboam-Gray, & Hannay, 2001). The results of this research are mixed, but some of the modified instruments, particularly Riggs and Enoch’s (1990) Science Teaching Efficacy Belief Instrument (STEBI), appear to measure teacher efficacy beliefs within the specific domain with more validity (Pajares, 1997; Bandura, 1997; Deemer & Minke, 1999).

The studies by Guskey and Passaro (1994), Pajares (1997), and Deemer and Minke (1999), as well as research conducted by Woolfolk and Hoy (1990) and Henson, Kogan, and Vacha-Haase (2001), have brought into question the use of the TES as designed by Gibson and Dembo (1984), and specifically the construct of general teacher efficacy (GTE). As a whole, this research suggests that GTE is not related to outcome expectancy, as Gibson and Dembo (1984) contended, and that the positive and negative phrasing of items in conjunction with the internal-external orientation of items has confounded the results of previous research using the TES and the teacher efficacy construct. It must be emphasized that this construct validation research has been conducted on the TES and that recent researchers assert that context-specific instruments that possess both positive and negative phrasing across both internal and external orientations are likely to produce more robust measurements of teacher efficacy.

The construct of teacher efficacy has clear relevance for character education. For character education to be effective, a teacher must believe in his or her own ability to build the character of students, as well as the ability of teachers in general to overcome negative influences from outside of the classroom. The character education literature suggests that many teachers lack such convictions (Lickona, 1993; Ryan and Bohlin, 1999; Berkowitz, 1999). Milson and Mehlig (2001) first applied the construct of teacher efficacy to character education with the development of the Character Education Efficacy Belief Instrument (CEEBI). Milson and Mehlig (2002) found that elementary teachers exhibited high levels of both personal teaching efficacy (PTE) and general teaching efficacy (GTE) for character education. They concluded that, “Elementary teachers feel confident in their abilities to serve as role models, to discuss issues of right and wrong with their students, and to use strategies that might lead to positive changes in students’ character. They also believe they can foster traits such as honesty, responsibility, courtesy, compassion, and respect” (p. 52). However, Milson and Mehlig (2002) did note a potential inconsistency in the efficacy beliefs exhibited by their sample. They stated, “Teachers …[seem] less confident with their ability to redirect students with certain shortcomings of character. In other words, teachers feel less efficacious when they are teaching precisely the students who need character guidance the most …We must question whether the global sense of efficacy that teachers seem to exhibit for character education is evident regardless of the type of students with whom the teacher works” (p. 52). Milson and Mehlig (2002) also examined the effect of a variety of teacher characteristics on teacher efficacy for character education. They noted no significant effects for age, level of education, grade level teaching assignment, or years of teaching experience, but they did find that teachers who earned their undergraduate degrees from private, religiously affiliated colleges or universities tended to have a greater sense of efficacy for character education than those who attended public or secular private institutions.

The present study builds on the work of Milson and Mehlig (2002) by administering the CEEBI to a sample of teachers from across the United States who teach a variety of subject areas, at a variety of grade levels, in a variety of communities, with a variety of training experiences. Specifically, the present study was designed to examine the level of teacher efficacy exhibited for character education by teachers in the United States and the effects of the following independent variables: 1) the type of institution of higher education attended by a teacher, 2) training in character education through university coursework, conference attendance, and staff development workshops, 3) the grade level at which a teacher teaches, 4) the subject area taught by a teacher, 5) the type of community in which a teacher’s students reside, and 6) the family socio-economic status of a teacher’s students.

The sample for this study consisted of 930 teachers who represent every state of the United States of America, including Guam and Puerto Rico. The sample was chosen from members of the Association of Supervision and Curriculum Development (ASCD). This organization was selected because it draws a nationwide membership of practicing teachers and the members teach a variety of subject areas at a variety of grade levels. It was also believed to be valuable to draw from an organization whose membership has a variety of levels of knowledge of and experience with character education.

The respondents consisted of 204 males, 716 females, and 10 respondents who did not indicate their genders. The respondents self-selected their racial/ethnic groups as follows: 3.2 percent African American, 1.0 percent Asian/Pacific Islander, 89.7 percent Caucasian, 2.7 percent Hispanic, 0.1 percent Native American, 0.3 percent “Multiple,” 0.9 percent “Other,” and 2.2 percent provided no indication of their ethnicity. The respondents reported their teaching experience as follows: 0.5 percent first year teachers, 1.9 percent with one to three years of experience, 6.0 percent with four to six years of experience, 8.3 percent with seven to ten years of experience, and 82.8 percent with more than ten years of experience. Finally, of the 923 respondents who completed the question, 87.7 percent reported their school type as “public school,” 7.3 percent reported teaching in a private school with a religious mission, 2.9 percent teach in a private, non-religious school, and 1.3 percent reported some “other” school type.

The instrumentation for this study consisted of Milson and Mehlig’s (2001) Character Education Efficacy Belief Instrument (CEEBI). The CEEBI consists of 24 statements to which the subject responds on a five-point Likert scale. The items were based on the two dimensions of Gibson & Dembo’s (1984) teacher efficacy scale (TES): personal teaching efficacy (PTE) and general teaching efficacy (GTE). Milson and Mehlig (2001) described the instrument development as follows:

Consistent with research on the teacher efficacy construct, the CEEBI was designed to achieve an appropriate level of specificity, with a balance of positive and negative phrasing, and a balance of internal and external locus orientation.

In their study of elementary teachers, Milson and Mehlig (2002) reported a bivariate correlation coefficient of .648 between PTE and GTE and reliability coefficients of a = .8286 for PTE and a = .6121 for GTE. For this study, a bivariate correlation coefficient of .690 between PTE and GTE and reliability coefficients of a = .8064 for PTE and a = .6545 for GTE were calculated. These results suggest that the instrument has maintained across administrations similar and acceptable levels of internal consistency as well as correlation between the scales.

A mailing list of 5002 members of the Association for Supervision and Curriculum Development (ASCD) was purchased from MGI Lists. The initial criterion for selection was ASCD membership data indicating that the member was employed as a teacher. The pool of members meeting this criterion was approximately 20,000. From this number, 5002 were computer-selected so that representation from all states in the United States was achieved. Each member on the mailing list received by mail a copy of the CEEBI, a cover letter requesting his or her participation, and a business reply, postage-paid envelope for returning the survey to the researcher. Each recipient was assigned a code number that was used to track non-respondents. As the surveys were returned, the data was coded on a manual scoring sheet and transferred to an SPSS spreadsheet. Approximately six to eight weeks after the survey was mailed, a follow-up letter was sent to fifty percent (N ~ 2000) of the non-respondents encouraging them to respond.

The final response of 1340 indicates a response rate of 26.78 percent. Of the 1340 respondents, 183 surveys were returned blank or only partially completed. Many of these respondents indicated that they were no longer teachers. Additionally, 227 respondents who completed the CEEBI indicated that they did not currently teach K-12 students because they had retired, moved to an administrative position, moved to higher education, or for similar reasons. These 227 respondents were excluded from the data analysis since they were not members of the target population of current K-12 teachers. These exclusions resulted in the final sample of 930.

A composite score for each respondent for GTE and for PTE was calculated by taking the sum of the twelve items on each scale. A respondent could thus receive a composite score of any number between 12 and 60. The mean of the composite scores for each scale was calculated to determine the relative position of the mean scores given the possible range. The mean composite scores of 46.89 (SD = 5.34) on the PTE scale and 42.44 (SD = 4.42) on the GTE scale are thus relatively high given the possible midpoint score of 36. The lowest PTE composite score was 28 (N = 1) and the highest was 60 (N = 2). The lowest GTE composite score was 26 (N = 1) and the highest was 56 (N = 1). These data indicate that teachers’ sense of efficacy for character education is generally positive. An exploration of the responses to specific CEEBI items, as well the statistical analyses conducted to investigate group differences should help to illuminate these findings.

The teachers’ responses to the CEEBI are presented in Table 1. The mean scores for each item can be used to determine the general level of efficacy exhibited for the item. With 5.00 representing the highest possible score and 1.00 representing the lowest possible score on each item, mean item scores between 1.00 and 2.99 are considered negatively efficacious, those between 3.00 and 3.99 are neither positive nor negative, and those scores above 4.00 are considered positively efficacious. Of the 24 items, teachers exhibited positive efficacy beliefs on eight items (1, 3, 4, 6, 7, 10, 19, 24), neither positive nor negative beliefs on 13 items (2, 5, 8, 9, 11, 12, 15, 16, 17, 18, 20, 21, 23), and negative beliefs on three items (13, 14, 22). These data indicate that the response to a majority of the items (54 percent) is categorized as neither positively nor negatively efficacious and that more items elicited positive responses than negative.

Of the eight items that elicited positive efficacy responses, five are PTE items and three are GTE items. The items with the three highest levels of positive response are PTE items: Item 3, “I am confident in my ability to be a good role model,” Item 1, “I am usually comfortable discussing issues of right and wrong with my students,” and Item 6, “I am usually at a loss as to how to help a student be more responsible.” The GTE items that garnered positive responses were Item 10, “Teachers who spend time encouraging students to be respectful of others will see little change in students’ social interaction,” Item 24, “Teachers who encourage responsibility at school can influence students’ level of responsibility outside of school,” and Item 4, “Teachers are usually not responsible when a child becomes more courteous.” The response to these items suggests that teachers feel exceptionally personally efficacious in the areas of serving as a role model, discussing moral issues, and developing responsibility in their students. They also appear to feel confident in teachers’ ability in general to encourage respectful, responsible, and courteous behavior. These results also indicate that teachers’ efficacy beliefs are not impacted by the personal versus general teacher efficacy distinction. They feel efficacious in both realms.

Of the three items that elicited negative efficacy beliefs, two were GTE items and one was a PTE item. In ascending order by mean score, these items were Item 22, “Teachers cannot be blamed for students who are dishonest,” Item 13, “Some students will not become more respectful even if they have had teachers who promote respect,” and Item 14, “When I have a student who lies regularly, I can usually convince him/her to stop lying.” Contrary to the items that elicited positive responses, all three of these items refer to students who exhibit problem behaviors. The other three items on the CEEBI that presented a situation in which the student might exhibit negative behaviors fell in the category of neither positive nor negative efficacy. These items were Item 2, “When a student has been exposed to negative influences at home, I do not believe that I can do much to impact that child’s character,” Item 11, “I am able to positively influence the character development of a child who has had little direction from parents,” and Item 16, “If responsibility is not encouraged in a child’s home, teachers will have little success teaching this trait at school.” All three of these items present a situation in which the student comes from a home environment that does not support positive character formation.

Statistical analyses were performed to investigate the effect of group differences. One-way analyses of variance and independent samples t-tests were conducted to examine the effects of various teacher characteristics on their PTE and GTE composite scores. The ANOVA results for Certifying Institution, Training, Community Type, Student SES, Grade Level, and Subject Area are presented in Table 2.

The respondents were asked to indicate whether they received their teacher certification from a “public college/university,” a “private, non-religiously affiliated college/ university,” a “private, religiously affiliated college/university,” or through “alternative certification (not through a college/university).” Of the 898 respondents who providedan indication of their certifying institution, 67.0 percent indicated “public college/university,” 7.8 percent indicated “private, non-religiously affiliated college/university,” 20.0 percent indicated “private, religiously affiliated college/university,” and 1.7 percent indicated “alternative certification (not through a college/university).” Due to the small number of respondents indicating “alternative certification,” the respondents in this category were eliminated from the ANOVAs. As presented in Table 2, ANOVAs on both PTE composite scores and GTE composite scores yielded no statistically significant differences among the groups.

The respondents were asked to respond “yes” or “no” to the question “Have you received any coursework or staff development in character education?” Of the 916 respondents who answered this question, 63.5 percent responded “yes” and 36.5 percent responded “no.” Two independent samples t-tests were performed to investigate the differences between these groups on their PTE and GTE composite scores. The results for PTE (t = 5.699, p < .01) and for GTE (t = 5.086, p < .01) indicate that those respondents who have received coursework and staff development training in character education have significantly higher composite scores on each scale than those who did not receive training.

Of those respondents who indicated that they had received coursework or training, 109 indicated that they have received “undergraduate coursework that addressed character education,” 180 have received “graduate coursework that addressed character education,” 286 have “attended a character education session at a conference,” 442 have “attended a staff development workshop on character education,” and 84 indicated some “other” form of training. In order to investigate the differences among respondents who received no training and those who received various types of training, the responses were grouped together to create five groups:

1. no training,

2. university-based coursework (undergraduate and graduate),

3. non-university-based training (conference session or staff development workshop),

4. some combination of university-based and non-university-based work, and

5. all four types of work.

The results of the one-way analyses of variance indicated significant differences in mean scores among groups on both of the scales (see Table 2). A Tukey’s HSD analysis was performed to investigate the source of the group differences. The results indicate that on the PTE scale those who received non-university-based training (p = .001), a combination of training (p < .001), or all types of training (p = .009), had significantly higher composite scores than those who received no training or coursework. The results for the GTE scale indicate that those who received non-university-based training (p < .001) or a combination of training (p < .001) had significantly higher composite scores than those who received no training or coursework. Interestingly, on both scales the scores of those who received university-based coursework did not differ significantly from those who received no training at all.

The recipients of the CEEBI were asked, “What grade level do you teach this year?” and were prompted to “choose all that apply” from a list of all grades from pre-kindergarten through grade 12 and “other.” As explained previously, those who selected “other” were excluded from all of the data analyses. Based on their responses, the respondents were grouped into the categories of “PK-4—elementary” (26.3 percent), “5-8—middle grades” (31.6 percent), “9-12—high school” (28.2 percent), “PK-8—elementary and middle grades” (8.0 percent), “5-12—secondary” (3.4 percent), and “PK-12—all levels” (1.1 percent). Given the small sample sizes in some groups, only those in the “PK-4—elementary,” “5-8—middle grades,” “PK-8—elementary and middle grades” and “9-12—high school” categories were included in the analysis of variance. The results of the one-way ANOVAs for PTE and for GTE indicated that the composite scores of the respondents differed significantly based on the respondents’ grade level teaching assignment (see Table 2). A Tukey’s HSD analysis was performed to investigate the sources of group differences. The results indicate that those who teach high school have significantly lower PTE and GTE scores than those who teach grades PK-4 [(PTE, p < .000); (GTE, p = .005)] and those who teach grades PK-8 [(PTE, p = .008); (GTE, p = .023)]. Additionally, on the GTE scale, those who teach grades 5-8 had significantly lower scores than those who teach grades PK-4 (p = .048).

The recipients of the CEEBI were asked, “What subject area do you teach most of the time this year?” and were provided with a list of subjects from which to select. The 822 respondents who completed this question selected their subject areas as follows: art—0.5 percent, foreign language—2.3 percent, language arts—19.7 percent, integrated curriculum—8.5 percent, mathematics—10.1 percent, music—1.2 percent, multiple subjects equally—14.5 percent, physical education— 1.0 percent, science—7.6 percent, social studies—7.1 percent, special education—4.3 percent, and other—11.6 percent. Given the small size of some of the groups and the broad diversity of those in the “other” category, the ANOVAs included only those who selected language arts, mathematics, science, social studies, integrated curriculum, and multiple subjects. The significant differences resulting for both GTE and PTE were examined using Tukey’s HSD. The results of the post hoc analysis indicate that those who teach science have significantly lower PTE composite scores than those who teach an integrated curriculum (p = .005) or those who teach multiple subjects (p = .042). Furthermore, science teachers GTE scores were significantly lower than those teaching multiple subjects (p = .042). Additionally, social studies teachers PTE scores were significantly lower than those who teach an integrated curriculum (p = .010).

Given that many elementary schools are not departmentalized, it is possible that few elementary teachers consider themselves to be science or social studies teachers. Furthermore, few high school teachers are likely to teach an integrated curriculum or multiple subjects equally. A cross tabulation between grade level and subject area was conducted to investigate this possibility. It was found that all subject areas were represented at each grade level, but that most of the science and social studies teachers were high school teachers and most of the integrated curriculum and multiple subjects teachers were at the elementary level. The subject areas language arts and mathematics were better represented across grade levels than science or social studies. Given that no significant differences were found among the language arts or mathematics teachers, it is likely that the differences for science and social studies teachers are not a function of grade level alone.

The recipients of the CEEBI were asked, “In which type of community do the students who attend your school live?” and were prompted to “choose all that apply” from “urban,” “suburban,” and “rural.” They were also asked, “How would you describe the socio-economic status (SES) of the students who attend your school?” and were prompted to “choose all that apply” from “low family income,” “middle family income,” and “high family income.” The respondents reported their community types as: 23.5 percent urban only, 44.4 percent suburban only, 23.5 percent rural only, 3.0 percent urban and suburban, 0.5 percent urban and rural, 3.1 percent suburban and rural, and 1.4 percent taught students from all three types of communities. The respondents reported their students’ family SES as: 25.1 percent low income only, 32.6 percent middle income only, 9.9 percent high income only, 15.4 percent low and middle income, 3.1 percent middle and high income, 0.5 percent low and high income, and 12.9 percent all three levels of family SES. The ANOVAs for PTE and GTE on community type included only those who indicated urban only, suburban only, and rural only. The ANOVAs for PTE and GTE on family SES included low only, middle only, high only, low and middle, and all levels. The ANOVA results presented in Table 2 indicate that neither community type nor family SES have a significant impact on the PTE or GTE scores of these teachers.

The results of this study generally support the findings of Milson and Mehlig’s (2002) study of elementary teachers’ sense of efficacy for character education. In both studies, teachers overwhelmingly agreed with statements such as, “I am usually comfortable discussing issues of right and wrong with my students,” and they consistently disagreed with statements such as, “I am usually at a loss as to how to help a student be more responsible.” Yet in both studies, teachers tended to agree with the statement, “Some students will not become more respectful even if they have had teachers who promote respect,” and disagree with the statement, “When I have a student who lies regularly, I can usually convince him/her to stop lying.” Teachers in both studies were also less than positive, although not entirely negative, about statements such as, “I am able to positively influence the character development of a child who has had little direction from parents,” and “If responsibility is not encouraged in a child’s home, teachers will have little success teaching this trait at school.” Taken together, both studies suggest that teachers appear efficacious for character formation when presented with statements that do not indicate the nature of the students’ character, yet they doubt their abilities to change the character of a disrespectful or dishonest student and are only slightly more confident in their abilities to overcome negative influences from the students’ homes.

One might argue that this research indicates that teachers feel efficacious about character education only when they are teaching “good kids.” Some critics of character education have argued that the initiative suffers from a “fix-the-kids” mentality that is derived from a politically conservative belief that social problems can be overcome through the exercise of good character (Kohn, 1997; Purpel, 1997). These critics contend that character education too frequently focuses on indoctrinating obedience and that the focus on good behavior distracts from the more suitable goal of promoting empathy or compassion. Following this line of thinking, one might interpret the results of this study as indicating that teachers doubt their ability to be successful with character education when they are faced with the task of teaching obedience, honesty, or responsibility to youth who are not already obedient, honest, and responsible. Others have countered that the portrayal of character education as ideologically-biased and narrowly conceived to “fix-the-kids” is a misunderstanding of the comprehensive nature of character education and a reaction to the inappropriate implementation of the initiative in some circumstances (i.e., Lickona, 1998; Milson, 2000, 2002). Although there may be some merit to the critique of character education offered by Kohn (1997) and Purpel (1997), I will argue that the results of this research provide direction for teacher educators rather than a rationale for abandoning the effort.

One primary purpose of this research was to identify teacher characteristics and components of character education that might undermine teacher efficacy for character education with the goal of providing direction to teacher education efforts. I use the term teacher education to refer to both university-based and non-university based development of teachers’ knowledge, skills, and dispositions. It seems that one objective for teacher education should be pursuit of questions such as, “how do you teach a dishonest, irresponsible, disrespectful young person from a morally bankrupt home environment that honesty, responsibility, and respect are important to personal happiness, the common good, and life as a citizen in a democratic society?” Wheatley (2000, 2002) has argued that teacher efficacy doubts theoretically may support educational reforms. He contends that positive teacher efficacy may lead teachers to feel content about their teaching, and thus unlikely to see the benefits of change. On the other hand, those teachers who feel doubt about their teaching efficacy are more likely to feel the disequilibrium that would foster reflection, motivation to learn, collaboration, and willingness to change. The evidence that teachers doubt their efficacy when working with potentially troublesome students then might in fact make them ripe for discussions related to such questions. Wheatley’s theoretical assertions and their implications suggest a need for further research. Most researchers employing the teacher efficacy construct, in accordance with Bandura’s social cognitive theory, view a low level of teacher efficacy as a source of low motivation.

Productive discussion of teacher efficacy doubts in the realm of character education should develop teachers’ understanding of the influence of culture, socio-economic status, peers, media, violence, drugs, and sex in the lives of today’s youth. Yet such discussions should also avoid over-emphasizing these influences and external factors to the point of accepting vicious, as opposed to virtuous, behavior as culturally or economically determined. According to the results of this study, teachers’ efficacy beliefs for character education are not significantly impacted by their students’ socio-economic status and community type. This suggests that the teacher in a middle income, urban school views her ability to foster character development in a similar way to the teacher in a low income, rural area. Both of these types of communities and all socio-economic levels contain young people with and young people without good character. One task for teacher educators then is to help teachers think about the challenge and explore methods for reaching those students who lack good character regardless of, or perhaps despite, where they live.

A second task for teacher educators is to consider how preparation for character education is different for elementary versus secondary teachers. The results of this study indicate that high school teachers feel less efficacious for character education than elementary school teachers. Although the importance of role models and the need to develop a positive moral school environment does not change over the years, the classroom activities associated with character education are necessarily different at the adolescent age level than they are for pre-pubescent children. For this reason, the character education programs of elementary schools are quite different from those in high schools. Generally speaking, early elementary level character education activities focus heavily on the reading of children’s literature from which moral lessons can be gleaned. Upper elementary and middle school character education activities tend to focus on developing an understanding of the meaning of key virtues through skits, readings, teacher-directed discussions, art, and similar focused, “hands-on” lessons. At the high school level, character education activities tend to become less teacher-directed as students engage in peer mediation, student government, and discussions of the “gray areas” associated with applying virtues to daily life when faced with the challenges of adolescence. Elementary and middle school teachers are understandably more comfortable with the activities associated with character education because these activities are similar in form to those used throughout the day.

The rationale for character education may also resonate more strongly with elementary and middle school teachers who tend to view character formation as an important component of their jobs. High school teachers, however, are more likely to view their primary role as the delivery of content knowledge from an academic discipline. This is not to say that high school teachers necessarily reject the need to continue the formation of character in their students, but they are more likely to see their primary responsibility as teaching history rather than teaching kids to be virtuous. There are at least three explanations, then, for the finding that high school teachers feel less efficacious for character education: 1) high school teachers do not feel confident implementing the activities associated with high school level character education, 2) high school teachers do not feel that character education is part of their job, and 3) high school teachers doubt their ability to have an impact on students’ character. All three of these explanations could appropriately lead to teacher education efforts to build knowledge, confidence, skill, and motivation among high school teachers for the task of character education.

Related to the task of identifying appropriate teacher education curricula for secondary teachers, is the need to consider how subject area specialization at the secondary level influences teacher efficacy beliefs for character education. Another finding of this study is that science and social studies teachers scored significantly lower on the CEEBI than teachers of integrated curricula and teachers of multiple subject areas. One might argue that this result occurred because many of the respondents who identified themselves as science or social studies teachers are also secondary level teachers and many of those who identified themselves as integrated curriculum or multiple subjects teachers were also elementary teachers. The fact that high school language arts and math teachers, though, did not differ significantly from elementary teachers of these subjects or from teachers of integrated curriculum or multiple subject areas clouds the explanation that this is an elementary versus secondary effect rather than an effect of subject area specialization. Thus this finding cannot be given much emphasis and the effect of subject area specialization in science or social studies on teachers’ efficacy beliefs for character education should be the subject of future research that focuses on secondary teachers.

A third task for teacher educators is to determine the appropriate content of pre-service and in-service preparation for character education. Teacher education for character education is often discussed as primarily the responsibility of institutions of higher education through programs in colleges and schools of education (i.e., Jones, Ryan, and Bohlin, 1998; Williams & Schaps, 1999; CAEC/ CEP1999). The results of this study indicate that teachers who have received staff development training or who have attended conferences in which character education issues were presented have higher levels of efficacy than those who received coursework through their undergraduate or graduate programs. More intriguing is the finding that those who received university-based coursework do not differ significantly from those who received no training at all. One explanation for this result is that in-service training is better and more effective than university-based coursework. Another explanation is that staff development and conference presentations on character education may tend to be more focused on specific practical implications for implementation within a school program, whereas university-based coursework is likely to cover character education, if at all, as one of many topics within a course and with less attention to specific implementation. Whether it is a matter of quality or a matter of focus, this study provides further evidence to support those who suggest that university-based teacher education programs are not meeting the task of preparing teachers for character education.

Most would argue that teacher education should be a constant process that begins when a teacher enters a teacher certification program and continues throughout a teacher’s career. Quality preparation for character education should exist at both the pre-service and the in-service levels. The results of this study suggest that in-service teacher education is producing a positive effect on teacher efficacy for character education. Clearly, university-based programs need to direct more attention toward improving the quality and/or focus of teacher education in this area. The type of program in which such training is likely to be most effective remains a question for future research, though. Milson and Mehlig (2002) found that teachers who attended private, religiously affiliated colleges and universities had significantly higher scores on the CEEBI than those who attended public or secular-private institutions. They speculated, as Jones, Ryan, and Bohlin (1998) did, that this might be a result of a more explicit, mission-level commitment to the moral character formation of students at religiously affiliated institutions. The results of this study do not confirm the findings of Milson and Mehlig on this issue. The type of institution of higher education at which a teacher earned teacher certification did not have a significant impact on the scores of teachers in this study.

Bandura (1986, 1997) theorized that there are four sources of efficacy expectations: mastery experiences, vicarious experiences, social persuasion, and physiological/emotional states. Each source provides information to a teacher that can serve to either lower or raise his or her efficacy beliefs. A mastery experience that would raise efficacy beliefs is one in which a teacher perceives that his or her performance was successful and that the success can be attributed to his or her own ability or effort rather than to random luck or the intervention of another person. Efficacy beliefs may also be enhanced when a teacher observes another teacher perform well. The efficacy enhancement through such vicarious experiences is particularly pronounced when the observer identifies with the model. Social persuasion from trusted, credible, and expert sources, such as supervisors, mentors, or colleagues, can also enhance efficacy beliefs when it motivates the teacher to act and to be persistent. Finally, physiological and emotional states, such as exhilaration or anxiety, can affect performance and can contribute to perceived competence or failure. One promising avenue for future research on teacher efficacy for character education is the investigation into questions such as: Which mastery experiences and which vicarious experiences at each level (elementary, middle school, high school) are most effective for enhancing teachers’ sense of efficacy for character education? What forms of social persuasion tend to enhance teachers’ sense of efficacy for character education? And which experiences at each level tend to result in efficacy-enhancing emotional and physiological states?

The premise of this research has been that improved teacher education for character education is desirable and necessary. This assertion is based on the widespread and considerable expectations placed on teachers across the United States to serve as character educators. For the last several years, researchers and character education advocates have noted and bemoaned the inadequate training provided for teachers in this area. The purpose of the research reported here was to apply the teacher efficacy construct to the field of character education in an attempt to guide teacher education efforts in character education. The results of this study suggest that teachers possess doubts about their abilities to provide character education for students who may lack good character. Furthermore, high school teachers, and perhaps teachers of science and social studies, seem to have greater levels of efficacy doubt related to character education than elementary teachers. This study also suggests that staff development training and character education conference sessions have a positive impact on teachers’ sense of efficacy for character education and that university-based coursework provides no significant boost in teacher efficacy. I have suggested three tasks for teacher educators based on the findings of this study:

1. provide opportunities for reading and discussion that help teachers think about the challenge of and the methods for reaching those students who lack good character,

2. design teacher preparation experiences that address the differences in philosophy, curriculum, and methods between elementary and secondary character education, and

3. identify successful approaches to both in-service and pre-service teacher education that improves the deficiencies apparent in university-based coursework.

Future research into these issues and an application to character education of Bandura’s theoretical model for sources of efficacy expectations would be valuable additions to the research literature in character education and teacher education.

This research was funded by a grant from the Baylor University Research Committee. I wish to express my gratitude to Sherie Carr, Krista Barrett, Karen Estes, and Terrill Saxon for their assistance with the data collection and analysis for this study.