The Relationship of Georgia s Rural Foreign Language Teachers Sense of Efficacy to Teacher Attrition

Foreign language teachers are in critical need in many parts of rural America. Using Bandura's conceptual framework of self efficacy teaching languages as a theoretical lens, the researchers created a scale to measure foreign language teacher efficacy and administered alongside a well-known efficacy survey to in-service rural teachers (N = 167) in Georgia. Data analysis indicates that the new instrument is psychometrically sound and there are two dimensions to language teacher efficacy: Content Knowledge and Facilitating Instruction. Positive correlations between the two surveys suggest that teaching languages requires more than just strength of content knowledge and FL teachers may need assistance engaging students. Additionally, it appears female novice Spanish teachers are more prone to attrition than teachers of other languages. This research holds implications for professional development opportunities as well as teacher preparation programs.

Nationally, the number of students enrolled in K-12 public schools in the United States (US) has been steadily increasing while the number of certified teachers willing to work in US classrooms has been decreasing. Such a phenomenon has contributed to a teacher shortage prevalent in many parts of the nation (American Association for Employment in Education, AAEE, 2008). Research on the shortage of teachers suggests a lack of consensus regarding the factors associated with the shortage. While Ingersoll (2001,2003) finds a revolving door of teacher attrition and turnover that helps explain the teacher shortage (Ingersoll, 2001(Ingersoll, , 2003, Darling-Hammond (2000) indicates that the shortage is exacerbated by a surplus of certified teachers who actively choose not to teach. Yet, others argue that a shortage of teachers in many parts of the country exists regardless of the available teaching pool from which to draw (AAEE, 2006;Fideler & Haselkorn, 1999), because some professionals tend to avoid employment in urban schools and small private schools. Further investigation reveals that an uneven distribution of teachers nationally appears to contribute to the current teacher shortage (Wilson, Darling Hammond, & Berry, 2001). Nevertheless, the literature clearly indicates there is a teacher shortage throughout the nation and among the areas of critical need are special education, bilingual education, math, science, and foreign languages (AAEE, 2008, Draper & Hicks, 2002National Center for Education Statistics, 2002).
While there is an abundant literature base describing the shortage of math and science teachers, there is a paucity of research discussing the lack of foreign language (FL) teachers, especially in rural schools. Such a finding is alarming because approximately half of the nation s 80,000 public elementary and secondary schools are located in rural areas or small towns, and nearly one in three of America's school-aged children attends public schools in rural areas or small towns (Johnson, 2003). Research indicates that in four states (Maine, Mississippi, Vermont, and West Virginia) the majority of the population lives in rural areas. Two other states, South Dakota and Arkansas, come very close to having most of their inhabitants residing outside of suburban areas (Beeson & Strange, 2003).
Characteristically, rural districts tend to have declining student populations, lower property value assessments, increased transportation expenses, a higher proportion of residents living in or near poverty levels compared to metropolitan areas, and difficulty attracting quality teachers (Dewees, 1999;Phillips, 2003).
The purpose of this research is to call attention to the lack of language teachers and investigate how rural FL teachers sense of efficacy plays a role in their decision to remain or leave the teaching profession at a time of critical shortage (Swanson, 2008). The authors first review the current situation facing FL teachers specifically and advance five factors that help explain the FL teacher shortage in the context of rural schools. They next describe the quantitative and qualitative methods and survey used (a) to assess rural FL teachers sense of efficacy teaching languages in Georgia and (b) to ascertain whether the FL teachers plan to remain in the profession. The article concludes with a discussion of the results and the implications from the findings.

Current State of Affairs for FL Teachers
Three decades ago researchers warned of a severe shortage of teachers in America (Boe & Gilford, 1992; Darling- Hammond, 1984;Haggstrom, Darling-Hammond, & Grissmer, 1988; National Academy of Sciences, 1987; National Commission on Excellence in Education, 1983) due to increased student enrollment and teacher attrition.
Unfortunately, the prediction was correct and there is a shortage of teachers in America's classrooms today (AAEE, 2006;Draper & Hicks, 2002; National Center for Education Statistics, 2002). Current research indicates that the shortage is not uniform across content areas and severe shortages are reported in special education, mathematics, science, bilingual education, English as a Second Language, and foreign language (AAEE, 2008; U.S. Department of Education, 2010). FL teaching positions are found to be the most difficult to fill, much more so than math, science, and special education (Murphy, DeArmand, & Guin, 2003). Moreover, rural schools continually have problems locating skilled teachers and now these schools are faced with a teacher shortage (Barley & Brigham, 2008, Darling-Hammond, Berry, & Thoreson 2001North Central Regional Educational Laboratory, 2003). Additionally, FL is an area currently facing a national shortage of teachers regardless of educational context (AAEE, 2009). Research on the shortage of FL teachers points to at least five factors that explain the shortage: retirement, attrition, increased enrollments, legislation, and perceptions of teaching (Swanson, 2008).

Retirement
The AAEE (2006) reported that 24% of elementary and 26% of secondary teachers in the US were at least 55 years of age in the late 1990s and that the same percentage of elementary and secondary teachers could be expected to retire between 2005 and 2010. Findings of the report indicated that if student enrollments remained constant, more than 24% of the teachers at each level would be needed in the next ten years. In a similar study, Radner (1998) found that 29% of the US population is considered part of the Baby Boomer generation and these individuals should be eligible to retire in the next ten years. Adding to Radner's work, Blair (1999) found that this generation of teachers who entered the profession 30 years ago are able to retire in masses (p. 22). Assuming that FL teachers are dispersed evenly from elementary to secondary contexts as well as in urban, suburban, and rural areas, a large number of FL teachers can be expected to retire soon. A factor equally serious as teacher retirement is teacher attrition.

Teacher Attrition
Nearly one-third of America's teachers leave the profession within their first three years of teaching, and almost half leave after five years (National Commission on Teaching and America s Future, 2002). Educators cite factors such as inadequate classroom management skills, large classes, work schedules, feelings of isolation in the classroom, social and disciplinary problems, low standardized test scores, serious financial budget issues, high dropout rates, higher absenteeism, and low salaries as reasons to leave teaching (Predmore, 2004;Schwartzbeck, Redfield, Morris, & Hammer, 2003;Strange & Silverman, 2005;Wilcox, 2004). For those people who enter teaching through an alternative route such as emergency certification, the attrition rate can be as high as 60% (Darling-Hammond, Berry, & Thoreson, 2001). Part of the attrition problem in rural schools has been associated with larger school districts, primarily in urban areas, recruiting rural teachers to suburban and urban schools by offering them substantial salary increases and better job benefits (Rebore, 2004).
While the research highlighting the attrition rates for FL teachers from a national perspective remains scant, researchers have reported double digital attrition statistics in two southern states. FL teacher attrition in North Carolina (22%) and Georgia (11%) was found to be higher than the rate of attrition for teachers in other content areas (15-18%) (Georgia Professional Standards Commission, 2006;Konanc, 1996). In North Carolina, FL teachers had the highest rate of attrition after the second year (21%), the fifth year (38%), the tenth year (49%), and the fifteenth year (57%) (Konanc, 1996). Further west in Wyoming, Stowers (2004) (Rural School and Community Trust, 2006). The largest increase has been found to reside in non-traditional immigrant states, particularly in the Southeast. In Georgia, enrollments of Mexican students swelled from 4% in 1990 to a 61% in 2005 in the rural Dalton district alone (Teague, 2007).

2)
Regrettably, these rules effectively exclude about 75% of the nation s rural and small-town schools and show racial, regional, and poverty bias by excluding some of the highest need rural schools in the country (Rural School and Community Trust, 2004 districts shall provide instruction in foreign language to all students in kindergarten through grade two in accordance with standards promulgated by the state board of education [sic] (School Improvement Act, 1999, p. 3). Later, the Wyoming House Bill 0170 extended the 1999 legislation to include grades 3-6. This new legislation requires elementary educators to teach an additional subject in an area for which many are not certified (Swanson & Moore, 2006).

Perceptions of Teaching
Lastly, teaching has been described for years as a dead end job with perceived low status, low salaries, lack of control over how schools are run, many classroom discipline issues, and ineffective administrative support leading to a lack of induction and mentoring (Boles, 2000;Boser, 2000;Brunetti, 2001;Stanford, 2001;Weld, 1998). For future educators, such a dismal description can serve to discourage adolescents and career changers from investigating and pursuing a career in teaching any subject, including FL. The same appears to be true for novice in-service educators because many times these individuals are assigned the most challenging duties with little to no professional support. These individuals tend to have fewer successes and a sense of failure may drive them out of the classroom (Ladson Billings, 2001). For rural districts, the perceptual problems appear more pronounced. Schwartzbeck, Redfield, Morris, and Hammer (2003) reported that the perception of low salaries, social isolation, and geographic isolation made it extremely difficult to attract and retain teachers in rural schools. Taken collectively, these five factors help explain the shortage of FL educators. The purpose of this research was to explore the role of an additional factor, teacher self efficacy, in FL teacher attrition.

Conceptual Framework
Efficacy is defined as an individual s belief or conviction that he or she can influence how well students learn, even those who may be difficult or unmotivated (Guskey & Passaro, 1994). It influences the personal decision to remain working in the classroom or not. Self efficacy is grounded in the theoretical framework of social cognitive theory, underscoring the notion of human agency,  Thompson & Daniel, 1996), the researchers used multiple criteria for selecting the number of factors: (a) Cattell s (1966) scree test, (b) Kaiser s (1960) rule for Eigenvalues greater than one, (c) parallel analysis (Horn, 1965;Turner, 1998), (d) the percentage of common variance explained by each factor using the weighted reduced correlation matrix, and (e) the interpretability of the rotated factors. Parallel analysis was chosen because it has been shown to be among the most accurate methods for determining the number of factors to retain (Zwick & Velicer, 1986) and generally superior to the scree plot and Eigenvalue greater than one rule (Henson, 2001a, p. 14).
The researchers submitted the data to principal axis factor analysis with a Direct Oblimin oblique rotation because items were assumed to be related. Factor analysis of the TSES items indicated that three factors with Eigenvalues greater than one were present, accounting for 74.33% of the total variance. Eigenvalues for the subsequent factors ranged from 0.13 to 0.70. The three factors were those identified by Tschannen-Moran and Woolfolk Hoy (2001). The first factor (Classroom Management) accounted for 51.95% of the variance while the student engagement and instructional strategy factors accounted for 11.75% and 10.63% of the total variance respectively. The overall commonalities for the three factors ranged from .58 to .81 (Classroom Management,(76)(77)(78)(79)58 .80;and Instructional Strategy,. In order to follow suggestions found in the literature about interpretation of the factors (Graham, Guthrie, & Thompson, 2003), both factor pattern and factor structure coefficients were interpreted.
Inspection of the matrices confirmed that the three factors were present and that each survey item loaded on only one of the three factors.
A second factor analysis was conducted next on the FLTES items only and two factors emerged with Eigenvalues greater than one, accounting for 68.01% of the variance. The remaining factors had Eigenvalues from 0.14 to 0.79. Here, 45.02% of the variance was explained by the Teacher as Facilitator factor and 22.99% of the variance was explained by the Content Knowledge factor. Overall commonalities for these two factors ranged from .35 to 73 for the Teacher as Facilitator factor and from .68 to 82 for the Content Knowledge factor. Again each of the items loaded cleanly on one of the two factors when examining both the structure and pattern matrices.
Once the factors were identified, the researchers conducted correlation analysis and found positive coefficients ranging from 60 to 63 indicating that the three subscales were intercorrelated. Further, the three subscales were highly correlated with the instrument itself (r=.81 to .89). These results were similar to those reported by Tschannen-Moran and Woolfolk Hoy (2001), which support their earlier findings. To examine the construct validity of the FLTES, the researchers conducted correlation analysis (see Table 1) between the FLTES and the TSES and found a positive relationship (r=.67, p < .001), which suggests that not only did the FLTES measure the construct of teacher efficacy, but it also measured a different aspect of teacher efficacy not assessed by the TSES.
Afterwards, the researchers conducted correlation analysis between the subscales of the two instruments and found the FL

Group Differences
Noting that mean differences in perceived efficacy between the two genders and for those who had earned graduate degrees were discovered, analyses of variance" (ANOVA) were conducted to evaluate the relationship between the other independent variables and the items from the FLTES and the TSES. No significant differences in efficacy were found for the variables of study abroad, highest level of degree attained, and future vocational plans. However, statistically significant differences were found for two items of the FLTES for gender. The ANOVAs indicated that males had more confidence helping students to learn at the highest level of language, F(1,162) = 4.48, p < 05, n°= .02, and more perceived confidence to use the language they teach, F(1,162) = 4.92, p < 01, n =.02 than did females.

The Identity of Teachers Leaving the Profession
To answer the second research question about the identity of the FL teachers planning to leave the profession in terms of self-efficacy and the demographic data, the authors began by computing frequencies to see how many FL teachers planned to remain in the classroom the next year and how many planned to leave the profession. Eleven teachers stated that they planned to leave the profession at the end of the academic year. Nine of these individuals reported having a clear renewable teaching certificate. One was going to retire after 30 years of teaching and ten stated that they were going to leave teaching. Of those ten stating that they intended to leave the teaching profession, all were female Spanish teachers, two-thirds (67%) of whom were between the ages of 24 and 27, held a bachelor's degree only, and were in their first five years of teaching.
To investigate differences in perceived efficacy between teachers who stated that they were going to remain in the profession and those who reported they were quitting teaching, the researchers conducted an ANOVA. While the efficacy tended to be lower on all items for the people leaving teaching than those remaining, no statistically significant differences were found between the two groups.
Such a finding can be attributed to such a small group of individuals intending to leave the profession.

Qualitative Comments
The final question of the survey requested participant comments. The researchers followed Patton s (1990) suggestion whereby "the first decision to be made in analyzing interviews is whether to begin with case analysis or cross-case analysis (p. 376). The authors began with cross-case analysis of the interviews, using a modified version of the constant comparative analysis (Glasser & Strauss, 1967) to group answers and make connections to common questions. Data analysis of the comments from 24 participants revealed themes of low pay, lack of support from parents, administrators, and the government, feelings of excessive work, and concern about too many false beginners (i.e., native Spanish speakers) in introductory level Spanish courses.
Feelings of earning low salaries were expressed by the majority of those individuals who commented. Males chose to comment on this issue more than females. Yet, collectively, participants expressed feelings of professional depreciation by administrators and colleagues in the tested areas . A 37 year-old male veteran Spanish teacher stated, I have two graduate degrees, I'm fluent in two languages, and because I'm a not math or language arts teacher, I don t seem to matter. Similar feelings were expressed from a 53 year-old male who mentioned that FL teaching is relegated to an inferior position by people who should know better.
Along similar lines, several participants (all women) cited a serious lack of support from various entities. Three veteran female teachers stated that they felt they were the only ones who care about the students success in the classroom. One stated that she was trying different methods to get students to value learning but she felt the students and parents just don't care so why should I?"Another noted that she felt a lack of support for FL instruction from both Additionally, findings from the present study offer new information about rural FL teachers specifically. Each of the ten who reported that they were going to quit teaching was a novice female educator. Even although only ten educators in the sample stated they were going to leave teaching, it is important to remember that these individuals represent 6% of the sample, which is significant for two reasons. First, such findings are important because they help focus attention from a rural Georgia perspective when looking at the 11% rate of attrition reported by the Georgia Professional Standards Commission (2006).
Second, and possibly even more important, if each of the ten teachers who stated she would be leaving at the end of the school year had an average of 20 students per Spanish class and taught six classes per day, 1,200 students would be affected by such attrition. Such findings clearly warrant further study in this area to help clarify any differences between those educators who decide to remain or to leave the profession. Additional research is required specifically focusing on the number of false beginners in the classroom (students who are native speakers of the FL being taught, yet may need instruction in English as a foreign language); how those individuals affect language teachers efficacy perceptions, and the difficulty of encouraging students to value learning. Such research could include participants from other areas, such as North Carolina, that have higher rates of teacher attrition than Georgia in order to discover other factors associated with high attrition rates.