2:00-2:50 p.m. HIGHER EDUCATION (Discussion Session) Salon A

Presider: Michele Jarrell, The University of Alabama at Birmingham

TRANSITIONAL ISSUES OF UNTENURED FACULTY

Carol A. Mullen and Sean A. Forbes, Auburn University

             A needs assessment for mentoring was conducted to evaluate the personal/professional needs of junior professors not being met by existing academic structures. The assessments identified critical issues of transition that can be used for developing effective mentoring programs. This study was based on the reflective and critical feedback of untenured faculty working in the United States, Canada, and Australia. The results can be used for developing effective mentoring programs.
             Testimonials of untenured faculty (Boice, 1992; Ducharme, 1993; Tierney, 1997) highlight the vulnerability of untenured faculty as an occupational group arising from a "crisis of professional self-identity" (Nixon, 1996, p. 5). An understanding of this issue was approached from the Carnegie Foundation's belief that effective mentoring offers the best preparation for faculty socialization and evaluation (Glassick, Huber, & Maeroff, 1997).
             Untenured faculty were defined for the purpose of this study and the survey as faculty who hold tenure-track and adjunct posts in research universities. Sixty untenured faculty members responded to a survey designed to elicit personal reflections on their experience of being socialized during the pretenure years. The survey centered on mentoring assistance, preparation, interactions with other faculty, unexpected connections, and informal channels of information. The data were coded using a comparative thematic technique for analysis (Miles & Huberman, 1994).
             Formalizing a mentoring process that responds to everyone's needs is difficult. Also, informal collegiality cannot be mandated. However, the data included recommendations for improving untenured faculty socialization in three areas: (1) defining the criteria for gaining tenure, (2) promoting collegiality as cooperation rather than competition, and (3) developing awareness of the academic power structure. For example, many respondents called for clarifying conversations with departmental/campus leaders about criteria for tenure.

MAKING THE MOVE: THE TRANSITION FROM GRADUATE STUDENT AT A PH.D.-GRANTING UNIVERSITY TO NEW FACULTY AT A SMALL, PRIVATE, LIBERAL ARTS COLLEGE

Vicki A. Wilson, Wilmington College

             Making the transition from graduate assistant to new faculty member is often difficult. Making the transition from graduate student at a Ph.D.-granting university to new faculty member at a small, private, liberal arts college in a remote location is even more difficult, involving change in both professional and personal arenas.
             In this study, 42 faculty members in their first, second, or third year of employment at small, private, liberal arts colleges in a nine-college consortium centered around eastern Ohio were asked to describe their initial impressions of faculty and administration and to reflect on those things that helped or hindered their transition into the college community and the local community. They were also asked whether or not they intended to begin their careers at small colleges and what surprises they encountered in their new positions. Finally, they were asked to make suggestions about how their colleges and communities could ease the difficulties encountered in their first years as faculty members.
             What emerged was a picture of earnest young professionals, most sincerely committed to teaching and many attracted to small colleges because of their own experiences as students, struggling to figure out the required balance of teaching, research, and service while making long commutes or trying to create satisfying lives in rural communities in which they feel they have little in common with their neighbors. Their suggestions included improving the pay scale, reducing course loads to allow for research and service, making candidates and new faculty aware of the abilities and expectations of students, and resolving issues that cause resentment and disgruntlement among the more senior faculty.

VALIDATION OF SITUATIONAL JOB SATISFACTION THEORY FOR PART-TIME FACULTY

Hae-Seong Park, University of New Orleans

             Over the past 30 years, American colleges and universities have markedly increased their reliance on part-time faculty instruction. Although job satisfaction of part-time faculty is seldom researched, job satisfaction has long been a well-studied concept in organizational theory. The result of a recent meta-analysis (Thompson, Mcnamara, & Hoyle, 1997) showed that Situational Model of Job Satisfaction was preliminarily supported through a synthesis of research findings. The purpose of this study was to examine the validity of the situational model of job satisfaction for part-time faculty.
             The data employed in this study were drawn from the 1992-93 National Study of Postsecondary Faculty (NSOPF-93). Faculty members who had part-time appointments and who marked that teaching was their primary responsibility were drawn for the analyses. The resultant sample size (n=5,703) was 22% of the total sample. Structural equation modeling using AMOS was employed to analyze the data.
             The findings indicated that the situational model, including a combination of typical task characteristics, organizational characteristics, and individual characteristics impact on job satisfaction, was not valid for part-time faculty. Contrast to theory task variables were not significant, whereas organizational and individual variables were statistically significant factors for part-time faculty. Among the measured variables, the quality of facilities/resources (beta = .33) and the faculty's preference for part-time appointment (beta = -.22) demonstrated a strong association with job satisfaction. Not only the overall job satisfaction (beta = .24) but also the perception of compensation (beta = .26) have significant impacts on the intention to remain in academia. Recommendations for future research were offered.

2:00-2:50 p.m. AT-RISK STUDENTS (Discussion Session) Salon B

Presider: Marie Miller-Whitehead, Tennessee Valley Educators for Excellence

UWA SECONDARY EDUCATION TUTORING PROJECT FOR THE WEST ALABAMA LEARNING COALITION

Gloria D. Richardson, Gloria J. Abrams, John Byer, and Tom W. DeVaney, University of West Alabama

             This study, conducted during the fall semester 1999, by tutors who were secondary education majors at the University of West Alabama sought to determine if tutoring on specific reading skills helped at-risk students improve their frustration levels. Of the initial 24 participants, 12 students completed both the pretest and posttest administration of the Informal Reading Inventory. Analysis of the Frustration Level scores was completed with a paired t-test. A T-value of -2.97 was obtained. This was significant at P-Value of 0.013. This finding confirmed the hypothesis: There will be an improvement in at-risk high school students, Frustration Reading Levels after 15 hours of remedial training in specific reading strategies and skills by UWA tutors. The qualitative portion of the study examined attitudes toward tutoring and found that the preservice teachers (tutors) were positive about their field experiences.

AN ALTERNATIVE SCHOOL WITHIN A SCHOOL: A CASE STUDY ON MEETING MOTIVATIONAL, CURRICULA, AND INSTRUCTIONAL NEEDS OF AT-RISK STUDENTS

Rebecca S. Watts, University of Louisiana at Monroe

             High dropout rates potentially threaten the social and economic status of our nation. Alternative schools are becoming increasingly popular as programs to prevent at-risk students from dropping out of school. Previous research has identified common characteristics among alternative schools (Peterson, Bennet, & Sherman, 1991), as well as risk factors that characterize students in danger of dropping out (Hahn, 1987; Barber & McClellan, 1987). Of significant interest and less emphasized in research on at-risk students are the motivational factors that lead to students becoming classified as at-risk and how the curricula and instructional programs of alternative schools accommodate the needs of at-risk students. Alternative schools differ in their organization and the student population they serve. This study focused on an alternative school program in north Louisiana functioning within the facilities and programs of a conventional high school and serving the at-risk students identified within the conventional high school. The purpose of this study was to identify the needs of students in the alternative program and determine how the program's curriculum and instruction differs from that of the conventional school in accommodating student needs. Data were gathered through classroom observations, analysis of documents reflecting the goals and objectives of the program, and interviews with students, teachers, and the alternative school program director. Results of this study suggested that self-esteem is a significant motivating factor among at-risk students. Curricula and instructional programs of the alternative school are implemented to accommodate student needs, yet these programs compliment the programs offered by the conventional school.

TEACHER'S PERCEPTIONS OF THE EFFECTIVENESS OF READING TUTORIAL PROGRAM FOR AT-RISK STUDENTS

Mimi Mitchell Davis, Educational Consultant, and Patricia Ashley, Daviess County (KY) Public Schools

             In response to young children that are at risk for reading failure, school districts are implementing tutorial programs (e.g., Reading Recovery) that use volunteers or certified teachers. Previous research has examined the effectiveness of these programs on reading achievement. However, little is known about what teachers think about reading tutorial programs. The purpose of this study, then, was to examine teachers' perceptions of the effectiveness of a tutorial program that used volunteers.
             The goal of the school district of this study was to support classroom reading instruction by implementing a tutorial program. To determine its effectiveness, surveys were developed and sent to 93 teachers at 12 elementary schools. The surveys, comprised of a structured rating scale and open-ended questions, focused on gaining responses that reflected the teachers' opinions. The rating scale ranged from highly effective to having a negative affect or not relevant. The open-ended questions provided an opportunity for the teachers to state their opinions about the program and to offer possible changes that would strengthen it. Frequencies of the effectiveness were tabulated, and the open-ended questions were examined for emerging patterns and trends.
             Of the 61 teachers (65.5%) who responded, most (68.9%) reported that there was no coordination between the tutorial program and classroom reading instruction. Only two (3.3%) teachers reported that there was coordination. This finding is interesting since the district's goal was to support classroom reading instruction. Other findings indicated that some students received less classroom reading instruction than students who did not participate in the program.
             This investigation suggested that teachers who understand their students' needs better than anyone should play a major role in the development and implementation of a tutorial program if it is to be effective. The findings have implications for educators who are interested in reading tutorial programs for at-risk students.

2:00-2:50 p.m. MATH EDUCATION (Discussion Session) Meeting Room 1

Presider: Vincent McGrath, Mississippi State University

ATTITUDES TOWARD MATHEMATICS INSTRUMENT: AN INVESTIGATION WITH MIDDLE SCHOOL STUDENTS

Martha Tapia, Berry College, and George E Marsh II, The University of Alabama

             Declining national test scores in mathematics and an increasing dislike of the subject by students have elevated student attitudes as an area of concern. Positive attitudes toward mathematics are critical for successful participation in the mathematics curriculum and achievement. The purposes of this study were to: (1) develop an instrument to measure students' attitudes toward mathematics (ATMI), (2) investigate a sample of middle school students, and (3) find the underlying dimensions comprising the ATMI. The sample consisted of 262 middle school students attending a private, bilingual preparatory school in Mexico City. Students were asked to indicate their degree of agreement with each statement, from strongly disagree to strongly agree. Variables considered were the value of mathematics as subject matter, anxiety, motivation, confidence, enjoyment, and adults' expectations. The alpha reliability coefficient for the scores on the entire instrument was .95. The nine weakest items were eliminated, which retained an alpha reliability coefficient of .95. A maximum likelihood factor analysis with a varimax (orthogonal) rotation was used to yield three scales designated as self-confidence, enjoyment of mathematics, and value of mathematics. The alpha reliability coefficients for the scores on the subscales were .94, .92, and 84, respectively. The ATMI psychometric analysis yielded sound properties and can be used by researchers and practitioners to measure students' attitudes toward mathematics in the middle school.

OFFERING DEVELOPMENTAL MATHEMATICS ONLINE

Gail H. Weems, Southern Illinois University, Carbondale

             This study compared two sections of beginning algebra: one taught online and the other by traditional lecture. The variable of primary interest was mathematical ability; however, the discussion also included students' attitudes toward mathematics and their reports of the effectiveness of the online course.
             Of the 28 students enrolled in each section, 19 completed the course. The reduction in class size is a bit large; however, it is within the range of expected sizes for developmental mathematics courses.
             The instructor met with the online section during the first week of classes, introducing students to the course and demonstrating the course web page. After the first week, students communicated with the instructor by email, telephone, course discussion board, and during a review session held on campus the week of each exam. Students used the textbook, video, CD tutorial, and contact with the instructor, by the aforementioned methods, to learn the material. Weekly online quizzes required students to keep the pace set forth in the syllabus, and homework problems were collected during each of the three exams. Students in the traditional section used the same textbook, video, and CD tutorial; however, they did not have access to the course web page and took their quizzes in class.
             While no significant differences between the scores of the two groups were found, and comments from the online students were positive, instructors should remain aware that this instructional method may not be appropriate for all students and all topics. Many students initially enrolled in the online section, only to learn of the instructional method, and changed to a traditional section. Certain topics were also more difficult with the materials used online, with factoring polynomials being particularly problematic. The findings of this study suggested implications for classroom practice, specifically for teaching mathematics.

THE EFFECTS OF FOUR SELECTED COMPONENTS OF OPPORTUNITY TO LEARN ON MATHEMATICS ACHIEVEMENT OF GRADE 12 STUDENTS IN NEW PROVIDENCE, BAHAMAS

Janet M. Collie-Patterson, The University of Southern Mississippi

             The purpose of this study was to determine if a single dimension of opportunity to learn (OTL) could be identified using four selected components of teachers, students, classrooms, and schools' characteristics and to determine if each of the four selected components of OTL was related to mathematics achievement.
             The primary sample of the study consisted of 1,015 grade 12 students from six public and six private schools in New Providence, Bahamas. Complete data were available for 463 students. The secondary sample consisted of 52 mathematics teachers who taught the participating students in the tenth, eleventh, or twelfth grade. Both the complete and incomplete data sets were analyzed.
             The findings of this study indicated that the model-data-fit was reasonable, suggesting that there was a relationship between opportunity to learn and three selected components of teachers' characteristics, students' characteristics, and schools' characteristics. The fourth component, classrooms' characteristics, was not significantly related to OTL.
             Each of the four components was significantly related to mathematics achievement. When the component indicators were considered individually, course taking, teaching strategies, professional development, educational background, affiliation, strength of climate, recognition, commitment, accomplishment, socioeconomic status, attitude toward school, and student's prior ability were significantly related to mathematics achievement. However, manipulative use, parental involvement, and years of teaching experience were not significantly related to mathematics achievement. Surprisingly, professional development, attitude toward school, strength of climate, recognition, and accomplishment were negatively significant to mathematics achievement
             In terms of effect size, students' characteristics made the largest contribution to mathematics achievement followed by classrooms, schools, and then teachers' characteristics. The set of students' characteristics (parental involvement not significant) explained about 60% of the variability in mathematics achievement, classrooms' characteristics (manipulative use not significant) explained about 36%, schools' characteristics explained about 12%, and teachers characteristics (teaching experience not significant) explained about 8%.

2:00-2:50 p.m. STATISTICS (Discussion Session) Meeting Room 2

Presider: Ernie Rakow, The University of Memphis

BIGGER IS NOT BETTER: SEEKING PARSIMONY IN CANONICAL CORRELATION ANALYSIS VIA VARIABLE DELETION STRATEGIES

Mary Margaret Capraro, University of Southern Mississippi

             Theory building and/or generation is often guided by the principle of parsimony, which suggests that given two equally viable explanations for a phenomenon, the simpler explanation is most likely replicable. As such, social science researchers often seek to explain the most for the least, such as maximizing the R2 value in a multiple regression analysis with the fewest possible predictors.
             Furthermore, even in multivariate contexts where researchers may purposefully examine omnibus relationships between multiple variables to model complex reality, parsimony is often preferred over obtuse models that include variables that do little to help the explanatory power of the model. Canonical correlation analysis, for example, can be used to examine the joint relationship between two sets of variables. However, either (or both) of these variable sets may include variables that are minimally (or simply not) useful in maximizing the overall effect size, or squared canonical correlation. In such cases, the researcher may be compelled to delete the variables from the final model. Given the flexibility of canonical correlation analysis in theory development, it is useful in both measurement and substantive contexts to determine which variables provide little information for the analysis.
             There are several variable deletion strategies available to facilitate parsimony in canonical correlation analysis, each of which focuses on slightly different aspects of the analysis. The purpose of the paper was to illustrate these methods in a manner accessible to applied researchers. To make the discussion concrete, data from a study of math attitudes and thinking processes in sixth grade students were used to illustrate each approach. Three methods of variable deletion were presented that focused on canonical communality coefficients, squared canonical correlation coefficients, and weighted communality coefficients.

CANONICAL COMMONALITY ANALYSIS AND THE COMMON MAN: UNDERSTANDING VARIANCE CONTRIBUTIONS TO OVERALL CANONICAL EFFECTS

Robert M. Capraro, The University of Southern Mississippi

             Multivariate statistical methods have increased in popularity over recent decades, largely because of increased access to computer technology and user-friendly software packages. Beyond ease of access, however, multivariate methods are critical in research because they allow researchers to more accurately examine the complexities of reality, where causes have multiple effects and effects have multiple causes. Accordingly, interest in omnibus model effects has increased.
             Even in the context of full-model effects, however, researchers are also commonly interested in determining from whence the full effect came. For example, in canonical correlation analysis (the multivariate general linear model that subsumes other parametric analyses as special cases), researchers may wish to determine what variables in the predictor set were able to contribute to the effect size. Furthermore, it may also be of interest to determine what useful/predictive variance was shared among two or more variables in a set. These unique and shared portions of variance can be partitioned and defined using a process called canonical commonality analysis.
             Essentially, canonical commonality analysis allows a researcher to determine what unique contribution a given variable made to the effect size. Furthermore, the analysis will also inform the researcher regarding what shared contributions were made among all combinations of the variables. While the overall multivariate effect is generally of interest to most researchers, defining the sources of variance for the overall effect may also be of interest.
             The purposes of the paper were to: (1) provide an overview of canonical correlation analysis and (2) illustrate canonical commonality analysis as a post hoc method of determining unique and shared variance contributions from variables in a predictor or criterion sets. Heuristic data were employed to make the discussion concrete and accessible to the applied researcher. Venn diagrams were used to build a conceptual framework for the statistical discussion.

INTERPRETING THE FOUR TYPES OF SUMS OF SQUARES IN SPSS

Jesus Tanguma and F. M. Speed, Texas A & M University

             A typical analysis of variance (ANOVA) table consists of, among other things, source, sums of squares, degrees of freedom, mean square, F-calculated, and p-calculated. Because each F-calculated is found by a ratio of mean squares, researchers need to be very careful as to how the sums of squares are computed. These sums of squares are computed by SPSS differently depending on the design being analyzed.
             Three research designs (balanced, unbalanced, and missing data) are illustrated by means of hypothetical examples. The purpose of the paper was to explain and illustrate the choices.
             When the design is balanced, all of the SPSS types of sums of squares yield equivalent results for testing the significance of the ANOVA models. Moreover, the hypotheses being tested are interpretable.
             When the design is unbalanced, the hypotheses being tested are dependent on the type of sums of squares being used. Thus, decisions made based on tests of significance when the design is unbalanced may differ depending on the type of sums of squares being used. Additionally, only the hypotheses being tested under Type III and Type IV sums of squares are interpretable.
             In case there are empty cells in the design, none of the types of sums of squares agree. Additionally, only the hypotheses being tested under Type IV sums of squares are interpretable. However, these sums of squares are dependent on the variability and pattern of the missing cells. Thus, if any conclusions are to be made based on these sums of squares, they should be made with caution.
             Each design was also analyzed using SAS. At each step, the results, testable hypotheses, and sums of squares were identical, within rounding errors, to those obtained by SPSS.

2:00-2:50 p.m. TECHNOLOGY (Display Session) Meeting Room 4

A COMPARISON OF STUDENTS' COMPUTER-GENERATED ART AND STUDENTS' ART CREATED WITH TRADITIONAL MEDIA

Barbara England, Freed-Hardeman University, and Dennie Smith, The University of Memphis

             As technology and computers become more accepted in the classroom, students and educators will be challenged to use technology in creative and expressive ways. It appears that in art education only a few inquiries have been conducted looking specifically at technology and the use of computers as art media. Publications, and forums for art educators have discouraged quasi-experimental studies in the field.
             This study explored the perceptions and processes of children when using computers as art media. Children were observed as they created art using computer media and traditional media (crayons, markers, and pencils). The study compared procedures to complete the drawings and the results of drawings created with the two different media. The individual elements of art were the basis for forms developed by the researcher to analyze each drawing.
             The analysis of the element space, value, and color indicated no significant difference in the drawings, and the presence of the element or the degree to which it was emphasized in the drawing. Additionally, there was not a significant difference in the recognizable subject matter in the two drawings. The significant differences in the drawings occurred in the use of lines and shapes. The computer drawings contained more geometric shapes than organic shapes and more lines drawn with the straight edge tool when compared to the lines drawn using a ruler in the traditional drawings.
             The perceptions of the children concerning their artwork and the two drawing experiences were favorable for the computer media. The perception that it was easier to change a drawing influenced the degree to which they enjoyed the experience. This study suggested implications for the art educator as well as the computer artist and encouraged further study in the area of technology when used as art media.

ELECTRONIC PORTFOLIOS: METHODS STUDENTS CONNECT WITH TECHNOLOGY

Vivian H. Wright and Joyce Stallworth, The University of Alabama

             With implementation of national standards addressing technology, teacher preparation programs are faced with the issues of preparing teachers to effectively use and to seamlessly integrate technology across content areas. A team teaching approach at one major southeastern university required its methods students to produce electronic portfolios during spring 2000. The teaching team consisted of secondary education language arts and social studies faculty, instructional technology faculty, and graduate students from both disciplines. This effort of modeling technologically best practices resulted from numerous team meetings, intensive planning, and consistent project evaluation. The preservice teachers were required to attend technology seminars in addition to regular classroom and methods work. Students were evaluated on their electronic portfolios, which consisted of websites, digitally-edited teaching episodes, databases, concept maps, and more. Through pretest and posttest surveys, the students were assessed on their perceptions of an electronic portfolio's value and their ideas of how technology can enhance teaching and learning in future classrooms. Results from these assessments and procedural details were presented, along with the challenges and successes experienced by the teaching team and the students.

MSERA AND THE MILLENIUM: SUBMITTING AND REVIEWING PROPOSALS ONLINE

Malenna A. Sumrall, The University of Alabama at Birmingham

             During 1999, selected MSERA members were asked if they would like to participate in submitting their proposals via email, and program committee members were asked if they would like to review papers via email. The response to this pilot project was overwhelmingly positive. Therefore, it was decided to allow any MSERA member to submit a proposal electronically for the 2000 annual conference.
             Web pages were created for collecting information from program committee members, for submitting proposals, and for reviewing proposals. All data from these web pages came to the author as email messages. The data in these messages were used to create a variety of files that helped automate the duties of the program chair. This display session described the methods used to facilitate this process.