Search

This study explored the way parents use information sources for information about their child's motor, social, and cognitive development. This study sought to identify: (1) variations in parental use of information based on the domain of the child's development for which the parent sought information; (2) the role of parent and child gender in patterns of parental information use; and (3) the extent to which parental information use varies with indicators of the child's development. Sixty parents of 3-year-olds were interviewed about the size of their information network, frequency of use of information sources, and usefulness of information sources. These items were related to the variables of parent and child gender, domain of child's development, and child's developmental level. Analysis showed that information use varied according to all four variables, most importantly for the domain of the child's development. Results also showed that: (1) parents had significantly larger information networks for information about their child's social development than cognitive and motor development; (2) fathers, more than mothers, used cognitive information sources, and found them to be more useful; (3) parents of girls who scored in the lowest third of the sample on a vocabulary test had significantly larger information networks than did parents of other girls; and (4) parents of boys used information sources more than did parents of girls. (TM)

This study investigated the robustness of the James second-order test (James 1951; Wilcox, 1989) and the univariate F test under a two-factor fixed-effect analysis of variance (ANOVA) model in which cell variances were heterogeneous and/or distributions were nonnormal. With computer-simulated data, Type I error rates and statistical power for the two tests were estimated. With data sampled from normal distributions, the F test was not robust to variance heterogeneity for equal or unequal sample sizes, but the James second-order test was robust in these situations. With normal distributions, equal variances, and equal sample sizes, the magnitude of power difference between the two tests was generally small when testing the main effects, but the magnitude of power difference between the two tests varied when testing the interaction effects. With data sampled from nonnormal distributions, although the James second-order test generally was liberal when the population distribution was skewed, the test was robust under several nonnormal distribution situations. Additionally, the robustness of the James second-order test in factorial designs may be affected by combinations of nonnormal distributions, sample sizes, and variance patterns. (Contains 22 references and 7 tables.) (Author/SLD)

This article documents the development of an instrument that would allow researchers and clinicians to assess the ways in which students differ qualitatively with respect to how they cope with the demands of evaluative situations. The Academic Anxiety Coping Scale identifies modal types of coping cognitions and behaviors that students employ palliatively and instrumentally in academically stressful circumstances to attenuate anxious arousal. The scale was administered to 215 university and community college undergraduates. Factor analysis revealed four factors: (1) emotional/physiological; (2) relaxation/letting go; (3) preparation and planning; and (4) worry. One way analysis of variance compared students on each factor, first by group (community college versus university) and then by gender. Females reported more worry than males, and conversely, males reported using more relaxation techniques than females. University students reported more preparation and planning and greater use of relaxation techniques than community college students. Seniors also reported using relaxation more than freshmen. Further research is needed to follow through on this pilot study to revise and consolidate scale items. (Contains 16 references.) (Author/SLD)

A method is presented for the simultaneous analysis of differential item functioning (DIF) in multi-factor situations. The method is unique in that it combines item response theory (IRT) and analysis of variance (ANOVA), takes a simultaneous approach to multifactor DIF analysis, and is capable of capturing interaction and controlling for possible confounding variables. It is referred to as the IRT-ANOVA method. The most salient feature is that the procedure used IRT to control for group ability differences and familiar inferential procedures to test the DIF effect. Residuals can be construed as item scores free from the effects of both personal ability and item difficulty. The use of ANOVA provides not only a test statistic based on a familiar distribution, but also descriptive measures of DIF magnitude in terms of group means and variances. Simulations in the one-factor, two-group situation reveal the usefulness of the approach and indicate error rates. Seven figures illustrate the simulations. (Contains 8 references.) (SLD)

Most training programs in education and psychology focus on classical test theory techniques for assessing score dependability. This paper discusses generalizability theory and explores its concepts using a small heuristic data set. Generalizability theory subsumes and extends classical test score theory. It is able to estimate the magnitude of multiple sources of error simultaneously, unlike classical theory, which enables only a single source of error to be considered at one time. Generalizability theory forces us to see that it is scores, not the tests themselves, that are reliable. Generalizability studies are the initial round of analyses that generate variance components for the sources of error in the study. Design studies use these variance components to answer questions about alternative measurement protocols. Generalizability analyses distinguish between decisions made in the context of cutoff scores (absolute decisions) and those that consider relative standing. An example involving 10 people who have taken a 4-item test each of 3 times illustrates application of generalizability theory. Six tables and two figures present details of the analysis. (Contains 11 references.) (SLD)

Much Catholic school and church rhetoric suggests that Catholic schools possess distinctive learning environments. However, little empirical evidence has been found to support this assertion. Research has been hampered by the lack of an appropriate assessment instrument. This paper describes the development of a seven-scale instrument to assess student perceptions of classroom psychosocial environments in Catholic schools. The instrument assessed student affiliation, interactions, cooperation, task orientation, order and organization, individualization, and teacher control. The instrument was used in a survey of 80 Catholic and 24 government classes in 32 schools. Findings revealed that the government schools scored significantly higher than the Catholic schools on the interaction and task orientation scales. A comparison of the environments of religion and science classes in Catholic schools indicated negligible differences on all scales except task orientation. The perceptions of 9th- and 12th-graders differed significantly in all scales except interactions. Girls perceived classrooms more positively than did boys. In general, the data suggest that the distinctive nature of Catholic schooling did not extend to all classroom environment dimensions deemed important to Catholic education. Moreover, government schools were perceived to be better organized, more conducive to positive teacher-student interactions, and more task oriented than Catholic schools. Four figures and four tables are included. (Contains 80 references.) (Author/LMI)

The effects of ethnicity, gender, and risk on locus of control were investigated for gifted children and for high ability nongifted children. Subjects (n=805) in grades 5 through 7 were African-American, Caucasian, Latino/Hispanic, or Fillipino and had been referred for evaluation of giftedness, which included evaluation of intellectual ability and six risk factors: (1) environmental risk, (2) economic risk, (3) language risk, (4) cultural risk, (5) emotional risk, and (6) health risks. Analysis of variance revealed significant main effects for giftedness, ethnic background, and level of risk, as well as a significant ethnic X risk interaction. In addition, a gender X giftedness X ethnic background analysis of variance yielded a significant main effect for gender. Overall, higher internal locus of control was associated with female Caucasians not at risk. However, in contrast to Caucasians, greater risk was associated with a higher internal locus of control in non-Caucasians. Results confirmed previous findings in showing a more internal locus of control in gifted children, and further indicate that, for gifted and intellectually bright non-Caucasians, risk is associated with, and may serve to strengthen, a greater internal locus of control. Results support the use of locus of control as an alternative in identifying gifted traditionally underrepresented groups such as African-Americans and Latino/Hispanics. (DB)

A study described the characteristics of sociocognitive conflicts and the discourse associated with such conflicts, and determined how the cognitive processes exercised during discussion were internalized. Participants were 97 fourth-grade students and 6 elementary classroom teachers. Peer-led and teacher-led discussions of texts were examined to determine the role of sociocognitive conflict in these discussions. Constant-comparative methods revealed three categories of sociocognitive conflict: conflict within self; conflict with others, and conflict with text. Sociolinguistic analyses revealed that students' discourse in peer-led discussions was significantly more complex than in teacher-led discussions. The Cognitive Conflict Scenario revealed that students in peer-led discussions were able to recognize and resolve sociocognitive conflicts better than students in teacher-led discussions. Findings suggest that peer-led discussions produced richer and more complex interactions than did teacher-led discussions and resulted in the internalization of the cognitive processes associated with engaged reading. (Contains 61 references, 5 tables and 6 figures of data; the Cognitive Conflict Scenario Task Pre/Postttest and the scoring rubric are attached.) (RS)

This paper describes an item response theory (IRT) based method of differential item functioning (DIF) detection that involves neither separate calibration nor ability grouping. IRT is used to generate residual scores, scores free of the effects of person or group ability and item difficulty. Analysis of variance is then used to test the group differences in residual scores. Simulation studies were conducted to examine the power and error rate of the method in terms of varying sample sizes, difference reference or focal group ratios, and varying degrees of group ability difference. Also investigated was the robustness of the procedure to moderate model-data misfit. Mantel-Haenszel statistics were also computed for comparison purposes. Desirable features of the new (IRT-ANOVA) method include: (1) concurrent calibration; (2) no ability grouping; (3) capability to examine interaction effects; (4) applicability to both dichotomous and polytomous items; and (5) applicability to relatively small sample sizes. Limitations and advantages of the method are discussed. Six figures illustrate the analyses. (Contains 2 references.) (SLD)

This paper presents a new computer program called Zigzag, suited for investigating readers' mental activity during text processing. The program permits data collection of two different kinds, namely reading time and reading errors for each of the words read. The second dependent measure, reading errors, provides a better control of the readers' text comprehension and permits the restriction of statistical analyses of the time data to correctly identified words. Subjects of this study were French-speaking third and sixth graders living in the greater Montreal (Quebec, Canada) area. One major goal was to examine the effect of visually presented contextual information on text processing at two developmental levels and to examine the readers' dependence on visual information in an interactional perspective. The paper analyzes the results for the experimental text as a whole, but also with respect to some linguistic variables, such as grammatical category, word position in the sentence, or other variables. Results indicate major differences between the data and the time data obtained by means of other on-line techniques with respect to the same linguistic variables. The paper concludes with a more detailed discussion of these differences. Four tables and two figures of data are included. Contains 27 references. (RS)

The purpose of this study was to determine if student teachers' attitudes toward their student teaching experiences differed significantly after the addition of professional development school experiences. Two groups of student teachers (Spring 1992 and Spring 1993) were compared using the Mississippi Student Teaching Attitude Inventory (MSTAI) constructed by Benton and Richardson (1990). Each group consisted of 35 students. A one-way analysis of variance revealed significant differences for 7 of the 38 items. These items consisted of the following statements, to which students responded using a Likert scale: "I enjoy teaching"; "My student teaching responsibilities do not restrict my nonprofessional responsibilities"; "I feel that I am prepared to teach"; "I feel prepared to meet the needs of individual learners"; "My students regard me with respect and have confidence in my professional ability"; "My students have the opportunity for enrichment activities daily"; and "Using more than two teaching methods during a class period is not difficult." Implications for the use of the MSTAI include: longitudinal group comparisons, restructuring of courses and programs, and utilization of findings to implement further professional development opportunities and changes. (Author/IAH)

The interaction of adult learning styles and the media, methods of instruction, and group functioning in a distance learning class using audiographics and computer-mediated communication was studied and compared with similar interaction in non-equivalent traditional classes. One graduate class in theory and practice of distance education, taught at a distance with 15 students, and 3 traditional on-campus classes were selected. The interaction of learning style and media, methods of instruction, group functioning, and support was determined by administering a questionnaire developed for the study and subjecting results to an analysis of variance. The major finding is that learning styles do not impact how students interact with media and methods of instruction, but do affect satisfaction with other learners. Class type impacts student satisfaction with media, methods, group satisfaction, goal setting, and group climate. Results of this study cannot be generalized because of the small sample in the distance class. (SLD)

This study integrates into one comprehensive Monte Carlo simulation a vast array of previously defined and substantively interrelated research studies of the robustness of analysis of variance (ANOVA) and analysis of covariance (ANCOVA) statistical procedures. Three sets of balanced ANOVA and ANCOVA designs (group sizes of 15, 30, and 45) and one set of unbalanced ANOVA and ANCOVA designs (groups of 15, 30, and 45) were simulated. Violations of normal shape and three degrees of heterogeneity of group variances were included. Each set of ANCOVA analyses included ANOVA violations as well. Each data set violation was simulated in isolation and in combination, resulting in 665 unique ANOVA or ANCOVA F sampling distributions. Unbalanced designs almost always produced statistically invalid F ratios in the presence of any data set assumption except those that only involved perturbation of shape. In balanced designs, a degree of robustness beyond that expected was found, even when heterogeneity of variance was coupled with heterogeneous regression slopes and a skewed covariate. Robustness was found even when the ratio of largest to smallest group variance was five. Implications of these findings for educational, social science, and behavioral research are discussed. Three tables are included. (Contains 29 references.) (Author/SLD)

The purpose of this study is to help define the precise nature and limits of the tolerable range in which a researcher may be relatively confident about the statistical validity of his or her research findings, focusing specifically on the statistical validity of results when violating the assumptions associated with the one-way, fixed-effects analysis of variance (ANOVA) and one concomitant analysis of covariance (ANCOVA) statistical procedure. Methodological and data set statistical assumptions that must be met for ANOVA and ANCOVA are discussed. Research results were obtained from an exploratory study of the effects of single and compound violations of the mathematical conditions (assumptions) underlying use of ANOVA and ANCOVA through Monte Carlo methods using randomly created data for a mathematical simulation (665 data set conditions). For all of the analyses, comparisons were made between the empirical F sampling distributions and the theoretical (i.e., nominal) F distributions expected when one uses normal theory. For balanced designs, the ANOVA and ANCOVA F tests were found to be remarkably robust when faced with most of the violations in the simulation. Research does reaffirm that ANOVA and ANCOVA should be avoided when group sizes are not equal. Specific recommendations are made for checking the ratio of largest to smallest group variances. (Contains 29 references.) (SLD)

By employing a concomitant variable, researchers can reduce the error, increase the precision, and maximize the power of an experimental design. Blocking and analysis of covariance (ANCOVA) are most often used to harness the power of a concomitant variable. Whether to block or covary and how many blocks to be used if a block design is chosen become important. This paper provides an historical review of the problem and recommends future research to examine the problem based on how subjects are assigned, how data are analyzed, and the distributions of the variables. In this study, subjects were randomly assigned to treatments ignoring the concomitant variable, and data were analyzed by one-way analysis of variance (ANOVA), post-hoc two-block, four-block, and eight-block ANOVA and ANCOVA. Distributions of the concomitant and dependent variables were normal. The Monte Carlo method was used to generate 20,000 data sets for 8 experimental conditions (2 levels of subject and 4 levels of correlation between concomitant and dependent variables. The five analysis procedures were examined under each experimental condition. Results show that ANCOVA is more powerful than post-hoc rank blocking. Eight tables present analysis results. (Contains 36 references.) (Author/SLD)

To determine if differing ability levels will affect the acquisition of problem-solving skills and self-esteem as a result of participation in two approaches to teaching problem-solving skills, a study was conducted with sixth graders in a posttest-only control group experimental design. Subjects were 102 sixth graders randomly assigned to 5 classes. Two classes participated in the Creative Problem Solving (CPS) for Kids approach to teaching problem solving. Two classes received computer-assisted instruction in problem-solving designed by the Minnesota Educational Computing Consortium, and one class was a control group. Both approaches consisted of five 3-minute lessons per week for 6 weeks. Results suggest that thinking-skills instruction does impact the development of creative and critical thinking and that the acquisition of these skills has a positive effect on self-esteem. The study also provides evidence that the length of training is an important consideration in providing thinking-skills instruction, and that such instruction should be an integral part of the curriculum rather than a supplementary or isolated program. In addition, thinking-skills instruction is appropriate for students at all ability levels. Seven figures and 12 tables present study findings. (Contains 17 references.) (SLD)

The literature is reviewed regarding the difference between planned contrasts, OVA and unplanned contrasts. The relationship between statistical power of a test method and Type I, Type II error rates is first explored to provide a framework for the discussion. The concepts and formulation of contrast, orthogonal and non-orthogonal contrasts are introduced. It is argued that planned contrasts are confused on thoughtful research questions of interest and reflect researchers' rational anticipation. An OVA test or unplanned contrasts, on the other hand, do not provide desired information in many situations. It is also explained that, to control for the possible inflated error rate for unplanned contrasts which usually test a large number of hypotheses, wanted or unwanted, some Bonferroni type of corrections are invoked. It is these corrections, usually built into statistical tables, that reduce the power of unplanned contrasts. This is demonstrated through a comparison of the critical values for planned contrasts and for some popular unplanned contrasts. (Contains 27 references.) (Author)

There are advantages to using a priori or planned comparisons rather than omnibus multivariate analysis of variance (MANOVA) tests followed by post hoc or a posteriori testing. A small heuristic data set is used to illustrate these advantages. An omnibus MANOVA test was performed on the data followed by a post hoc test (discriminant analysis). A second MANOVA using Helmert coding as a means for a planned comparison test was run and the results were compared to those from the MANOVA followed by the unplanned tests. The results provide concrete examples of the different possible outcomes using various procedures. Two advantages of using planned or a priori contrast over post hoc tests are presented. In addition to increasing the power against Type II error rate, planned comparisons force the researcher to be more thoughtful in the design of the research. Nine tables and an appendix containing the questionnaire. (Contains 24 references.) (Author)

It is argued that analysis of variance (ANOVA) and related methods should be taught using a general linear model (GLM) approach, rather than a classical ordinary sums of squares approach. The GLM approach emphasizes the linkages among conventional parametric methods, emphasizing that all classical parametric methods are least squares procedures that implicitly or explicitly use weights, focus on latent synthetic variables, and yield effect sizes analogous to "r" squared (are correlational). The case for teaching statistics using a GLM conceptual framework is based on the following four contentions: (1) a GLM instructional approach provides a unifying conceptual framework that better enables students to understand analytic methods; (2) a GLM approach provides a better match between the researcher's analytic model and the researcher's model of reality; (3) the GLM emphasis on planned contrasts helps students understand the critical role of reflective thought in good research; and (4) a GLM instructional approach helps students see that focusing on variance-accounted-for (or other) effect sizes, rather than statistical significance, is important in all analyses. Three tables present analysis examples. A 68-item list of references is included, and an appendix lists Statistical Package for the Social Sciences control cards. (Author/SLD)

Potential gender-related differences in the process of mathematical word problem solving were investigated. Sixth and eighth graders (n=302) solved 2-step mathematical word problems using a 9-step problem solving plan. The nine steps were: (1) identify the facts; (2) identify the question; (3) draw a diagram; (4) choose the operations; (5) write an open sentence; (6) estimate the answer; (7) compute the answer; (8) state the answer; and (9) validate the answer. A 2 x 2 between-subjects multivariate analysis of variance was computed to determine the effects of gender and grade level on the subjects' scores on each of the nine steps of the problem solving plan. There was a significant gender-related difference in the overall process of problem solving in favor of females. Univariate "F"'s for each of the nine steps were not significant. Although there were no significant differences at any of the individual steps, the combined use of the same steps produced a gender-related difference. It is concluded that the use of a step-by-step problem solving plan might be gender biased. (Author/SLD)

This Monte Carlo study examines whether, given various numbers of variables, treatments, and sample sizes, in a one-way multivariate analysis of variance, Type I error rates of the test approximations provided by the BMDP program, the Statistical Analysis System (SAS), and the Statistical Package for the Social Sciences (SPSS) for Roy's largest root, Hotelling's trace, Wilks' likelihood ratio, and P. C. S. Pillai's trace meet the stringent criterion of J. V. Bradley (1978) for robustness. For each of 85 conditions, the Type I error rate for each of the above statistical tests was estimated based on 100,000 random samples per situation. Results indicate that in multivariate analysis of variance studies with relatively small numbers of subjects of around 15 per treatment level, or fewer, the current probability levels reported by SAS and SPSS are conservative for the F approximations based on Pillai's trace and liberal for the F approximations based on Hotelling's trace. The BMDP 4V program does not report Pillai's trace and reports accurate probability values for Hotelling's trace. All of these programs report accurate values for the F approximations based on Wilks' likelihood ratio and for Roy's largest root (SPSS does not report Roy's largest root). Recommendations are made for specific conditions. Two tables and five figures present analysis results. (SLD)

Applied researchers frequently precede analyses of interest with one or more preliminary tests, used to help researchers determine which variables to examine more closely, or whether there are anomalies in the data set. These texts can be classified into three categories: omnibus tests, tests for model fit, and exploratory tests. Fifty-four applied journal articles, 31 statistical textbooks, and the statistical literature are reviewed to discuss some limitations associated with uses of some preliminary tests. The focus is on the following topics: (1) the omnibus F-test in analysis of variance; (2) tests for variance equality; (3) tests for equality of regression slopes in analysis of covariance; and (4) tests for sphericity in repeated measures designs. In general, it is concluded that many preliminary statistical tests are not useful. In many contexts, omnibus tests do not answer questions of substantive interest. Preliminary analyses in tests for model fit are unnecessary because alternative less restrictive models can be used, and because many tests for violations of data assumptions lack adequate statistical power, or are overly sensitive to another assumption violation. More focused analyses are advocated, using less restrictive analytical models, and an increased use of exploratory analyses is recommended. One table illustrates the discussion. Lists of the journal articles and textbooks reviewed are appended. (Contains 43 references.) (SLD)

Student perceptions of changes in the mode of instruction used by teachers who use alternative assessment techniques and changes in student attitudes and beliefs about mathematics after experiencing alternative assessment were studied. Eighteen teachers surveyed their elementary school and secondary school students three times during the school year to measure attitudes about mathematics and participation in classroom activities. At time 1, there were 387 students, with 384 students at time 2 and 137 students at time 3. An attitude score and traditional and non-traditional classroom activity scores were the dependent variables in a repeated measures analysis of variance (ANOVA). Traditional activities were things such as watching the teacher and textbook problems. Non-traditional activities included working in groups and having students make up their own problems. Gender, ethnicity, and ability were blocking variables in a two-way ANOVA. For attitude, there were significant main effects for teacher, time of observation, and ability. Attitudes improved over the year, as did non-traditional activities. There were significant main effects for all groups except ability. Girls, African Americans, Hispanics, and students in the lower and upper grades reported more involvement in classroom activities than did boys, Anglo Americans, and seventh and eighth graders. Five tables and seven figures illustrate the findings. The student survey is also included. (Contains 16 references.) (SLD)

The validation of key contextual features of algebra word problems was studied in two phases. In the first phase, five experts were asked to assess the appropriateness of the concepts in the problems and the adequacy of the assignment of the contextual features to the problems. In the second phase, construct validity was established by having 6 judges rate each of the 16 word problems in random order on the contextual features of familiarity, imageability, and variable type (discrete or continuous). A repeated measures analysis of variance for the construct validity of the key contextual features showed that when one rater or judge and one of the problems were removed, agreement between problems and the criterion were extremely high. When a step-down analysis on each key context feature and variable type was done without one judge, the results indicate a convergence on the constructs devised. In effect, judges agreed with each other, and were correct on 93.5 percent of the ratings, which is strong evidence for both construct validity and reliability of the 16 problems. Seven tables present study findings. (Author/SLD)

To generate a useful means of analyzing think-aloud protocols of individual students, and to determine the extent to which practical information might be derived from transcripts and videotapes, this project focused on the problem-solving behaviors of 17 sixth-grade students videotaped while solving routine textbook word problems. Using a checklist derived from prior research, problem-solving strategies were grouped in the following categories: orientation; organization; execution; and verification. Incidents of strategy use were calculated as percentages across individuals in the sample, and each category was submitted to a separate analysis of variance. Analysis allows determination of the effect that strategy use within a category had on strategy use across individuals. Within a category, particular strategies proved more natural or accessible for all students. For example, use of vocalizing calculations was the most common of the execution strategies. An opportunity for self-correction seemed to play a significant role in determining problem-solving behaviors. Overall, application of the strategies checklist proved effective for analyzing the protocols, and actions were effective for classifying behaviors. Checklist data were suitable for analysis, and the analysis provided useful information. Two figures and three tables provide study data. (Author/SLD)

This study considers the problem of performing all pairwise comparisons of column means for a two-by-four additive nonorthogonal factorial analysis of variance (ANOVA) model where cell variances are heterogeneous. Extensions of the following procedures are considered: (1) Games-Howell (1976) procedure; (2) the C. W. Dunnett (1980) T3 and C procedures; (3) the B. S. Holland and M. D. Copenhaver (1987) technique; (4) the A. J. Hayter (1986) procedure; and (5) the James (1951) second-order test. Using computer simulated data, Type I error rates and statistical power for these multiple comparison procedures are estimated. Sixty-six combinations of sample size and variance patterns are examined. The results suggest that these procedures maintain the familywise Type I error rate under the nominal 0.05 level. In terms of statistical power, the Games-Howell procedure generally provides the greater any-pair power, but the extension of the Hayter technique provides greater average power per contrast as well as in identifying all significant pairwise differences. Fourteen tables present details from the analyses. (Author/SLD)

The model of feedback processing proposed by R. W. Kulhavy and W. A. Stock (1989) was studied in a traditional classroom setting in which methods of assessing students' response confidence as predictors of postfeedback performance were also examined. The relationship between confidence ratings at the time of the test and confidence assessed prior to delayed feedback was explored. Subjects were 6 male and 21 female undergraduates assigned to confidence or interest conditions who rated their confidence in responses or interest in each questions. Data for 25 students were used. Students were asked to predict their scores, received feedback, and completed the examination again. An analysis of variance compared the performance of the confidence and interest groups, and regression and correlation analyses explored the predictability of postfeedback performance. There were no significant differences between postfeedback performance of the interest and confidence groups. Increase in elaborative processing due to students' rating their confidence does not appear to affect postfeedback performance any more than does rating the interest level. Results indicate that the Kulhavy and Stock model can be applied to the classroom. Use of students' estimates of test scores is not recommended as a measure of response confidence, as it accounted for very little variance in postfeedback performance. Implications for prediction of students' feeling-of-knowing are explored. (SLD)

Two ethnic groups of students, majority and minority, were compared on variables of a Psychological Productivity Model. Black and Hispanic students were grouped as minorities, and White students were considered majority students. Participants were 120 high school students from an urban high school in the Southwest. The Psychological Productivity Model entailed seven variables (home, social environment, time on task, motivation, quality of instruction, peer group, and media). Sex differences on the variables were also examined. Descriptive statistics and seven separate two-way analyses of variance were performed on the data. The variables Ethnicity and Sex both showed significant main effects on the variable Motivation, favoring the majority ethnic group and males, respectively. The results, however, fail to reveal any significant interactions. No significant main effects for the variables Ethnicity and Sex were found on the remaining six dependent variables. The results of the study reflect the need for further and more intensive, in-depth research on ethnic and sex differences on motivation. Results have implications for curriculum design. Five tables present study findings. (Contains 24 references.) (SLD)

Given that theory is a crucial component of path analysis and that major theories in the social sciences either directly or by inference assume interaction, it appears that interaction has to be considered in path analytic models that reflect those theories. The use of interaction within the framework of path analytic methodology is investigated by sampling articles reporting the use of path analysis over the last 5 years. The Type VI error, the focus of the investigation, has been identified as the incorrect use of post hoc procedures and the examination of a research question using one or a limited number of designs, studies, or analyses. Fifty articles on path analysis were identified in the ERIC educational reference system. Of the 50, 19 did not test for interaction in any way, and they did not compare separate models. Twenty-six used a multiple model approach, although it was not made clear that the use of multiple models was a reflection of concern for possible interaction effects. In eight studies, analysis of variance was used prior to the inclusion of variables, and in some cases, interactions, in the model. Sixty-eight percent of the articles appeared not to consider interaction at all, even though it is fair to assume that interaction is implied or explicitly stated in the theories underlying all of the models tested. This suggests that Type VI errors are being made in a majority of studies using path analysis. An extensive table lists the articles reviewed and the methods they use. (Contains 49 references.) (SLD)

Two analyses of variance (ANOVA) models for item scores are compared. The first is an items by subject random effect ANOVA. The second is a mixed effects ANOVA with items fixed and subjects random. Comparisons regarding reliability, Cronbach's alpha coefficient, psychometric inference, and inter-item covariance structure are made between the models. When considering the inter-item covariance structures for the two ANOVA models, brief comparisons with factor analysis models are also made. It is concluded that inference from a sample of items to a population of items requires homogenous inter-item covariances, that reliability has different meanings under the two models, and that while coefficient alpha is a lower bound for reliability under the second model, it is not under the first. (Contains 51 references and two tables). (Author/SLD)

To help teachers discover effective strategies appropriate for ethnic minority students, a study examined whether readers better understand material related to their own culture, and whether thematic units are an effective method for building culturally related schemata. Within a pre- and posttest design, a class of 25 Scottish pupils and a class of 21 Texan students (ages 8 to 10 years) listened to a same-culture passage and answered a 10-item multiple-choice test. The groups then listened to an other-culture passage and answered a similar test. Students/pupils participated in a one-week other-culture thematic unit and took the other-culture test. T-tests showed no significant difference between same-culture and other-culture pretest scores for the Texan group, but the Scottish group scored significantly higher on the same-culture test than the other-culture test. Repeated measures of analysis of variance supported the effects of instruction with higher posttest that pretest scores for both groups. (Three tables of data are included; 23 references and 5 appendixes of data are attached.) (Author/RS)

This paper reports on a study investigating the effects of studying alone or in cooperative-learning groups on the performance of high and low achievers. The effects of completing computer-based instruction using either learner or program control are also examined. A total of 92 sixth-grade students were classified by Stanford Achievement Scores and randomly assigned to group or individual treatments, stratified by achievement scores. Students completed training to enhance small group interaction before completing the computer-based tutorial and posttests. Both high and low achievers in the cooperative treatment increased achievement on program-controlled and learner-controlled computer lessons. The learner-controlled cooperative learning group made more options while checking their concept learning and spent more time interacting with the learner-controlled, computer-based tutorial than the learner-controlled individual learning group. Six statistical tables are appended. (Contains 13 references.) (Author/KRN)

This study investigated the reading preferences of American Indian and Anglo children with respect to ethnicity of main character, gender of main character, and setting in children's stories. The 160 subjects consisted of 80 American Indians and 80 Anglos from Grades 2 and 3, divided evenly by gender and grade level. The study yielded significant preference differences for ethnicity of main characters, gender of characters, and type of setting. Both Indian children and Anglo children preferred to read about Indian characters over Anglo characters. A majority of all subjects indicated a preference to read stories with both an American Indian and an Anglo as main characters, rather than stories with two Indians or two Anglos as main characters. Males preferred to read stories with at least one male character over stories with only females; girls preferred to read about two boys instead of a boy and a girl. Both Indian and Anglo subjects preferred stories set in a suburban-type setting over reservation-like settings. Three statistical tables and sample illustrations used in the stories for ethnicity, gender, and story setting are appended. (Contains 14 references.) (Author/KRN)

The purpose of this study was to investigate the effect of cooperative learning and type of reward on performance and continuing motivation. Subjects used either a cooperative or individual learning strategy while receiving information, examples, practice, and feedback from an instructional television lesson. Subjects (126 undergraduate education majors) were also provided with one of the three rewards (i.e., task, performance, none). Results indicated that subjects who worked alone performed better on the posttest and expressed more continuing motivation than those who worked cooperatively. Type of reward did not affect performance or continuing motivation. Implications for reemploying cooperative groups in settings that were originally designed for individual learning are provided. Two tables containing performance and motivation test score data are appended. (Contains 18 references.) (Author/KRN)

This research examines the performance, achievement, and attitudinal effects of a dissection alternative, an interactive videodisc-based (IVD) simulation in two ways: as a substitute for dissection and as a preparatory tool used prior to dissection. Sixty-one high school students enrolled in three general ability high school biology classes participated in this research over a 4-day period. On the substitution issue, findings suggest that the IVD simulation was at least as effective as actual dissection in promoting student learning of frog anatomy and dissection procedures. On the preparation issue, it was found that students using the IVD simulation as a preparation performed a subsequent dissection more effectively than students receiving no preparation and more effectively than students viewing a videotape as preparation. Students using the IVD simulation as preparation also learned more about frog anatomy and dissection procedures than those who dissected without preparation. Students in all groups evidenced little change in attitudes towards dissection. All students reported a significant gain in dissection self-efficacy, but no between-group differences were found. Findings are discussed relative to their implications for educational practice and future research. Six tables containing results of a survey on student attitudes towards dissection, achievement, attitude, self-efficacy and dissection performance test scores, and analysis of variance results are appended. (Contains 16 references.) (Author/KRN)

This study compared the effects of instructor-provided vs. learner-generated analyses of semantic relationships between major concepts on structural knowledge acquisition in an introductory psychology course. One group explained the relationships on instructor-provided graphic organizers; the other group classified relationships between concepts by completing skeletal maps provided by the instructor. Increase in structural knowledge was related to practice effect for both groups; the group that generated analysis scored higher at a near-significant level. Regardless of group, there was a significant relationship between total exam score and frequency of completing the treatment exercises (an indicator of motivation). Eight graphs displaying the statistical findings are included. (Contains 19 references.) (Author)

The present study was conducted to investigate the effects of peer collaboration and individual study as modes of practice, using the lesson content of a required introductory accounting course. The study also examined the effects of using content organizers on the practice worksheets. Data were analyzed using a 2 x 2 analysis of variance comparing posttest performance across practice modes (peer collaboration and individual) and content organizer treatments (organizers versus no organizers). Data from the attitude questionnaire were analyzed for variations in responses between the peer collaboration and individual work groups. Students showed a significantly higher level of achievement using the practice mode of peer collaboration, as well as having significantly more positive attitudes towards this method of classroom practice in problem solving. Several factors may have contributed to the positive results for peer collaboration. These include providing and explaining the answer, obtaining feedback and discussion, and playing an active role in learning. (Contains 8 references.) (Author)

Using a hypothetical data set of 24 cases concerning opinions on contemporary issues on which Democrats and Republicans might disagree, concrete examples are provided to illustrate that canonical correlation analysis is the most general linear model, subsuming other parametric procedures as special cases. Specific statistical techniques included in the analysis are "t"-tests, Pearson correlation, multiple regression, analysis of variance, multiple analysis of variance, and discriminant analysis. The discussion is aided by an initial explanation of the logic of canonical analysis. The equivalence of results, common features of the statistical methods, and the superiority of canonical methods are highlighted. Similarities between the canonical technique and other univariate and multivariate procedures are emphasized. Twelve tables present analysis results. An appendix lists the Statistical Analysis System commands for all parametric tests. (Contains 22 references.) (Author/SLD)

The Gang Membership Inventory (GMI) was developed for use with urban high-risk youth and administered as part of a larger survey. The GMI consists of 15 true/false items asking about the respondent's involvement in gang activities in the year just past. Activities range from peripheral involvement to more serious involvement in gangs (e.g., being an actual gang member). The GMI was administered in English or Spanish to fifth through eighth graders at two Chicago (Illinois) elementary schools in predominantly Latino neighborhoods. Usable surveys were obtained from approximately 650 children. Forty percent of the respondents self-identified themselves as Puerto Rican, 28 percent as Mexican, 5 percent as both Mexican and Puerto Rican, and 7 percent as other Latino ancestry. Principal components analysis with varimax rotation produced three factors with eigenvalues greater than 1.0 that together accounted for 50 percent of the total variance. Because the factors do not appear to be well-defined, it was decided to retain the measure as a single scale. Validity of the GMI was assessed through a one-way analysis of variance and other measures. Also examined was the relationship between gang involvement and the subscales of the Teacher-Child Rating Scale. As expected, youth rated high or moderate on acting-out behavior were significantly more gang involved than were youth low in acting out. The GMI has been shown to be easy to administer and useful in determining different levels of involvement in youth gangs. One table displays GMI factor loadings, and six figures describe relationships of various variables to gang involvement. An appendix contains the GMI. (SLD)

The Test of Spoken English (TSE) is an internationally administered instrument for assessing nonnative speakers' proficiency in speaking English. The research foundation of the TSE examination described in its manual refers to two sources of variation other than the achievement being measured: interrater reliability and internal consistency. Because the reported data were drawn from a 1980 study of the reliabilities based on two raters, newer data and more extensive analyses were needed. This study uses data from recent administrations of the TSE examination. Analysis of variance examined the effects of scale, section, examinee, and rater, as well as the interactions of these factors. Reliabilities were reported for item, section, and scale scores. Common and unique variance estimates were developed for each scale score for which a section was rated. Estimates of the effects of altering section lengths suggested that some sections should be lengthened and others omitted if reliability were to be maximized. Others suggestions are offered for improving reliability. Thirteen tables present study findings. An appendix describes TSE scale points. (Contains 16 references.) (SLD)