Search

The 6+1 Trait[R] Writing model (Culham 2003) emphasizes writing instruction in which teachers and students analyze writing using a set of characteristics, or "traits," of written work: ideas, organization, voice, word choice, sentence fluency, conventions, and presentation. The Ideas trait includes the main content and message, including supporting details. The model has not been adequately studied using experimental methods. In order to provide evidence on the effectiveness of this approach, the study reported here was designed as a large-scale effectiveness trial (Flay 1986). The study addressed the following confirmatory research question: What is the impact of 6+1 Trait Writing on grade 5 student achievement in writing? It also investigated two exploratory research questions: (1) What is the impact of 6+1 Trait Writing on grade 5 student achievement in particular traits of writing?; and (2) Does the impact of 6+1 Trait Writing on grade 5 student achievement vary according to student gender or ethnicity? Data for the cluster-randomized experimental study were collected from participating grade 5 teachers and students in 74 Oregon schools. The benchmark estimates indicate that use of the 6+1 Trait Writing model significantly increased student writing scores during the year in which it was introduced to schools. After controlling for baseline writing scores, the estimated average score of students in the treatment group was 0.109 standard deviations higher (p = 0.023) than the estimated average score of students in the control group. An intervention with this effect size would be expected to increase the average level of achievement from the 50th to the 54th percentile. The findings remained stable when tested using alternative choices for the analytic sample and the model specification. In addition to the analysis of holistic writing scores, exploratory analyses found statistically significant differences between control and treatment group students on three of the six specific outcome measures of particular writing traits--organization, voice, and word choice-- with effect sizes ranging from 0.117 to 0.144 (p = 0.031 to 0.018). For the other three traits-- ideas, sentence fluency, and conventions--the mean outcome score of students in the treatment condition was higher than that of students in the control condition, but these differences were too small to be considered statistically significant given the size and sensitivity of the experiment. Additional exploratory analyses of holistic writing scores found no differential effects of the intervention based on student ethnicity or gender. Appended are: (1) Professional Development Institute and Workshop Objectives and Agendas; (2) Statistical Power Analysis; (3) Test Administration Directions and Sample Student Essay Booklet; (4) Teacher Survey; (5) Scoring Rubrics for Student Essays; (6) Correlations between Holistic and Writing Trait Scores; (7) Interrater Reliability and Coding Details for Student Essays; (8) Technical Note on Multiple Imputation of Missing Data; (9) Alternative Analysis of Implementation Fidelity; (10) Complete Multilevel Model Results for Confirmatory Research Question; (11) Technical Note on Effect Size Calculations; (12) Complete Multilevel Model Results for Exploratory Analyses; and (13) Technical Note on Pooling of Effect Estimates. (Contains 3 boxes, 1 figure and 52 tables.)

This secondary analysis of survey data from a voluntary sample of 11,561 grade 3-8 students examines the prevalence and distribution of aggression, victimization, and approval of aggression, both overt (verbally and physically aggressive behavior intended to threaten or harm) and relational (behavior intended to harm someone's relationships with others). The report presents both descriptive statistics and the results of hierarchical linear modeling (all results are reported at the 0.05 level of statistical significance). The following are key findings: (1) On the Normative Beliefs about Aggression Survey, 1-12 percent of girls and 4-20 percent of boys in grades 3-5 reported that retaliation was "sort of OK" to "perfectly OK;" (2) For the study sample, school factors were associated with 1-7 percent of the variation in student survey scores, and student characteristics, such as gender and grade level, were associated with 93-99 percent of the variation; (3) On the Normative Beliefs about Aggression Survey, boys approved of aggression more than girls did both in general social situations and in situations involving retaliation; (4) On the Peer Experiences Questionnaire, students in grades 7 and 8 reported higher agreement with beliefs that endorsed aggression than did students in grade 3 when asked whether bullying "pays off," whether a student who gets bullied "deserves it," and whether a student should intervene if others are fighting; (5) For overt victimization, 12-61 percent of girls and 17-60 percent of boys reported being victimized at least once during the last 30 days, and 2-10 percent of girls and 3-14 percent of boys reported being victimized once or more a week, with the percentage varying by the behavior; (6) For relational victimization, 41-48 percent of girls and 31-42 percent of boys reported exposure during the last 30 days, and 4-6 percent of girls and boys reported exposure once or more a week, depending on the behavior; (7) For overt aggression, 3-37 percent of girls and 7-44 percent of boys reported perpetrating such acts during the last 30 days, and 0.4-2 percent of girls and 1-5 percent of boys reported perpetrating such acts once or more a week, depending on the behavior; (8) For relational aggression, 21-28 percent of girls and 20-24 percent of boys reported perpetrating such acts during the last 30 days, and 0.8-1 percent of girls and 1-2 percent of boys reported perpetrating such acts once or more a week, depending on the behavior; (9) Boys reported more overt victimization, overt aggression, and relational aggression toward other students than did girls; (10) No significant differences between girls and boys were found in the frequency of relational victimization; and (11) Differences between overt and relational aggression were also associated with grade level. Appendices include: (1) Study sample and methods; (2) Item responses for the total sample and by gender for the Normative Beliefs about Aggression Survey; (3) Item responses for the total sample and by gender for the Peer Experiences Questionnaire; (4) Results of hierarchical linear modeling of gender and grade-level differences associated with survey scale tables; and (5) Sensitivity analyses for imputed data. (Contains 32 tables, 1 box and 1 note.) [For "Student-Reported Overt and Relational Aggression and Victimization in Grades 3-8. Summary. Issues & Answers. REL 2011-No. 114", see ED518228.]