The project aims at helping middle school students understand the relationship between energy, economics, and climate change by monitoring home-energy consumption. Students use energy-monitoring equipment to assess the amount of stand-by power consumed by their home appliances and entertainment devices when these units are powered off. Service-Learning is the theoretical foundation underlying the project, which is based on a student empowerment model that values their contributions to solve problems. The project seeks to investigate the necessary conditions to expand learning opportunities and outcomes from a previous project with sixth graders in new learning environments with diverse ethnic groups, rural areas, alternative schools, and different climate zones. The scale-up plan (Dede, 2006) addresses depth (deep belief change), sustainability (maintaining the innovation over time), spread (diffusion of the innovation to a large number of classrooms), shift in reform ownership (comes to be owned and maintained by the local school), and evolution (project revision by a reflective community of practice over time). To accomplish its goals, the University of North Texas partners with Numedeon Inc. to provide curriculum development services through Whyville, a learning-based virtual environment about the science and mathematics of energy.
The key research question is: What are the necessary conditions to scale-up the positive impacts of STEM content knowledge, dispositions, and affinity for careers from a previous project to a wider audience of diverse populations and learning environments representative of the national student population? The setting of the project is 24 middle schools in seven states (Maine, Vermont, Virginia, North Carolina, Louisiana, Texas, and Hawaii). The G-Power, version 3.1, is used to estimate power of the proposed analyses using individual repeated measures ANOVA analyses. Assuming the projected sample size of 1,400, with á=0.05 and ES=0.05, power is estimated at 0.84. In Year 1, teachers (n=8) and students (n=192) will be the treatment group, while 8 new classrooms with the same sample sizes will be the comparison groups. In Year 2, teachers (n=16) and students (n=384) will be the treatment group, and 8 classrooms will be the comparison groups. In Years 3 and 4, treatment groups will consist of teachers (n=24) and students (n=576), and comparison groups will include teachers (n=16) and students (n=384). The impact of the project is analyzed at the student and classroom levels using a quasi-experimental design. To reduce the risk of confounding variable due to individual differences, each student from the comparison group will be matched with a corresponding student from the treatment group using propensity score matching techniques. The project performs a logistic regression analysis to produce a propensity score for each subject using a weighted combination of covariates chosen for the theoretical influence that they could have on the systematic differences between the treatment and control groups. Treatment and comparison classrooms are also matched so that subsequent analyses may be performed at the classroom level. Because sample sizes are small, the project use covariates, including school location, school-level SES, percentages of students by gender, and average science achievement to match closely equivalent classrooms in the treatment and comparison groups. Data gathering using valid and reliable instrumentation includes (a) the STEM Semantics Survey, a 25-item semantic differential instrument containing five scales assessing perceptions of Science, Technology, Engineering, and Mathematics; (b) a Likert-type Career Interest Questionnaire of 12 items on three scales; (c) a second Likert-type instrument from the TIMSS 2007 Mathematics Study (IEA, 2007); (d) the National Geographic Vampire Power Quiz to assess STEM content gain; and (e) the Career Aspirations Survey to detect changes due to project activities (Nolte & Harris, 2010). Data are analyzed at the matched-student and classroom levels using within-subjects analyses. Differences between measures collected before and after participation in the project activities are analyzed to determine the level of impact that the project has on participants. Data interpretation strategies include quantitative data analysis techniques, such as repeated measures analysis of variance for comparing gains by treatment students to gains by comparison students. Also employed are Hierarchical Linear Modeling to separate out the differential effects of the intervention at differing school sites, as well as multivariate techniques, such as General Linear Modeling for comparisons of the time series (three-point) growth lines of treatment students to the same time-series points for trend lines of the comparison groups. Qualitative data gathering strategies from interviews and surveys data by site, date, and project objective are used to explain the why-type of questions emerging from quantitative data analyses.
Project outcomes are: (a) a research-informed and field-tested prototype focused on energy conservation for middle school students; (b) data on the knowledge, skills, and dispositions needed by middle school students to become stewards of Planet Earth; (c) relevant instructional materials; and (d) a set of valid and reliable instruments.
