The SEELS sample, research agenda, and data collection schedule make SEELS an especially ambitious study. The study must be equally ambitious with regard to analysis and dissemination so that the generated information will be of maximum use to as many audiences as possible. Specifically, the SEELS analysis and dissemination strategies need to address the following issues:

• Range of audiences. SEELS will create a wealth of new information that will be of interest to many audiences, including parents, teachers, administrators, related service professionals, policy-makers, advocacy organizations, and researchers. SEELS will need to consider both the content and presentation of information that suits particular audiences best.
• Range of information needs. Related to the variety of audiences, the study will need to address a range of information needs to maximize its usefulness. For example, reports documenting the study’s technical details, comprehensive reports, executive summaries, briefing materials, one-page descriptions, peer-reviewed journal articles, a World Wide Web presence, and special topic reports all are likely to play important roles in communicating the study’s results.
• Types of analyses. Data generated from SEELS will need to support a range of analytic purposes:

• Descriptive. One of the most important analytic tasks of the study will be to describe students in special education at the applicable ages, their background and characteristics, the education they receive, and the outcomes they achieve. These descriptions will need to be weighted so as to represent the national population of students in special education in elementary schools as a whole and by age and disability group.
• Relational. An additional analytic purpose will be to explore relationships between various contexts, characteristics, practices, and outcomes. These relationships may need to be examined by a variety of subgroups defined by disability, age or grade, gender, ethnicity, or other factors.
• Explanatory. An important purpose of the SEELS analysis approach will be to explain the contribution of individual factors and combinations of factors to variation in outcomes.
• Comparative. Many findings are most powerfully understood when placed in the appropriate comparative context. SEELS findings will gain important perspective in their ability to be compared with other national data collections, particularly for peers without disabilities. Items that have been used in national surveys are being included in the SEELS design so that comparison between databases will be possible. Currently, items in SEELS protocols have been borrowed from the Schools and Staffing Survey (SASS), National Health Interview (NHI), National Early Intervention Longitudinal Study (NEILS), and Early Childhood Longitudinal Study (ECLS).
• Longitudinal. Repeated measures over time offer the opportunity to examine and explain changes in student behaviors and outcomes, as well as changes in factors that could influence them, such as school programs and family contexts.

• Range of media. The variety of ways in which people access information has increased greatly over the last decade. This represents a great opportunity for SEELS to communicate both progress on study activities and study findings. Possible avenues include CD-ROM and the World Wide Web to supplement traditional media.

Given the range of likely SEELS data analysis and dissemination activities, it is appropriate to refer to SEELS analysis and dissemination as consisting of multiple interrelated strategies rather than a single strategy. Thus, the approach here is to propose analysis and dissemination strategies that link likely study analyses and their products to the study’s data collection timeline. This approach provides a general framework that will allow for thoughtful planning about analysis and dissemination, but one that is sufficiently flexible to accommodate changes in terms of particular topic areas that may be of interest over the course of the study. Thus, a variety of documents are suggested that are linked to specific waves of data collection, including comprehensive reports, technical reports, and special topic reports. Table 5-1 shows illustrative topics for the reports, as well as their possible timelines and data sources. These topics are placeholders; the actual content addressed in many of the documents may well be different in the implementation of the study.

* PGI=parent interview; LATS=teacher survey; SPS=school program survey; SBS=school background (principal) survey; DA=direct assessment; SS=student survey.

In addition to these formal reports, professional journal articles that report segments of analyses from the larger reports also should be produced.

SEELS will use a variety of statistical and analytic methods to generate this variety of products. Likely methods to be used include:

• Weighted frequencies, cross-tabulations, and summary statistics. These tools provide descriptive information in conjunction with standard errors to estimate their degree of precision.
• Exploratory data analysis. The graphical tools used in EDA are especially useful for uncovering patterns in datasets and among subsamples of the data.
• Correlational analyses. Simple and multiple correlation coefficients for continuous, dichotomous, and ordinal data allow investigation of relationships among variables in comparison with both statistical standards and the relative strength of specific relationships across subgroups.
• Multiple regression. The linear combination of variables to predict and explain variation in a continuous dependent variable.
• Logistic regression. The linear combination of variables to predict and explain variation in a dichotomous dependent variable.
• Hierarchical linear modeling. Multi-level HLM allows the construction of models that sort factors into conceptually logical strata.
• Structural equation modeling. This method may be useful to construct and test models among underlying (as opposed to measured) factors.

BACK

NEXT