This memo documents discussions that occurred during two initial meetings for Task 6 of the contract for Technical Support for the Department of Educationís Initiative to Link Research and Practice to Improve Results for Individuals with Disabilities. Task 6 involves designing a longitudinal study of students in special education who are ages 8 through 12 at the outset of the study. The first of the two meetings involved OSEP and Task 6 staff and focused on Task 6 alone, whereas the second afternoon meeting involved staff of Tasks 7 (design of a longitudinal study of the state and local implementation of IDEA) and 8 (professional development of professionals serving students with disabilities) and focused on coordination among the three longitudinal studies.

Attendees: Renee Bradley, Lou Danielson, Judy Holt, Kelley , OSEP

Jose Blackorby, Mary Wagner, SRI

Tom Fiore, RTI

Renee Bradley opened the meetings with introductions. Lou Danielson then gave background information on OSEPís intentions to think through and implement a program of longitudinal research in the coming years that would provide information on the full age range of students with disabilities. The ongoing National Early Intervention Longitudinal Study (NEILS), being conducted by SRI and partners, including RTI, provides a look at infants and toddlers with disabilities who are in early intervention through age 5 (or perhaps 7). The elementary longitudinal study, which subsequent discussion dubbed the Special Education Elementary Longitudinal Study (SEELS), will illuminate the experiences of 8 through 12 years olds in special education over time. Lou described OSEPís intention to consider how these studies might fit within a broader program of research and the need for the elementary study design to reflect this bigger picture and to be flexible to the extent possible in linking with or feeding into other studies in the program.

Subsequent discussions are organized around the design subtasks.

The meeting itself and this report constitute completion of this subtask. However, given a start date for the task of August 24, the delay of the meeting until September 21 makes these activities and those of several subsequent tasks out of compliance with the original schedule. Judy suggested recasting the deliverable schedule in terms of an "effective start date" 2 weeks after the contractual start date, or September Costs will still be covered from the contract start date, but the deliverable schedule will be referenced from the effective start date. Judy and Tom will discuss any contract modifications that may be required to revise the deliverable schedule.

In a telephone discussion on September , Judy provided Jose and Mary feedback on OSEPís preferences regarding potential advisory panel members. SRI staff have allocated responsibility among themselves to contact and recruit preferred members in the coming week.

In the same telephone conversation on September , Judy provided Jose and Mary feedback on OSEPís preferences regarding the task force membership. SRI staff have allocated responsibility among themselves to contact and recruit preferred members in the coming two weeks. The group discussed alternative task force meeting dates and determined that October 6 met the schedule requirements best.

Timeline. Considerable discussion focused on some key parameters of SEELS, particularly length and funding level, and their implications for data collection cycles. Lou reported that SEELS could be assumed to extend beyond a single 5-year contract period, with the optimum total length being informed by upcoming work on thinking through the larger program of longitudinal studies. Total funding for the entire program of longitudinal research (potentially as many as four concurrent studies) is unlikely to exceed $8M per year. Given the interest in multiple studies with nationally representative (therefore, large) samples, and this funding limit, it is unlikely than yearly data collection would be affordable for SEELS. Biannual data collection is possible, as long as the studies stagger their heavy data collection years.

Another consideration regarding timeline is the expected start date for awarding the contract to conduct SEELS. Lou reported that a January 1999 start date was currently projected. Mary pointed out the importance of starting very early in the school year to permit both the lengthy process of selecting the student sample from participating districts and the first wave of data collection to be carried out in a single school year. After discussion, OSEP staff agreed that effort would be made to move up the procurement process but that there was no guarantee that it could be completed by September 1999. A fallback position that is an attractive option is to amend the task order to include selecting the student sample in the task so that it may proceed in September, and proceed with a January award of the contract to implement the study, with data collection beginning as early as March 2000.

Sampling Plan. Mary and Jose briefly outlined an intention to follow the NLTS sampling approach but mentioned that a larger initial LEA sample would be needed because of the high probability of districts refusing to participate. Other sampling issues were discussed in the study coordination meeting (see below).

Data collection plan. We considered the issue of measuring studentsí academic performance and the importance of that outcome domain for SEELS. The preceeding longitudinal study of secondary school students (NLTS) had educational outcome measures such as grades and school completion that are not appropriate for the age group in SEELS. Lou expressed an interest in seeing some form of direct student assessment included in SEELS, at the same time acknowledging the high cost of extensive assessment being conducted by contractorsí field staff. Jose and Mary discussed an option of paying teachers (or other district staff, as suggested by Tom) to conduct brief assessments of students, such as a reading test (given that reading ability in elementary grades is a powerful proxy for academic achieve in those grades and predictors of academic achievement in later grades). Design work will consider options for assessing academic performance, within the cost parameters mentioned above. Other key issues mentioned were access to the regular education curriculum, the use of supplemental aids and services, and the relationship of both these factors to outcomes for students.

Another domain that was discussed as needing further development that had previously been accomplished involved functional abilities. All agreed that multiple domains would need to be assessed. Lou suggested that the PASS instrument might be a helpful guide in determining dimensions of functioning that could be assessed by a knowledgeable adult (e.g., teacher, parent), without direct assessment.

Discussion of outcomes was brief, generating agreement that the work of the National Center on Educational Outcomes was a useful framework for identifying outcome domains. Jose mentioned SRIís interest in having Martha Thurlow on the task force; OSEP staff agreed on the importance of someone from NCEO serving in that capacity and offered to help persuade NCEO staff of the importance of accepting our invitation, if needed.

Analysis Plan. Analysis issues were not discussed. We assume that the general outline of this aspect of the study that was presented in the proposal is acceptable as a starting point for further development.

The importance of coordination among the three longitudinal design tasks was stressed by Lou and recognized by all participants. Mary suggested that the upcoming special education expenditure study also be considered as an important point of linkage, recognizing that the later timeline for that study might make careful integration difficult.

Mary mentioned that this subtask had begun already, with an emphasis on the national education studies (e.g., NELS, NHES) and the National Health Interview Survey.

As mentioned above, Mary pointed out the extended time period likely to be needed to secure the participation of the required number of LEAs, with implications for starting implementation of the study early in school year 1999-2000. Other issues of coordinating contacts with LEAs were discussed in the afternoon session (see below).

Mary reiterated the discussion in the proposal of an intention to collect data from parents, teachers (about themselves and individual students), and principals, as well as needing data at the LEA level. Comprehensive identification of content areas for instruments must await the task force meeting. However, initial work on instrument development has begun, based on previous SRI experience and an initial review of other national surveys.

No specific discussion of this task ensued. (See below for discussion in the context of cross-study coordination.)

Discussion of this topic focused on the importance of assuring that all parties interested in potentially bidding on the contract to conduct SEELS have fair access to all the relevant information on the design of the study as soon as that information becomes available. Web presentation will be key to that access; Judy stated that the approach to the Web site specified in the task order proposal was acceptable.

Beyond Web access to materials, however, Tasks 7 and 8 include briefings of potential bidders by the design team on aspects of those designs and we discussed the possibility of including that approach in Task 6. In our discussion, we all agreed to the importance of assuring fair, impartial, and timely access to design information for all potential bidders. We agreed that one or more briefings for potential bidders could be included in Task 6, although they were not commissioned in the RFP or budgeted by SRI and RTI, as long as they could hold down costs by "piggybacking" on currently planned trips to Washington, specifically those for the data coordination meetings (months 4.5 and 7.5). We agreed to add briefings at these time points. However, discussions in the afternoon meeting regarding coordination (see below) resulted in a focus on data coordination through monthly conference calls among relevant teams and OSEP staff. Thus, we revisited the appropriate timing and number of briefings and determined that one briefing after the final products for this subtask are due (month 5.5) is the optimal briefing approach.

Representatives of Tasks 6, 7, and 8 met with OSEP representatives (above attendees plus Pat Gonzalez, Helen Thornton, and Dave Malouf) to discuss the possibility and ways of coordinating the three longitudinal studies. Lou suggested there were at least four levels of coordination possible: interactions with districts that are sampled for more than one study, sample overlap or integration, data collection integration and data sharing,

The entire group agreed on the potential benefits coordination and pledged good faith efforts to maximize coordination to the extent possible. However, at this early stage in the design work, it was not possible to identify with certainty the kinds of coordination that would make sense. One sampling parameter that supports the possibility of sampling overlap or integration is that all tasks involve an LEA sample as part of their approach. However, LEA sample requirements differ in that Tasks 7 and 8 require a sample of all LEAs, regardless of grade level, whereas SEELS requires a sample of LEAS that serve only 8 through 12 year olds (elementary and unified school districts). The group discussed various ways that a broader sample of all districts might be selected, from which a subsample of elementary and unified districts could be selected for SEELS. However, none of those present wanted to commit to any particular approach without doing the more thorough design work each study requires.

A logistical obstacle to coordination is the different timelines for the studies, which put such activities as sampling and contacting districts for SEELS ahead of similar activities for the other studies. It did not seem possible to speed up the timeline for the other studies to match SEELS because of the significant amount of conceptual work that needs to be done for them. Similarly, it did not seem possible to delay these activities for SEELS because of the importance of getting into the field to select the student sample early in the 1999-2000 school year (see timeline discussions above). Mary suggested that the SEELS staff stick with their timeline and draw the nationally representative sample needed for that study and that other studies could scaffold on it, adding other kinds of or more districts in the cells of whatever sampling frame they devise when they reach that stage in their work. The group agreed that was theoretically possible but again, could not to commit to any particular approach.

Given the inability to determine at this point any particular aspect of the studies that could be coordinated, the group focused finally on a process for coordination. We agreed that the two meetings during the design tasks called for in the RFPs would be insufficient to develop the kinds of coordination desired. We agreed on monthly conference calls on the third Thursday of each month and set dates for the coming months of October, November, December, and January. It is unclear whether some in-person meeting might be desirable at a later point, but all agreed that budgetary tradeoffs might limit this possibility (given that travel funds for Task 6 are being diverted for a briefing, additional trips may not be possible).