INTRODUCTION

As part of a comprehensive OSEP program of longitudinal research related to the 1997 Individuals with Disabilities Education Act (IDEA97), the Special Education Elementary Longitudinal Study (SEELS) is intended to involve a large, nationally representative sample of students in special education who are ages 6 through 12 when the sample is selected (fall 1999). The SEELS student sample will be generated through a two-stage process: (1) selecting a nationally representative sample of LEAs stratified by region, size, and district wealth, and (2) selecting students in specific disability categories from the selected LEAs to participate in SEELS. The SEELS LEA sample was selected in January 1999. This report was intended to be the final product that documents the process and outcome of the LEA recruitment from the SEELS sample. However, for a variety of reasons, the recruitment process has taken longer than originally planned, and this report must be an interim sampling report because the recruitment will need to be extended through summer 1999 and into the next school year to reach the target number of LEAs. The report addresses a number of topics, including a description of the desired LEA sample characteristics, LEA recruitment activities, current and projected levels of LEA participation, problems encountered, and planned activities.

The SEELS LEA sample must be drawn from a sampling frame that reflects the nation with regard to region, LEA size, and LEA wealth. In addition, the size of the LEA sample must be sufficient to generate the needed student sample (n=14,375). Our analysis of LEAs’ estimated enrollment across district size, and estimated sampling fractions for each disability category, suggested that 297 LEAs (and as many state-sponsored special schools as will participate) would generate sufficient numbers of students. However, previous experience with the NLTS suggests that significant numbers of LEAs either decline to participate because of concerns related to confidentiality of student records or simply never respond to communications from researchers. Although considerable time and effort were expended in recruiting LEAs for the NLTS, approximately 55% of the LEAs invited to participate declined, did not respond, or introduced procedures that unacceptably lengthened the recruitment process. Thus, assuming a similar level of nonresponse/refusal for SEELS, we selected a sample of 765 LEAs (and a reserve sample of 380 LEAs) in order to obtain our target of 297 LEAs who will agree to participate in the study. In addition, the universe of state schools for deaf students and for blind students was included in the sample.

To generate the appropriate number of students, the LEA recruitment effort was carefully designed to be a sequential, two-stage process resulting in a student sample in time for the first planned data collection (the SEELS parent interview). The timeline of the SEELS LEA and student sample is described in Table 1. Note that the dates below reflect the timeline as originally planned. Each of the activities experienced a number of delays and took longer than planned.

LEA recruitment proceeded sequentially from OSEP communication to states, to SRI/Westat communication to states, to SRI communication directly with LEAs. In preparation for these communications, informational materials were developed for SEELS, including letters, a brochure, and a Web Site.

• Initial SEA contact from OSEP. A memo from OSEP was sent to chief state school officers announcing the three studies associated with the national assessment of IDEA97 and soliciting state support for the studies (see Appendix A).
• Initial state special education director contact from SEELS and SPeNSE. This task was a coordinated effort between SRI and Westat. A letter was sent to state directors of special education describing SEELS and SPeNSE. The letter included study brochures, lists of selected districts for both studies, a list of very large LEAs selected for both studies, and a form requesting contact information (e.g. special education director name, telephone, mailing address) for individual LEAs. This letter was followed by telephone contact with the state director or his/her designee by senior project staff from SRI or Westat. During these conversations, directors could ask questions or provide state-specific information, preferred LEA contact procedures, and state-level support. In addition, states were presented with the possibility of augmenting the national sample with additional LEAs to form state-level samples. Interested states were later provided with estimates of the number of additional LEAs that would be required, as well as estimates of the costs associated with such samples.
• Direct mail to LEAs. Informational packets containing an introductory letter, a study brochure, a letter from the director of the Family Educational Rights and Privacy Act (FERPA) of the U.S. Department of Education allowing contractor access to student level data, and participation agreements were sent to all LEAs (Appendix B). LEAs had the option of returning participation agreements by fax or by prepaid mail. Because state contacts took somewhat longer than expected, the first mailing was conducted in two stages on April 24, 1999, and May 1, 1999. To increase LEA response, a second mailing was sent to all uncommitted LEAs on June 10, 1999.
• Telephone follow-up. The procedure of following mail contact with a telephone contact from project staff–providing personal contact, answering questions, clarifying issues–can be an effective method of increasing positive responses. However, the large number of districts in the SEELS LEA sample precluded telephone follow-up with each sampled LEA. Therefore, our telephone recruitment efforts have focused on very large LEAs, which are relatively few in number but from which a relatively large proportion of sample students will be selected. Smaller LEAs are more numerous and therefore have a large number of potential replacements for refusing districts. Although this strategy is likely to be most efficient in selecting the LEA sample quickly, there is a risk that smaller LEAs who refuse to participate differ systematically from other LEAs in terms of the types or effectiveness of programs that they offer to students. Thus, detailed tracking will be necessary to identify potential patterns that emerge with regard to LEA refusal/nonresponse.
• Tracking. A relational database was created to track various aspects of the state and LEA recruitment efforts, including contact information, current recruitment status, log of contacts with state or local personnel, notes, and recruitment summary fields.

Approximately 2 months have elapsed since our initial contact with LEAs. Table 2 describes the current status of the SEELS LEA recruitment effort in terms of the number of uncommitted LEAs, LEAs who have agreed to participate, and LEAs who have declined to participate. These data are presented separately for very large LEAs and for state schools for deaf students and blind students. Small, medium, and large LEAs have been grouped together for this report. In general, our assessment is that the recruitment process is going relatively well but that significantly more time will be required to meet the goal of 330 LEAs. In fact, the rate of SEELS LEA recruitment is very similar to that of the NLTS in the amount of time required to complete the process. The current status of the recruitment effort (6/24/99) can be summarized in the following points:

• Special schools. Recruitment has been very successful among special schools, where 26% of the universe have agreed to participate and only 1% have declined. This is much higher than we had expected.
• Small, medium, and large LEAs. Approximately 10% of sampled LEAs in these strata have agreed to participate in SEELS, and 4% have declined.
• Very large LEAs. Approximately 11% of sampled LEAs in this stratum have agreed to participate, and 3% have declined.
• Nonresponse. The largest category is that of nonresponse (80%). These districts’ ultimate decisions to participate or decline are unknown.

In addition to judging our progress in LEA recruitment, it is important to make projections regarding the amount of additional time that will be required to complete the recruitment process. To accomplish these projections, we estimated the number of LEAs who will agree to participate on a monthly basis according to the following assumptions:

• Early agreements are easier. We have assumed that the LEAs who agree to participate early in the process require less effort than those who decide to participate later. This assumption reduces our expectation of the number of participation agreements we expect to receive in the out months.
• Summer is a slow time. We know that many LEA business offices remain open for at least some of the summer months. However, some staff who are in decision-making positions may take vacation or may prefer to wait until school resumes in the fall to consider participation.
• School openings are slow for recruitment. The beginning of new school years are hectic times for teachers, students, and administrators. During this time, it can be difficult to get the attention of district personnel to consider participation decisions.
• Increases in October and November. We have assumed that our recruitment rate in October and November will improve as the school year progresses.
• Linear increase in proportion of LEAs who decline. We have assumed that the proportion of LEAs who decline to participate will increase linearly over time.

According to these assumptions, Table 2 suggests that it will be necessary to continue the recruitment effort into fall 1999 to reach the target number of LEAs. Of course, these estimates represent our "best guesses" about the how the recruitment process will proceed. The timeline will vary to the extent that our assumptions are correct. Below we analyze the implications of the extended time required to select the sample.

• Shift in the nature of requests to LEAs. Current communication to LEAs requests only a commitment to participate in SEELS and to provide student rosters in the fall of 1999. As LEA recruitment continues into the fall, it will be necessary to request LEA participation and student rosters at the same time. Although this may have some impact on refusal rates, this process was used in the NLTS and was successful in that case.
• Contractor coordination. The original schedule specified the LEA recruitment process as the responsibility of SRI and RTI during the SEELS design phase. The student sample selection was to be the responsibility of the SEELS implementation contractor. The overlap in the two selection processes will require that the contractors coordinate these recruitment efforts closely.

We believe that the LEA recruitment process is well under way and will ultimately generate the required number of students for the sample. However, we believe that an explanation of events and processes that led to the delay is desirable and will be informative for future efforts.

• State and LEA variability. SEAs and LEAs were quite variable in their response to our requests. In some cases, a single phone call or mailing was sufficient. In other cases, additional procedures, mailings, presentations, and telephone contacts were required. In some, we were unsuccessful in establishing contact at all.
• Application LEAs. A number of LEAs, usually from the very large stratum, have lengthy and detailed application processes for entry into the LEA for research purposes. We have underestimated the level of effort that has been and will be required to gain participation from these LEAs.

To achieve the LEA recruitment goals, we propose the following activities and decision rules:

• Continue telephone follow-up. We will continue efforts at telephone follow-up in very large LEAs. We will consider further telephone follow-up for smaller LEAs if the rate of participation declines significantly.
• Monitor progress. We will continue to evaluate the success of our efforts on a monthly basis to establish the reasonableness of our assumptions and to adjust them accordingly.
• Evaluate fit of LEA sample to sampling frame. An important step in evaluating the adequacy of the LEA sample is to establish how well it fits the sampling frame from which it was selected. It is currently premature to conduct such an analysis because too few LEAs have agreed to participate. However, we will conduct this analysis when we have approximately 75% of the LEA sample, which will be in mid to late September, according to our assumptions. If we observe that participation rates are uneven in particular strata, we can intensify follow-up activities for LEAs in those strata and/or consider using the reserve sample.
• Reserve sample. At the same time that the core SEELS sample was selected, a reserve sample was also selected that contains half the number of LEAs in the regular sample. If our participation rates fall significantly and we believe that we are in danger of not reaching our recruitment goals, we may elect to use the reserve sample. The disadvantage of doing so is that additional state contact would be required to obtain at least some of the contact information, which was a time-consuming process in and of itself.
• Report to OSEP regarding progress. We will provide monthly updates to OSEP regarding progress in recruitment, problems encountered, and corrective actions proposed.

The SEELS LEA recruitment process is proceeding more slowly than originally planned in the design task, but approximately at the same rate as our previous experience in the NLTS. We expect that the proposed strategy will result in an LEA sample by the end of 1999 that will both generate sufficient numbers of students and not delay the SEELS data collection timetable.