Sampling approaches to data collection are indicated when there are limited resources (financial and staff) and many sampling units (schools, students, and parents). With more than 435,000 students receiving special education services in over 8,400 campuses in Texas, the use of a sampling approach to examine indicators within the State Performance Plan (SPP) is essential.
Importantly, the sampling approach must still provide valid and reliable information. It is noteworthy that sampling can actually increase the value of information given the extreme variance in district and student characteristics that change from region to region and by age grouping. The use of purposive sampling while still adhering to a stratified random sampling approach allows for consideration of these variances.
The Texas sampling plan for SPP indicators has been approved by the federal Office of Special Education Programs (OSEP). The current plan considers prior experience with sampling within the special education program in Texas.
Specific Sampling Plans for SPP Indicators
There are over 1,200 school districts in Texas ranging in size from 20 to over 200,000 students, also with an expected large difference in the number of students receiving special education services. Following the OSEP approved sampling plan, districts will be included in sampling for the various indicators once every sixth year. Those included within each year were determined using consideration of geographic region and student demographics to yield representative findings. Large districts, with over 50,000 total number of students, must be included each year.
As already noted, there are over 8,400 campuses within Texas with some districts having over 100 separate campuses. As with district sampling, a framework for including campuses based on grade span (e.g., elementary, secondary) and type of population served (e.g., early childhood) has been established. This framework, however, is based on a three-year cycle within the original six-year district sample framework.
For campus sampling to be employed there must be at least six campuses of the same grade span within the district. This means that while there are many districts with sampling allowable at the elementary level, there are far fewer that have at least six high schools. Thus there may be a combination of campus-level sampling at the elementary level and a district-wide sample at the high school level. Again, the intent of sampling is, while maintaining an acceptable level of validity, to reduce the burden on individual campuses to the greatest extent possible.
Following the already stated purpose for sampling, a certain number of students (or student’s parents) are selected each year. In general, the number of students selected is far greater than necessary to establish acceptable levels of statistical representation at the state level. For example, the parent involvement survey includes forms for 18,000 parents statewide. Statistical bounds are, of course, somewhat lower at the regional level and the district level. In many cases, there are simply too few students to be included in reporting due to FERPA considerations.
In order to increase return rates and up the representation of information, certain groups are included at a higher than expected (based on population) ratios. For example, there are relatively few deaf-blind students so their proportion of the sample is greater than in the population in general. Likewise, prior history indicates that certain other groups have lower return rates. Increasing percentages from these eligibility categories and groups will tend to increase the validity of the sample.
Monitoring return rates
TEA will monitor the survey return rates to determine if reasonable efforts are being made by district staff to encourage returns. While it is understood that the district cannot directly force survey returns, the historic range of response rates indicates that some districts are likely making a good faith effort in this regard while others are, perhaps, not as attentive. Consideration is given to historic return rates among various groups already noted in purposive sampling.
Division of IDEA Coordination
1701 North Congress Avenue | Austin, Texas 78701-1494
Telephone: 512.463.9414 | Fax: 512.463.9560