Institute of Education Sciences
National Center for Education Statistics
Purpose: Please provide support for the focus of the study on measurement error issues specific to the age group 18-24.
Rationale: There are a few reasons for focusing on this age group. Substantively, the age range represents the period in life when most Americans who are going to obtain a high school credential, will have done so. For students who went through the U.S. elementary and secondary school system after having started kindergarten on time, and who were not held back or promoted over a grade, most will have earned a regular diploma by age 18. Age 22 represents the upper end of the age range for which most states provide free public elementary and secondary education so those who will graduate late will have had to have done so by this age, or find alternative credentialing options.1 Age 24 represents the age by which most people who pass the GED exams will have done so (71% according the most recent report on the subject released by the GED Testing Service).2 Also, this age range is a regular focus for high school completion statistics published by NCES.3 Operationally, the age range should provide enough cases from outgoing rotations of the CPS to provide meaningful statistics. Also, we think that the upper age range means that schools and school districts would need to work backward through archives that are less than 10 years old. We were concerned that requesting information in the format that we need for even older individuals could cause problems for school and school system staff retrieving information.
Procedural:
Why did NCES choose to submit this collection separately from the main CPS school supplement, thereby requesting a new OMB control number?
Rationale: There is no specific need for a separate OMB number for this study. Given that it is exploratory in nature, we were concerned that it might raise questions or require modifications such that it could not be fielded in October 2008. We did not want to couple this clearance with the annual clearance request as a result.
A1 of the SS indicates that the "Department of Education" sponsors the CPS, along with two statistical agencies. Should this say "NCES" instead?
Response: We can modify this to indicate that NCES is the cosponsor.
Please provide a brief description of the role played by the individuals cited in A8.
Response: The individuals listed in section A8 were individuals that attended meetings where the concept of the study was discussed and who had substantive interest in the study. The first meeting was the August 2006 KIDS COUNT Education Indicators Panel Meeting where high school completion and graduation rate measures were discussed in detail. Bill O’Hare organized the meeting and Brian Harris-Kojetin was one of the individuals with whom the concept of the study was discussed. At that time, the study had a much simpler design approach that was changed at the request of the Census Bureau’s legal affairs staff in light of concerns of their enabling legislation. The second was the November 2005 annual Private School Meeting sponsored by NCES where the concept of the study was presented to representatives from many private school associations. Individuals listed from the private school associations provided feedback regarding the ability and willingness of private schools to participate (Jack Klenk, Jennifer Groover, Abba Cohen, Dale McDonald, and Allen Blau were among approximately 12 association representatives at the meeting). Clyde Tucker and Tom Nardone are senior staff at BLS that agreed to the conceptual approach being taken and the purpose of the study. Matthew DeBell helped design the instrument and sampling approach. He is currently at Stanford University and is a project director on the American National Election Survey.
Confidentiality/privacy:
Please include a Title 13 confidentiality reference on the questionnaire and letter, rather than solely ESRA.
Response: We will include the Title 13 confidentiality reference. It was in an earlier draft, but removed to help reduce the potential for possibly intimidating respondents.
Please clarify how NCES or Census would "work with the GED Testing Service to verify some of the alternative credential reports." Would the testing service be privy to who is in sample for the CPS under this scenario? What strategies would mitigate against this?
Response: The Census Bureau could share information about CPS respondents with the GED Testing Service that would allow the GED Testing Service to cross walk the CPS data with the GED data. The most efficient way to do this would be to have staff at the GED Testing Service performing this work to obtain special sworn status from the Census Bureau to access the Census data allowing the GED staff to know who was in the Census data.
Has NCES experienced past difficulties in obtaining complete student directory listings when students had opted out under FERPA? Can this be quantified at all? Is there any concern that Census may have the same or greater difficulties? If so, how would this impact the validity of the findings?
Response: I do not know if we will have difficulties related to students opting out of student directory listings. Given details in the accompanying regulations, 34 CFR 99.31 – schools can disclose information without consent under several situations including when, "The disclosure is, subject to the requirements of 99.35, to authorized representatives of - The Secretary of Education."
99.35 indicates that we can get the data if they are being used to evaluate federal or state education programs. We have to avoid disclosure.
Sample Design
Please clarify who will be in sample for this collection. Specifically, B1 indicates that it will be about 3,000 18-24 year olds in the outgoing sample rotations in October 2008, yet this would seem to yield only about ½ of the sample cases planned (i.e., ½ of 3000). Are the month 4 cases also being used?
Response: Good point. Yes, the outgoing rotations include both the 4th and the 8th rotations.
Please provide additional detail regarding the power analysis alluded to in B2.
Response: We have attached a spreadsheet that works through the approach taken for deriving the power of the proposed sample.
Per previous conversations with OMB, please modify the sentence in B2 that indicates that "The statistical properties of these supplemental items will fall within those associated with the CPS itself," to ensure no misunderstanding.
Response: The sentence now reads, “The supplemental items will be administered for individuals in the outgoing sample rotations described in the attachment (rotations 4 and 8).”
Operational
Please provide approximate rates at which this age group's responses are reported by proxy.
Response: Approximately 68 percent of 18-24-year-olds in the study have proxy reported data under normal CPS data collection protocols.
Will any special procedures be employed to promote direct reporting versus proxy responses?
Response: Yes, two attempts will be made to field the supplemental items with the sampled individual before proxy reporting will be accepted.
How will the incidence of proxy reporting impact the validity of the findings?
Response: We will not know until we field the study. If reporting error rates for high school credentialing are significantly different between the two sources, we make will make this clear in the analysis and discuss its implications for how the federal government collects such data. We will also explore characteristics of the individuals for which we are seeking verification to see if proxy reporting correlates with other key characteristics of the individuals as this could affect different subgroups to varying degrees.
1 De Souza, A. (2007). ECS State Notes: School Attendance Age Limits. Education Commission of the States. Denver, CO.
2 American Council on Education, GED Testing Service. (2007). 2006 GED Program Statistical Report. Washington, DC: Author.
3 For example, please see Laird, J., Cataldi, E.F., KewalRamani, A., and Chapman, C. (2008). Dropout and Completion Rates in the United States: 2006 (NCES 2008-053). National Center for Education Statistics, Institute of Education Sciences, U.S. Department of Education. Washington, DC.
| File Type | application/msword |
| File Title | NHES:2005 Sample |
| Author | NCES |
| Last Modified By | #Administrator |
| File Modified | 2008-09-10 |
| File Created | 2008-09-10 |