6.1. May a sample be used for the baseline survey of 17 year-olds?
No. The baseline population includes every youth who is in foster care at some point during the 45 days after their 17th birthday, as per 45 CFR 1356.81(b) and 1356.82(a)(2), in a specified year. Sampling may be used only to determine the follow-up population after the State has surveyed the entire baseline population (45 CFR 1356.84).
From: NYTD Webinar Briefing - April 10, 2008
6.2. Please re-visit the explanation of how smaller States not using a random sample can determine a minimum sample size. Will these States be required to have a 100 percent sample for the baseline?
The regulation requires States to use either a simple random sample or another accepted methodology that has been approved by the Administration for Children and Families (ACF). A State's actual sample size will depend on the number of youth in the baseline population who participate in the outcomes data collection at age 17 (the sampling frame) and the application of the correct formula pursuant to 45 CFR 1356.84(c) and Appendix C. We acknowledge that, depending on the number of 17-year-olds in foster care who participate in the outcomes data collection at age 17 in the State, the resulting sample may be the entire baseline population.
From: NYTD Webinar Briefing - April 10, 2008
6.3. How do States go about getting their sampling methodology approved? What is the minimum sample required?
States should submit their sampling methodology to their Children's Bureau Regional Office specialist who will consult with the ACF statistician. The minimum sample size will depend on population of youth who participate in the outcomes data collection at age 17 and the application of the sampling requirements at 45 CFR 1356.84(c) and Appendix C.
From: NYTD Webinar Briefing - April 10, 2008
6.4. What do "Py" and "Pn" actually stand for in the calculation for determining sample size for the follow-up population?
"Py" is the estimated proportion in the population with the characteristic of interest and "Pn" is (1 - Py). If Py = 0.5, then Pn = 1 - 0.5 = 0.5. A value of 0.5 for Py maximizes the sample size and (Py)(Pn) = 0.5(0.5) = 0.25.
From: 2008 NYTD Technical Assistance Meeting - July 23-25, 2008
6.5. Why do we have to use a Z coefficient of 1.645?
For the purpose of sampling, ACF has determined that a 90 percent confidence interval provides a balance between larger and smaller sample sizes for National Youth in Transition Database (NYTD) data reporting. When a 90 percent confidence level is desired, the coefficient from a standard normal probability distribution (a Z-value) is 1.645.
From: 2008 NYTD Technical Assistance Meeting - July 23-25, 2008
6.6. Is a single sample drawn from the baseline population for the entire Federal fiscal year (FFY) or from each of the two reporting periods during the FFY?
The State is to draw the sample from the entire FFY's baseline population that has participated in the outcomes data collection (See 73 FR 10345) and not separately for each reporting period.
From: 2008 NYTD Technical Assistance Meeting - July 23-25, 2008
6.7. Is it permissible or advisable to use the formulas in Appendix C of the NYTD regulation to estimate the size of a State's follow-up population sample? Our State has a large child welfare population and anticipates opting to sample youth for follow-up NYTD outcomes data collection and wants to begin forecasting the costs associated with locating and surveying such a sample of youth in Federal Fiscal Year 2013.
States are welcome to use the formulas provided in Appendix C of the NYTD regulation for planning purposes to estimate the number of youth the State will need to locate and survey at ages 19 and 21 as part of its sample. However, the actual sample size will depend on the number of youth in the baseline population who participate in the outcomes data collection at age 17 (Q&A #6.2).
From: 2010 NYTD Technical Assistance Meeting - July 21-23, 2010
6.8. Please provide more guidance on the option to select and follow a sample of youth for follow-up outcomes data collection at ages 19 and 21. Is it accurate to say that any State that determines a sample size by using one of the formulas in Appendix C of the NYTD regulation and uses simple random sampling procedures may sample? Must States coordinate this process with ACF in any manner?
For States that opt to sample, the follow-up sample must be selected using simple random sampling procedures. States must work with the Children's Bureau Regional Office specialist, however, to approve any sampling plan where a States chooses to use a sampling methodology other than simple random sampling procedures. We will provide additional guidance on sample selection at a later date.
From: 2010 NYTD Technical Assistance Meeting - July 21-23, 2010
6.9. Our State that has opted to collect and report information on the follow-up population using sampling in accordance with 45 CFR 1356.84. What should the State report for element 36 (foster care status-outcomes) for youth in the follow-up population who are not part of the sample and were never required to be surveyed?
For States that have opted to conduct sampling in accordance with 45 CFR 1356.84, the State should report in element 36 the foster care status of a "non-sample" youth as of the first day of the report period in which the youth turned age 19. The State also should report "not in sample" as the reason for non-participation for element 34 for such youth. As previously explained in NYTD Q&A #2.38, the State is not required to report further follow-up population information specified at 45 CFR 1356.83(e) for "non-sample" youth at age 21.
From: Questions & Answers released on July 15, 2013