Transcript for 2026 Summer Training Series Session 1: Overview of NDACAN Administrative Datasets 
Presenters: Garrett Baker, Ph.D., Duke University, and Alyssa Lindsey, MPH, University of California, San Francisco
National Data Archive on Child Abuse and Neglect (NDACAN)

[MUSIC]

[VOICEOVER]
National Data Archive on Child Abuse and Neglect.

[ONSCREEN CONTENT SLIDE 1]
Welcome to the 2026 NDACAN Summer training series! The session will begin at 12pm EST. This session is being recorded. Please submit questions to the Q&A box!

[Alyssa Lindsey]
All right. Hello everyone. Welcome to our 2026 NDACAN Summer Training Series. My name is Alyssa Lindsey. I'm a graduate research associate here at NDACAN. Before I kick off our time together today, I just want to give y'all a couple of housekeeping items. We're recording this session so that we can post the recording, transcript, slides, any related materials on our website several weeks following each presentation. I'm going to announce their availability on the "child maltreatment research L" or "CMRL" list serve and I'll put that information on how to join the list serve in the chat. Most of y'all are probably already on that list serve. Please let me know if you have any questions by using the Q&A box. And speaking of Q&A, because we're recording the session, we're using a webinar format on Zoom which means that we would like y'all to ask any questions you have throughout the presentation using the Q&A box. At the bottom of the Zoom screen, you'll see a box with a question mark that says Q&A. And you can click that box to submit your questions throughout the presentation. And after the presentation's over, we'll go through those questions at the order in which they arrive. So we should have plenty of time for questions at the end of this presentation. Next slide, please.

[ONSCREEN CONTENT SLIDE 2]
NDACAN Summer Training series
National Data Archive on Child Abuse and Neglect
Duke University, Cornell University, UC San Francisco, & Mathematica

[Alyssa Lindsey]
Great. So, welcome again to the NDACAN summer training series. NDACAN is the National Data Archive On Child Abuse And Neglect housed at Duke University, Cornell University, UC San Francisco, and Mathematica. Next slide. 

[ONSCREEN CONTENT SLIDE 3]
Laying the groundwork foundational skills for using administrative data in child welfare research.
Logo for the Children's Bureau features an image of overlapping blue and white silhouettes of children next to red and white stripes. To the right is the text "Children's Bureau: An Office of the Administration for Children & Families."
Logo image which features a semicircle of icons representing people holding hands positioned on top of the acronym NDACAN, and next to the text National Data Archive Child Abuse and Neglect.

[Alyssa Lindsey]
NDACAN does two learning offerings every year. We have our monthly Office Hours Series during the academic year and in the summertime we have a Summer Training Series. And the theme of this year's 2026 Summer Training Series is "Laying The Groundwork: Foundational Skills For Using Administrative Data In Child Welfare Research". This series is designed to strengthen foundational skills necessary to conduct high-quality impactful research with administrative data. We're intending this to be useful both for folks who are new to this area and also for those who are just seeking to strengthen existing skill sets. So throughout these summer presentations, we're aiming for each session to be useful and thorough starting point for conducting research on child welfare using administrative data. 

[ONSCREEN CONTENT SLIDE 4]
NDACAN Summer Training series schedule
July 1st: Overview of NDACAN administrative datasets
July 8th: Data cleaning and management
July 15th: Linking NCANDS and AFCARS
July 22nd: Handling missing data
July 29th: Data presentation and visualization

[Alyssa Lindsey]
And this is just a preview of our schedule for the next 5 weeks. We'll be meeting every Wednesday at 12 Eastern for an hour. These are the topics that we'll cover. We'll start with an overview of NDACAN administrative data sets today and then the following sessions will cover data cleaning and management, linking NCANDS and AFCARS, handling missing data and end with a session on data presentation and visualization.

[ONSCREEN CONTENT SLIDE 5]
Session Agenda
Brief NDACAN overview
Discuss three main administrative datasets:
National Child Abuse and Neglect Data System (NCANDS)
Adoption and Foster Care Analysis and Reporting System (AFCARS)
National Youth in Transition Database (NYTD)

[Alyssa Lindsey]
So today's session is going to start with a brief NDACAN overview and then a discussion of the three main administrative data sets housed within NDACAN. The National Child Abuse and Neglect Data System also referred to as NCANDS, the Adoption And Foster Care Analysis And Reporting System or AFCARS and the National Youth And Transition Database or NYTD. 

[ONSCREEN CONTENT SLIDE 6]
Brief NDACAN OVERVIEW

[ONSCREEN CONTENT SLIDE 7]
MISSION
The National Data Archive on Child Abuse and Neglect (NDACAN) promotes secondary analysis of child abuse and neglect data by providing researchers with high quality datasets, documentation, and technical support, and encourages collaboration within the scientific community.

[Alyssa Lindsey]
So to begin, I'll briefly give an overview of NDACAN, our mission, background, and the activities we do. This is our organizational mission statement. The National Data Archive On Child Abuse And Neglect, NDACAN, promotes secondary analysis of child abuse and neglect data by providing researchers with high-quality data sets, documentation, and technical support, and encourages collaboration within the scientific community. Next slide. 

[ONSCREEN CONTENT SLIDE 8]
Background.
Located at two institutions: Duke University and Cornell University
Founded in 1988
Supported by contract with the Children’s Bureau,  Administration for Children and Families

[Alyssa Lindsey]
For some quick background as I mentioned earlier NDACAN is located at Duke University and Cornell University and founded in 1988. Our data archive is funded by the Children's Bureau in the Office Of Administration For Children And Families. Next slide. 

[ONSCREEN CONTENT SLIDE 9]
NDACAN activities.
Acquire: Collect and assess child welfare data
Protect: Protect confidentiality
Transform: Create documentation to make data usable
Disseminate: License and distribute datasets in multiple formats
Support: Support child welfare researchers
Expand: Expand the scope of child welfare research

[Alyssa Lindsey]
We participate in various activities ranging from collecting and assessing child welfare data, ensuring and protecting the confidentiality of that data, creating documentation to make the data usable, licensing and distributing data sets in multiple formats, supporting child welfare researchers, and expanding the scope of child welfare research. Now, I will pass it off to Garrett to lead the next few sections on NDACAN's three and main administrative data sets. 

[ONSCREEN CONTENT SLIDE 10]
National child abuse and neglect data system. (NCANDS)

[Garrett Baker]
All right. Thanks so much. This so I want to say quickly well first I should say who I am I'm Garrett Baker I'm formerly a graduate student now Post Doc at Duke University in the Department of Sociology Research Associate at NDACAN. I just want to say quickly before I get started talking about these different data sets I just want to re-emphasize that this is intended to be a fairly brief kind of more introductory or higher level overview of these three data sets. And also stress and I'll say this again a few times throughout the presentation that the NDACAN website has a ton of resources to learn or dive more into each of them in detail. There's recordings of past Office Hours or Training Series presentations. There's user guides, technical manuals, and just a bunch of resources that so I encourage everyone to go to the NDACAN website. And again, I'll sort of mention that a few times throughout. I guess I'd also say that for the Q&A, I'd say also probably this session's questions is probably best suited to sort of broader rather than really specific technical questions or things about your specific project. Of course, we'll try to answer whatever you have the best we can, but more specific questions there are, you know, our emails are at the end of the slides and there's also a bunch of emails for other NDACAN experts on the website. And you know, everyone is is pretty good about getting back to you if you have, you know, really specific questions  about your specific research ideas or projects.  Feel free to email us or any of the folks listed on that website. Okay. So that all said let's start by talking about NCANDS or the National Child Abuse And Neglect Data System. 

[ONSCREEN CONTENT SLIDE 11]
Overview.
A federally-sponsored national data collection effort created for the purpose of tracking the volume and nature of child maltreatment reporting each year within the United States.
Data are collected annually through voluntary participation of states.
This means users should always be wary of potential data reporting differences across states

[Garrett Baker]
So, NCANDS is a federally sponsored national data collection effort which is designed to track the frequency and the nature of child maltreatment reporting every year. Sort of main high-level thing to know about NCANDS is that even though it's often used to generate national estimates, from a user perspective I always like to stress that you know especially people starting out with NCANDS is that you should really think of it as a bunch of different state data sets and I say that for a few reasons. One is that states are not required to submit data to NCANDS. They do so voluntarily. And so their sort of funding levels, their staffing, their general sort of infrastructure surrounding data collection, data processing and then reporting really varies quite a lot by each state. And then even putting that sort of capacity aspect aside, states just differ a lot in how they sort of report or characterize specific elements of the data, specific variables etc. And so, you know, the sort of first thing I always want to make sure I emphasize about NCANDS is to and I'll say this again a few times throughout the next few slides, but I always tell people when they're getting started with NCANDS to make sure you look at state-specific trends over time if you're looking at trends over time or just breakdowns, you know, crosstabs or or whatever by state and and look for maybe potential discrepancies or ways in which states might be changing things over time. It's just really sort of best best practices. 

[ONSCREEN CONTENT SLIDE 12]
DATA DETAILS
Child File: Case-level records
Agency File: State-level aggregates

[Garrett Baker]
So there's two main NCANDS files the Child File and the Agency File. We'll talk about both of those here. The Child File contains case-level records. I've got a lot more to say about unit of analyses in a moment but basically it gives much more detailed information about specific children, specific reports, maltreatment allegations, disposition services, related constructs. The Agency File on the other hand we'll cover much more briefly because it's a lot simpler. It's basically just state-level aggregates and we'll talk about that at the end. 

[ONSCREEN CONTENT SLIDE 13]
Child file.
Contains many reports, where each report:
Is a notification of suspected child abuse
May involve one or more children (records)
May be substantiated, unsubstantiated, alternative response, or other dispositions
Has information on the alleged maltreatment, the alleged perpetrator(s), child and caregiver risk factors, and services received

[Garrett Baker]
So the Child File contains records connected to reports of suspected reported child abuse or neglect. A report could involve one child or multiple children. It could be it's just a report, right? So it could be unsubstantiated, it could be substantiated, there could be some alternative response or disposition, but you know, it's just thought of as a report. The file also includes information on details about the maltreatment types, the alleged perpetrators, risk factors, services received, etc. So it's a quite rich it's quite rich data set but it should be you know always remembered by users that it only contains information reported to and then collected by the agency. 

[ONSCREEN CONTENT SLIDE 14]
CHILD FILE.
Consist of all investigations or assessments of alleged child maltreatment that received a disposition in the reporting year
Records are provided at the level of each child on a report, also known as the report-child pair
This means a given child could show up multiple times if they are the subject of multiple reports.

[Garrett Baker]
So something that we'll show a sort of visual or schematic for in a couple slides is to remember that it's the Child File is not necessarily a child-level data set and by that I mean it's not like each child just has one single row, right? It consists of all investigations or assessments of alleged reported malreatment. And so records are really best thought of at the level of each child on a report or a report-child pair. And this means that a given child could show up multiple times in a given year if they're the subject of multiple reports. And I promise I'll keep stressing this in the next couple slides and give you some visuals if you're a visual learner. 

[ONSCREEN CONTENT SLIDE 15]
VARIABLES.
Report variables include:
Report Data (report ID, report date, disposition, investigation start date)
Child Data (child ID, demographics)
Maltreatment Data (types, individual dispositions, fatalities)
Child Risk Factors (substance abuse, diagnosed disabilities)

[Garrett Baker]
So the Child File also includes several broad categories of variables. We have things like report you know the an id an identifier for the specific report. We have the date, disposition, the start date for when the investigation began. We have specific information about the children. So there's child-specific ids which will come in handy. And again we'll show a visual in a moment. We have sort of basic demographics about the child. Then we have specific information about the malreatment. So the actual type, physical abuse, some type of neglect includes fatality information. Then we also have some some child level kind of risk factors if there's any diagnosed disabilities, substance related things. Those are also typically available as well. 

[ONSCREEN CONTENT SLIDE 16]
VARIABLES.
Child variables include:
Caregiver Risk Factors (substance abuse, financial problems)
Services (foster care, adoption, counseling)
Perpetrator Data (relationship to victim, demographics, sex trafficking)
Additional Fields (AFCARS ID, date of death, plan of safe care, referral to CARA-related services)

[Garrett Baker]
We also have information on caregivers. So caregiver risk factors potential financial issues or substance abuse services that have been engaged with foster care, adoption, some type of counseling. We also have information about the perpetrators. So what is their relationship to the victim and then some demographics about the perpetrator as well. And then there's a bit of sort of additional potential fields as well. 

[ONSCREEN CONTENT SLIDE 17]
Child File Entity Relationships.
Report-to-Child Relationship Diagram: Report A is connected to Child 1 and Child 2.
Child-to-Report Relationship Diagram: Child 1 is connected to Report A and Report B.
A child can be the subject of more than one report in the reporting period.

[Garrett Baker]
Okay. So this is what I was mentioning about sort of how to think about the structure of the data. And I'll say at the onset that especially if you're new to NCANDS, I think the best thing to do is always be really really clear at the start about what your goal is, what your research question is or what you're sort of trying to investigate. Are you trying to investigate things about reports? Are you trying to investigate something about children? Are you trying to investigate something about the perpetrators? Right? Because all of those things can exist sort of on different levels. And so having a sense of your unit of analysis at the start will make this part I think a little simpler. But so this shows just one example of how the Child File structure can get complicated and should be thought about carefully. So on the far left here you have you know one report that could involve multiple children, right? So you could have two children included in one report. On the flip side, you could have one child who is the subject of multiple reports, right? So we're not just again working with a single child per row file. So this is, you know, this is useful if you're thinking about potentially looking at repeated reports or report level patterns. But you know just again make sure you're always matching sort of your question or what you're investigating with the actual unit of analysis or structure of the data.

[ONSCREEN CONTENT SLIDE 18]
Child File Entity Relationships: Child 1 is connected to Maltreatments 1-4, Perpetrator A connected to Maltreatments 1-3, and Perpetrator B connected to Maltreatments 1-2.
For a child in a report, each perpetrator can be associated with each instance of reported maltreatment.

[Garrett Baker]
Here's another way in which the sort of unit of analysis or structure of the data can get more detailed. For each child in a report each perpetrator can be associated with multiple particular malreatment allegations. Right? So a single child report record may involve multiple maltreatment types and multiple alleged perpetrators. Right? So someone who is reporting to CPS they could say that you know this report is about a single child but they are suspecting that this child has been the subject of multiple forms of malreatment from multiple perpetrators. So you can get this sort of hierarchical or sort of multi-layered structure in a single report pretty quickly.  And so again, I'll just say, you know, this is is powerful from a sort of analysis and detail standpoint, but again, you just want to make sure you're you're clear on sort of what you're studying. 

[ONSCREEN CONTENT SLIDE 19]
Child File Entity Relationships: Perpetrator A is connected to Reports A and B. Report A is connected to Child 1 and Child 2. Report B is connected to Child 2 and Child 3.

[Garrett Baker]
Then finally, you know, another layer to be aware of is that a perpetrator, if you're sort of thinking about it from the perpetrator side, a perpetrator can be associated with more than one report and more than one child in a report. So you see here on the left you know this perpetrator it could be the subject or the subject as a perpetrator of multiple reports and there could be multiple children within those reports. And so just with one perpetrator you could have multiple reports and multiple children nested within those reports. So the broad takeaway here is that NCANDS has a lot of really rich and detailed the Child File I should say has a lot of really rich and detailed information on the malreatment the child and the perpetrator and that you can get really detailed data at multiple levels of all of those different aspects. But you just need to always make sure you're not just treating this as a simple child level sort of flat data set without first understanding both your own goals and then what you're actually studying as a unit of analysis when you're working with the data.

[ONSCREEN CONTENT SLIDE 20]
AGENCY FILE.
The unit of observation here is the state
So N=52 observations
Child Abuse Prevention and Treatment Act (CAPTA) required items

[Garrett Baker]
Okay. The Agency File is much much simpler. Basically it shifts from these sort of case-level child report case-level records to just state-level aggregate data. So the unit of analysis or the unit of observation is the state. So you have 50 states plus DC and Puerto Rico. So you end up with 52 observations total. This includes sort of useful information for understanding kind of state-level patterns, counts, summary measures. But it doesn't necessarily give you the same sort of answers or level of detail that you get from the Child File. This is really if you're interested in sort of state-level patterns. 

[ONSCREEN CONTENT SLIDE 21]
AGENCY FILE.
Summary data
Prevention Services
Referrals and Reports
Additional information on child victims reported in the Child File
Part C of IDEA reporting
Child Fatalities

[Garrett Baker]
So some of the information you're going to get in the agency file includes information on prevention services, referrals and reports, additional data information on victims that are reported on which would be part of the Child File, child fatalities part C of IDEA reporting. So this is you know again a sort of higher level kind of overview at the state-level. 

[ONSCREEN CONTENT SLIDE 22]
Child Maltreatment report.
Contains state-level summary statistics are published annually by the Children’s Bureau in the Child Maltreatment Report
Link to 2024 report:
https://acf.gov/cb/report/child-maltreatment-2024
Image Description: Cover of the “Child Maltreatment 2024” report, marking the 35th year of national data collection. The background features silhouettes of adults and children standing on a multicolored puzzle floor, symbolizing the complexity and interconnectedness of child welfare. A circular emblem in the top right reads “35th Year of Reporting.” The bottom includes logos and text for the U.S. Department of Health & Human Services, Administration for Children and Families, Administration on Children, Youth and Families, and the Children's Bureau.

[Garrett Baker]
One thing to plug here is that the Children's Bureau actually puts out a report which has these sort of state-level summary statistics. These are published annually the 2024 report we link to here very easy to find on the NDACAN website as well. It's also I included it here because I think it's also useful especially if you're just getting started with the data to use as sort of a check against your own analyses or your own sort of initial understanding of the data. If you're for example working with the Agency File and you're coming up with numbers that are far different from what you're seeing here then you know you suggest that you may want to go double check the codebook or your code or just sort of how you're looking at or or analyzing the data.

[ONSCREEN CONTENT SLIDE 23]
Adoption and foster care Analysis and Reporting System (AFCARS)

[Garrett Baker]
Okay. So next we'll move on to the adoption and foster care analysis and reporting system more commonly referred to as AFCARS. AFCARS is the sort of main national administrative data source on foster care and adoption as the name suggests. So just to make explicit whereas NCANDS focuses on reported child maltreatment, AFCARS is directly focused on children's experiences in foster care, out-of-home care, and adoption. 

[ONSCREEN CONTENT SLIDE 24]
OVERVIEW.
Federally mandated data on all children covered by the protections of Title IV-B/E of the Social Security Act (Section 427)
States are required to collect data on:
All adopted children placed by the state’s child welfare agency or by private agencies under contract with the public child welfare agency
All children in foster care for whom the state child welfare agency has responsibility for placement, care, or supervision

[Garrett Baker]
So AFCARS is federally mandated, it covers children connected to Title 4 B and E child welfare programs as part of the Social Security Act. States are required. So again, as a sort of difference with NCANDS where NCANDS data are often a little bit messier and and sort of more varied at the state level because there's not that sort of required there's not that required reporting mechanism. It's different with AFCARS where AFCARS has a requirement for states to collect and report data on two things. One, all adopted children placed by the state's child welfare agency or by private agencies who are under contract with the child welfare agency and all children in foster care for whom the state child welfare agency has responsibility for their placement their care their supervision. 

[ONSCREEN CONTENT SLIDE 25]
DATA DETAILS.
Foster Care (FC) data files contain information at the child-level
Contains information on:
Child demographics (gender, birth date, race and ethnicity)
Their experience in foster care
Removal reason, the number of previous stays in FC, service goals, availability for adoption, dates of removal and discharge, funding sources
Demographic information about the child’s caregiver

[Garrett Baker]
Foster care data files FC files as they're sometimes called contain information at the child level. So again a sort of difference from NCANDS. Here we are talking about child-level data. And these child-level data contain information on demographics, things like gender, race, ethnicity, when a child's born, their experience in foster care, and this has a whole wealth of data. Some include removal reasons, number of previous stays if any in foster care, service goals, availability for adoption, dates for removal and discharge, potential funding services, and many others. And there's also some demographic information available about the child's caregiver. One key point I always want to stress whenever I talk about dates is that you know users should always be really careful in how they're defining entries, exits, sort of time can be a bit tricky if it's something you're not used to working with. 

[ONSCREEN CONTENT SLIDE 26]
DATA DETAILS.
Adoption data files contain information on
Demographics of the adopted child (gender, race, birth date, ethnicity)
Prior relationship with the adoptive parents
The date the adoption was finalized
And dates parental rights were terminated
Characteristics of birth and adoptive parents
Whether the child was placed from within or outside of the US

[Garrett Baker]
Then the adoption data files contain information on the demographics of the adopted child. So again, gender, race, ethnicity, when child's born, birth date, prior relationship with the adoptive parents, if any, the date that the adoption was actually formally finalized, as well as dates that parental rights were terminated, characteristics of both the birth parents and the adoptive parents. And then it also contains international information or really just whether the child was placed from within or outside of the U.S.. 

[ONSCREEN CONTENT SLIDE 27]
Recent changes.
This image shows a timeline of the AFCARS transition from old to new format beginning FY2023.

[Garrett Baker]
So something that is really useful to know especially if you are a new user of AFCARS is that there have recently in the last few years been some fairly big changes to how AFCARS is structured. You know this doesn't change you shouldn't think of this as sort of a holistic or you know that you can't possibly use old versions of AFCARS at all. But there's a few sort of key things that that users should know about. I will note as well that we have done some past presentations exclusively focusing on the AFCARS changes. And this will you know like the rest of this presentation be a pretty brief and high level overview that just gives you some initial information and then you know if this is something that you think is going to be relevant for you or that you don't or that you want to learn more about I'd recommend you go to the website there's again plenty of user guides and manuals but also some recordings of past presentations that we've done on the AFCARS changes. So an important shift is that there's a shift from the older files which were typically referred to as annual files to newer files that are referred to as the AB files. Basically, states submit AFCARS data and sort of two periods in the fiscal year, two sort of six-month periods. The older annual files used the A and B periods for that fiscal year plus they waited for the following year's initial period before sort of synthesizing and reporting out the data which helped capture sort of late submissions, corrections, cases that had not been fully resolved by the end of the fiscal year. But that also delayed you know typically delayed reporting a bit. So now the newer AB files only use the A and B periods for the nominal fiscal year. The new files then you know they're not conceptually different. They don't you know totally change things but there could be some sort of aggregate counts that are sort of systematically different right before and right after because of this sort of shorter reporting window. 

[ONSCREEN CONTENT SLIDE 28]
Recent changes.
States report all removals, permanency plans, and placements
From 66 elements to 186
Foster care and Adoption files are now combined into the Out of Home Care file
Some variable definitions changed

[Garrett Baker]
There's a few other sort of important changes to know. One is that the data elements or you could think of this kind of as variables expanded dramatically. The old file AFCARS file system has 66 elements whereas the new system now has 186. So you get a ton more sort of detailed information. We sort of show some in past presentations we show some examples of what this kind of means that there's now a lot more detailed variables and definitions as this final bullet point shows some variable definitions change and so you know you can have some cases there are some that are totally the same and the again the sort of meaning or the underlying foundation of the data are not conceptually different but you might have you know a variable that had a few only, you know, say two or three different response items previously and now has eight different ones or or maybe they changed or combined or or split up you know, one into multiple variables. And so you should just anytime you're working over time with AFCARS data, you should just make sure that you know sort of when these years overlap when this shift happened and make sure you're plotting things over time and looking at you know is there a jump and and maybe investigating that more to see if that's a sort of real jump that happened in the real world or is that sort of an artifact of this sort of administrative process or procedural change. The other thing to now know is that foster care and adoption files are now combined into the out of home care file.

[ONSCREEN CONTENT SLIDE 29]
Recent changes.
Best practices:
Pay particular attention to years
Plot variables longitudinally (by year) and look for any jumps or discontinuities
Always check missingness and cross-tabs (and by state!)
Read codebook and variable definitions carefully

[Garrett Baker]
Okay. So again best practices for AFCARS particularly if you are using these data across years if you're only looking at just you know this most recent year and that's all you care about then this will not be as relevant for you. But if you are trying to look at longitudinal trends or compare years make sure you plot the key variables longitudinally look for any jumps or discontinuities and then go investigate those more rigorously. Always check missingness and crosstabs again by state you know I will say always with with any NDACAN data that you know you should think of it as state-level data in a sense and make sure you're looking to see if there's any sort of systematic differences across states and across time with AFCARS in particular and then make sure you read the codebooks and variable definitions carefully like I said that these these did update quite a bit and so make sure you know sort of the underlying structure of the variables that you care about. 

[ONSCREEN CONTENT SLIDE 30]
National Youth in Transition Database.

[Garrett Baker]
Okay. And then the third and final administrative data set that we'll discuss is the National Youth In Transition Database. This is often referred to as NYTD. NYTD is quite different than NCANDS and AFCARS. It focuses specifically on older youth as youth are transitioning from foster care towards adulthood and it's survey-based. So whereas NCANDS is centered on malreatment reports and is you know agency administrative state based or county or state agency-based data as well as AFCARS which is specifically in foster care. NYTD is more of a survey-based data set that I think of as sort of asking you know questions of youth about their experiences and themselves as they're transitioning into adulthood out of foster care. 

[ONSCREEN CONTENT SLIDE 31]
Overview.
Data collection is mandated by the John H. Chafee Foster Care Program for Successful Transition to Adulthood (Section 477 of Social Security Act, as amended by P.L. I 15-123).
Unique compared to NCANDS and AFCARS because it is derived from survey-based data
Goal is to help improve our understanding of understand how (and how well) states are preparing youth in the child welfare system for adulthood

[Garrett Baker]
So NYTD data collection is mandated by the John H. Chafee Foster Care Program for Successful Transition to Adulthood Section 477 of the Social Security Act. It's again unique because it's derived from survey-based data and the goal is really to help improve our understanding of how and how well states are preparing youth in the child welfare system for adulthood. 

[ONSCREEN CONTENT SLIDE 32]
Data details.
Data are at the case-level
Two files:
Services (updated annually, data on all children who received services)
Outcomes (survey-based)

[Garrett Baker]
So these data are at the case level but there's two files and something to keep in mind. So there's a Services file which I will say is sort of the I think less used data set. Typically people who are interested in it are using the Outcomes file which is the survey based data that I was just talking about. The Services File is updated annually and it basically just contains sort of basic information on all children who received services. Again the Outcomes File is this sort of survey-based accelerated or staggered cohort designbased survey that I'll talk a bit more about in a second. 

[ONSCREEN CONTENT SLIDE 33]
Data details.
Services
Academic supports, career preparation, budgeting, mentoring, health education, housing education, financial assistance, etc.
Outcomes
Financial self-sufficiency, educational attainment homelessness, high-risk behaviors, connection with adults, access to health insurance, incarceration, etc.

[Garrett Baker]
So again, services is talking or is containing information on academic supports, career prep, budgeting, mentoring, health, education, housing, education, financial assistance, those kind of things. Outcomes data, you're getting things on financial self-sufficiency, educational attainment and homelessness, high-risk behaviors, connection with adults and social systems, access to health insurance experiences with the juvenile or criminal justice system. And so NYTD can be useful for asking whether youth who are receiving the Outcomes data I should say of NYTD can be useful for asking whether youth are receiving these transition-related supports and for describing how these folks are fairing across multiple dimensions of sort of late adolescence or early adulthood.

[ONSCREEN CONTENT SLIDE 34]
Data collection.
States began reporting in October 2010
Cohort-based for Outcomes file
Initial wave for what’s referred to as the “baseline population”
All youth in foster care who reach their 17th birthday in FY2011 and then every third fiscal year following
They are then followed up with at age 19 and 21

[Garrett Baker]
So states began reporting information for NYTD in October 2010. It's also a fairly new fairly newer source of data. Like I said or at least mentioned a couple slides ago, the outcomes file, the survey is cohort-based with a bit of a complex structure that I'll show you a table for in a moment. There's a sort of initial wave for what's referred to as the baseline population. This is all youth who are in foster care who reach their 17th birthday in the given fiscal year. So it started in fiscal year 2011. And then is followed up with at ages 19 and 21. 

[ONSCREEN CONTENT SLIDE 35]
Data collection.
Table titled “NYTD Data Collection Schedule, FFYs 2021–2030” showing when different youth data are collected across fiscal years FY21 through FY30. It lists four reporting populations: information on youth receiving services, youth outcomes at age 17 for youth in foster care, youth outcomes at age 19, and youth outcomes at age 21. Check marks indicate annual collection for youth receiving services, while the other rows show cohort-based collection in selected years across the decade.

[Garrett Baker]
Again for our visual learners, I think this is a useful a useful sort of schematic though can be a little confusing at first. This is a more recent so even projecting into the future for cohorts. You can see you know if you look at cohort five that started as 17-year-olds in fiscal year 2023 you can see that basically two years later they're followed up with at age 19 in fiscal year 2025 and then they will be followed up with again at age 21 in fiscal year 2027. So you can do that again, you know, with the current baseline population of 17-year-olds or six, that's this current fiscal year 2026, they will be followed up with two years from now. In 2028 and at age 19, and then again at age 21 if you just sort of look through the diagonals going top left to bottom right there, you can track sort of one cohort as they age across time. And then you see in that top row the information on youth receiving services the Services File is is collected every year. So as with you know any data set but especially with NYTD I think you know users need to be especially attentive to the sort of baseline population and cohort definitions. You know are you trying to track one cohort in particular over time? Do you only care about people, you know, who were 17 in a given fiscal year and you're just trying to track them over time? Do you care about looking at changes in trajectories over time? Sort of longitudinal modeling, in which case you want to be able to compare multiple cohorts, but then you know, again, just making sure that your your sort of question or your sort of research or data goals are well aligned with your actual data structure and what you're sort of doing with it. 

[ONSCREEN CONTENT SLIDE 36]
Research examples.
canDL is a database of references related to NDACAN datasets and to secondary research. It is maintained as a publicly accessible Zotero database.
Link: https://www.ndacan.acf.hhs.gov/candl/candl.cfm

[Garrett Baker]
Okay. And then the to wrap up here, I want to re-emphasize that the NDACAN website has a ton of resources you know, user manuals, code books, past slides, presentations, recordings that will cover each of these data that do cover each of these datasets in much more respective detail than we were able to get into during our time today as as today was sort of meant to be a high-level overview. And then in particular I want to draw attention to what's referred to as canDL c a n d l which is a database of basically research or references related to NDACAN datasets. It's maintained as a publicly-accessible Zotero database. The link we have in here but it's also very easy to find if you just search if you just search online. And it's really useful. I always recommend to people who are just getting started with new data that they go read actual research that's been done with these data. I find that that's sort of the best most intuitive way to kind of figure out what you can do with the data is to look at how others have used it and how others have described it. You know, most published research studies will give a pretty good detailed holistic overview of the data they use. And I think that that's often a really kind of simple intuitive summary of what's contained in the data and how you can use it. So I encourage everyone to go look at the canDL repository. 

[ONSCREEN CONTENT SLIDE 37]
Questions?
User support: NDACANsupport@cornell.edu
Garrett baker: garrett.baker@duke.edu
Alyssa Lindsey: Alyssa.lindsey@ucsf.edu

[Garrett Baker]
Okay. So we've hit time for questions. My email and Alyssa's are both here. There's the NDACAN support email and again on the NDACAN website there's a bunch of folks who are not mentioned here that are NDACAN associates, experts who are also available to answer questions. 

[ONSCREEN CONTENT SLIDE 38]
Next week…
Date: July 8
Topic: Data cleaning and management
Instructor: Alexander Roehrkasse

[Garrett Baker]
Also want to mention next week's topic will be data cleaning and management and alex warassie will be leading that presentation. So you want to get us into some questions? 

[Alyssa Lindsey]
Yeah, it looks like we have one question. Does NYTD include follow-up data on where individuals are living after exiting foster care such as living with family, living alone, living at college, etc.?

[Garrett Baker]
There is some some data indeed on living arrangements in NYTD at the at the follow-up studies at age 19 and 21. Yes, you do get some information on on the sort of living arrangements of people. So, yes, that is available. 

[Alyssa Lindsey]
Great. And that looks like the only question so far, but please feel free to put questions in the chat. I've also in the or in the Q&A and in the chat, I've put several resources throughout the presentation while Garrett was speaking about where to find various information that he shared throughout.

[COMMENTS IN THE ZOOM CHAT]
The NDACAN website is https://www.ndacan.acf.hhs.gov and contains the following:
The NDACAN archive recorded a special version of its Office Hours on using 'R' for data analysis. It is a series called 'LeaRn with NDACAN’ and is available on the NDACAN website.
More information on the AFCARS changes can be found in the NDACAN website FAQ page.
The Child Abuse and Neglect Digital Library (canDL) is a list of publications created using NDACAN datasets and is available on-line here: https://www.zotero.org/groups/421939/candl/library

[Alyssa Lindsey]
Great. We have a question. Is the child specific ID consistent for the same child across multiple reports? 

[Garrett Baker]
Correct. Yes, that is one of the goals of that child ID is that you will be you know if a child is a subject of multiple reports in a given year, you can ascertain that by tracking them through their their child-specific ID. Good question. 

[Alyssa Lindsey]
Great. And then we have a question. How to get raw data set? Is there any eligibility? 

[Garrett Baker]
Yes. So you can just go to the NDACAN website. There is not really any restriction or eligibility. If you go to the website for the particular you know NCANDS or AFCARS or NYTD it will have more information on how to get it.

[Alyssa Lindsey]
Great. A couple more questions. Are code books and actual variables with descriptions available to browse publicly?

[Garrett Baker]
Uh, yes, they are indeed. NDACAN website will have really detailed information. Will have really detailed information. Yeah, 

[Alyssa Lindsey]
Great. And then this question says, I know you mentioned there are presentations/resources available about the transitions to AB files, but could you give a quick overview on how this will affect entry exit counts? Do we expect for there to be a systematic change or will this depend on how states report? 

[Garrett Baker]
Yeah, great question. So the the short answer well you know I will say we have a bunch of actual examples in some of those presentations that I mentioned. So I will encourage you to go look at those because we actually show some trends over time in particular variables and and counts and by state. So there so to give again a quick answer there is some systematic changes but it varies by state. And so I think it yeah I think you you really just have to be able to look you have to really just look at the state-specific trends. But the short answer is that yes there are some systematic changes as well as to variables too because the sort of elements change. We pull out a couple examples that show you know not massive but there is a little bit of a sort of discontinuity or jump in either direction. So yes the short answer is there are some systematic changes but it depends. 

[Alyssa Lindsey]
Okay. Then we have a question do any of the data sets have data on the state that youth reside in given child welfare policies for states differ so much? 

[Garrett Baker]
Yes the data sets do indeed have state-level identifiers. 

[Alyssa Lindsey]
Great. And then this question says for AFCARS data you just submit a request for NCANDS data you have to submit a request and need to get approval for NYTD is it more like AFCARS or NCANDS? 

[Garrett Baker]
I might kick that to Andres. I believe it's more like I believe it's more like AFCARS but I'll let Andres give a better answer than I will to that. 

[Andres Arroyo]
Yes, it is like AFCARS a non-restricted data set. 

[Garrett Baker]
Yeah, thanks Andres.

[Alyssa Lindsey]
Great questions, everyone. Feel free to put some more in the chat or in the Q&A. Great. Another couple questions for the data sets with state identifiers. Would those be in the restricted versions?  

[Garrett Baker]
I'm again going to maybe kick this to Andres, but I believe the state identifiers are available for all. I'm not I don't I can't think of a data set that would not have state identifiers. Andres, is that right? 

[Andres Arroyo]
I'm not too familiar with the contents of the data sets. Yeah, I think they all have state identifiers, though. 

[Garrett Baker]
Yeah. Yeah. I think would those yeah, I believe I don't I'm not aware of any restrictions in getting state identifiers, I guess, is the the shortest answer to that question. 

[Alyssa Lindsey]
Great. And then are the mentioned presentations that provide examples of discontinuities in the AB transition available online? 

[Garrett Baker]
I believe they are. The most recent one we did was an office hours over the winter and so I'm I'm almost positive that's made its way online already. 

[Alyssa Lindsey]
I'll do some digging and see if I can put that link in the chat as well. Thanks for that question. Great. Thanks. Did you talk about how AFCARS and NCANDS can be combined? 

[Garrett Baker]
I guess it depends a little bit about what you mean by combined. They can be combined in the sense that you can look at child maltreatment reports and you can look at reports and of malreatment and foster care placements in a given year and even in a given state. I guess it depends a little on what you mean by combined.

[Alyssa Lindsey]
Okay. And then a question saying I saw NIS on your website as a data set. Is that still being collected? 

[Garrett Baker]
That I do not know. I do not. The national that's the national study of incidence of child abuse and neglect. Andres, anything to know about that?

[Andres Arroyo]
Oh, I'm not sure if that data collection effort is still ongoing. 

[Garrett Baker]
Yeah, I don't believe it is. I'm not involved with it. If it is. 

[Alyssa Lindsey]
And then I think this question is being addressed already, but this question says, "Are there additional restrictions for media use of the AFCARS files outside of the terms and conditions", which I believe your answer should be in text form. Great. Yeah. So the question around additional restrictions for media use of the AFCARS files outside of the terms and conditions the answer to that is the Terms Of Use Agreement which is a PDF which governs the use of the data applied to all researchers and are not different from media's use of the data. So if that answers your question. And then regarding the question around previous events, I'm just going to put in the chat our events page where previous event recordings can be found, including office hours. 

[Andres Arroyo]
There is a correction. We don't record the Office Hours. So, Office Hours materials can be requested from the NDACAN data archive and the presenters for those Office Hours can share those materials directly with people who request it. But in general we don't record Office Hours but we do record Summer Training Series and some other events.

[Alyssa Lindsey]
Amazing. Thank you, Andres. Great. Thanks everyone for all your questions. If there are any additional ones or specific ones that you'd like to reach out to us about, I'll put the user support email in the chat, Garrett's email, and my email as well. And then Andres just put in the chat our archive of one special Office Hours that you can look at as well. But it looks like we've reached the end of the questions that have been coming in. I think thank you all for coming in again. We have our next session next week July 8th over data cleaning and management. But thank you everyone.

[VOICEOVER]
The National Data Archive on Child Abuse and Neglect is a joint project of Duke University, Cornell University, University of California, San Francisco, and Mathematica. Funding for NDACAN is provided by the Children's Bureau, an office of the Administration for Children and Families.

[Music]