[Voiceover] National Data Archive on Child Abuse and Neglect
[Erin McCauley] Alright it is noon so I’d figure we’d get started. Good afternoon everyone and thank you for tuning in for our second session in the second annual summer training series from the National Data Archive on Child Abuse and Neglect. We’re really happy to have you. And we have a really exciting session today we have Madonna who’s going to be introducing us and then Kristen and Michael are going to be leading us through the session but more on them later I just want to start off by welcoming you and thinking our presenters for being here today. So as I said this is the second summer training series that we’re really excited about it it’s kind of another way were hoping to help connect our audiences and then get more information about our data out there. Those of us who joined last year know that that session was focused just specifically on the NYTD data set and then at the end of the year we sent out a survey asking for feedback and we had a lot of requests to do the whole administrative cluster and so that’s the focus of this year’s training series and today were going to be going in depth from the first one. The first data set in that cluster. And as I said this is the summer training series from the National Data Archive on Child Abuse and Neglect. we’re housed in the Bronfenbrenner Center for Translational Research at Cornell University. I’m the host of the series my name is Erin and you’ll be hearing from me each week just kind of getting us rolling and closing us out. And I’m a graduate researcher in the archive. So here’s our summer series overview. Last week we heard from John and he gave us an introduction to NDACAN and then today were going to be going in depth on our first set of data and then next week we’ll be going over the latter two and then throughout the summer then we’ll be learning some more concrete strategies on how to use the data and then how to link them and then concluding with a research example. And so now I’m gonna pass over to Madonna from the WRMA who we’re really lucky to have here today. And she’s going to be introducing our session. 
 [Madonna Aveni] thank you Erin. Okay so as Erin said my name is Madonna Aveni and I am the project manager for the National Child Abuse and Neglect Data System. Today I’m presenting the the introductions to NCANDS on behalf of Elaine Stedt who is the director of the Office of Child Abuse and Neglect at the Children’s Bureau and Elaine is also the NCANDS federal project officer or COR. So I am doing this introduction on her behalf. She sends her apologies for not being available today. So just to give you a brief background on what NCANDS is, it was created as a voluntary system. That voluntary part is very important to NCANDS, and it was created in response to a 1988 Child Abuse and Neglect Treatment Act or CAPTA Amendment. NCANDS data are a critical resource for many publications, reports, and activities of the federal government child welfare personnel, researchers, and many others. We collect two submissions from each state case level data are collected for all children who received a response from a CPS agency and aggregate data are collected from mostly agencies outside of CPS, such as fatality data from vital statistics or law enforcement agencies. Data are due each year at the end of January. The data collection period is the federal fiscal year. All 50 states, the Commonwealth of Puerto Rico and the District of Columbia submit the case level data in a file known as the Child File, aggregate data in the Agency File, and an encryption certificate certifying that all of their data are encrypted. Once a state uploads their data, a technical review occurs and the data are validated. This usually takes a few months between January and March and one NCANDS technical team member works with each state to ensure that their data are validated and reviewed prior to acceptance. Every year the NCANDS data are publicly released in the annual Child Maltreatment report. As you can see we have a picture here from the 28th edition of Child Maltreatment so this has been going on since the beginning. So when we collect the data and validate them that occurs from January to March, we analyze the data in early spring, then we draft text and download population updates from the U.S. Census Bureau and that occurs during the summer. And then the the report is laid out by a graphic designer and goes through Federal clearance in the fall, and then the report is released in winter, usually about January or February. So and the reason that I’m telling you all of this is so that you can understand that we collect, validate, analyze, write the text for the report, and release the report all within 12 months. We collect the data in January, the report including those data are released in the following January. It’s one of the faster federal reports that exists. It’s also a very important report as I’m sure many of you will agree. All the reports going back to 1995 are available on the Children’s Bureau website at the link at the bottom of this slide. (https://www.acf.hhs.gov/cb/research-data-technology/statistics-research/child-maltreatment). So now I’m going to pass it over to Kristen Stafford who is a senior research manager at WRMA, and the task lead for two important tasks on the NCANDS project. Thank you. 
[Kristen Stafford] okay hello everyone as Madonna mentioned I’m Kristen Stafford I’ve been working on the NCANDS project as senior research manager for several years. I am going to give a little bit more detailed information about the NCANDS. So we’re going to start with the Child File which as we previously mentioned is the case-level data. It’s currently has is an ASCII text file containing child-specific records for each report that received a CPS response. So these are every response that is screened in by a state and completed report. The reports must be completed during the reporting year so they result in a disposition or finding. A report-child pair is what we call on NCANDS record. This combines a report ID and a child ID to uniquely identify that single record. When a completed report results in a disposition, we end up with victims and non-victims within that report, essentially. A child with at least one maltreatment level of substantiated or indicated is considered a victim. And any other children any other children with any other maltreatment level are considered non-victims. In the past we included ‘alternative response victim’ as a victim in the Child File, and that is no longer the case in keeping with the philosophy of alternative response. All alternative response children are considered non-victim. So report characteristics, and again a report is can contain it’s a notification of suspected child abuse and it can contain more than one child so multiple records within that report. Each one of those children may be substantiated, unsubstantiated, or other disposition, as well as the report itself has a disposition. A report may also involve multiple perpetrators as well as multiple children. A single Child File record, as separate from the report, this is data related to only one child in a given report and it represents the victim or nonvictim so it represents a disposition for that one child. There are data in all fields for victim records, including perpetrators, and that includes up to three perpetrators per victim. It also includes demographic information, services information if available, risk factor information if available. So the entities, what we consider the entities in the Child File, these include report data, so data there are linked to the report, this includes an ID for each report, a date a report date for each report, the overall disposition, and the investigation start date and time. Child data includes things like the ID for each individual child, and like I said various demographics including race, sex, and age. There’s maltreatment data, so for each child the types, individual dispositions for each type of which is up to four, and whether this particular record was a fatality. We also include child risk factors such as substance abuse and/or diagnosed disability. There are caregiver risk factors, services that are provided, perpetrator data including demographics and the relationship to the victim. And we also have a category called additional fields and this includes things like the AFCARS ID which can link NCANDS Child File to the AFCARS file, the date of death if the record is a fatality, plan of safe care, and referral to what we have called CARA-related services that these are appropriate services for infants with prenatal substance exposure as well as the plan of safe care that is also linked to these infants with prenatal substance exposure. This is in response to legislation that was passed in 2016 to capture data related to infants with prenatal substance exposure. So those are actually the plan of safe care and the referral to appropriate services, those are actually two new fields that were added to the Child File just last year, so FFY 2018 is the first year of data collection on those two new fields. So to do a sort of a graphical representation of Child File entity relationships, you can see on this screen that we have Report A. Report A has two children in it, Child 1 and 2, and then Child 1 was also in Report B so this is two ways of looking at this. A report can pertain to more than one child but a child can also be the subject of more than one report during that federal fiscal year reporting period. We also have the relationship between children, maltreatment, and their perpetrators. For a child in a report each perpetrator can be associated with each maltreatment so in this example, Child 1 had four maltreatment types. Each one of those maltreatment types will have an associated level or disposition and each one of those maltreatment types should be linked to a perpetrator if the disposition is substantiated or indicated. So Perpetrator A is linked to Maltreatment 1, 2, and 3. Perpetrator B is linked to just Maltreatment 3 and 4 so you’ll see that Perpetrator A and B are both linked to Maltreatment 3. Looking at it from the other point of view, from the side of the perpetrator point of view, any given perpetrator can be associated with more than one report, and one child in a report. So in this example Perpetrator A is in Report A and Report B. If you look over on the far right of your screen you’ll see that Report A had Child 1 and 2, Report B also contained Child 2, and Child 3. The Child 2 appears in two separate reports during the reporting period. 
So moving on to the Agency File this is again the aggregate data that the states submit annually. This includes several CAPTA required items, as well as summary data well it’s all summary data. It includes several categories including prevention services so this is number of children and families served under different funding streams, referrals and reports. This includes the number of screened out children and reports that the CPS agency receives each year. It also includes information about workforce, the number of workers under different categories. It includes an average time to investigation that the state calculates for each year based on reporting period. There’s a new field in the referrals and reports section related again to the CARA legislation. This is the number of screened out infants with prenatal substance exposure. Some states either it doesn’t arise a screened out report of an infant with substance exposure prenatal substance exposure either doesn’t rise to the level of an investigation and its screened out or the state actually has separate programs dedicated to those reports. And they,re screened out of the CPS system to another program. Additional information on child victims reported in the Child File this is information that cannot be reported information on victims that are in the Child File for the same reporting period but cannot be captured in the Child File. So for example child victims in the Child File who also received family preservation or family reunification services in the previous five years. So that’s a summary that’s an aggregate count that the state provides. There’s a category a section on child fatalities for states who cannot report fatalities in the Child File or cannot report all of their fatalities in the Child File. So the child fatality section some states can only report an aggregate count of child fatalities in the Agency File. Other states report in both files they they there’s for whatever reason cannot report all of their fatalities in the Child File. So they have to report some this happens from validation errors or fatalities that come to the attention of the state outside of the CPS system, such as from the coroner’s office or law enforcement. The last section of the Agency File is part C of IDEA Reporting, and this is the number of children who are eligible for IDEA and the number of children who are actually referred for IDEA services. Okay. Now we can move to the next one and I’m going to turn it over to Michael Dineen to talk about a little bit about NDACAN. 
[Michael Dineen] file in the agency file from the NCANDS. The purpose of NDACAN is to promote the scholarly exchange information and data among researchers in the child maltreatment field. We acquire data from researchers and from the national data collection efforts like NCANDS and AFCARS and NYTD and we make the data sets available to the research community and for secondary analysis. The secondary analysis is analysis of a data set that’s already been analyzed or already been collected the the researcher doesn’t do the data collection themselves for secondary analysis. We support information sharing and other ways we have a Child Maltreatment List Serve which is pretty popular. We have an E newsletter called Updata and we provide data analysis opportunities to researchers through conferences and workshops and an annual summer research Institute. We also do not listed here but it’s the we do special data requests when we have an application if people want to if they don’t have the data themselves or they don’t want to do the data analysis themselves they don’t have to download the data they can make a special data request to NDACAN. And that that application is reviewed by the Children’s Bureau and approved, and if approved then we can do the analysis and produce tables or graphs or whatever the people want. That’s called a special data request and you go through our website for that. Then internally to NDACAN the entities what the data look like if it were done in a regular database. Just to show you that the NCANDS Child File has many what would be called tables in database language. A table is just a set of rows and columns. Or you can call them columns are variables and rows are entities. You know the Child File is a the entity the the a record the variables that define the record or the columns that define the record as unique are the report ID and the child ID as Kristen pointed out. And Kristen also described all these entities but each what I wanted you to what I wanted to illustrate here was that each entity has its own variables and each entity could be abstracted from the Child File in its own table. And those can be those can be pulled out of the data by a researcher and then analyzed separately. So you see these connections between like reports and perpetrators is called report per and the entity relationship itself the report and perpetrator has has variables associated with it like the perpetrator age would be at the time of a particular report if the perpetrator a couple years later they would have a different age. So the age is not unique to the perpetrator it might be be different on different reports. And then the relationship between the child and the report through maltreatments. The child on a particular report will have certain maltreatments but on a different report for they might have different maltreatments. So and then the maltreatments themselves have characteristics like the level of the maltreatment and the type of the maltreatment and so forth. So anyway I just wanted to show you in a more graphic way what the entities in the Child File would look like when abstracted into individual tables. Kristen was describing all the things that happen at her firm Walter R. McDonald who takes care of the data for the Children’s Bureau and when Walter R. McDonald goes through all those steps that Kristen was describing they send the data to us and we are the firm that makes the data available to researchers. We do add a couple variables to the Child File on our end and one is called I call it RptVictim means that child was a victim on that report. And I put that in there just to save the researcher having to compute that themselves because they would have to look at each maltreatment and see which if any of the four of up to four maltreatments were substantiated. And that’s all done for you you just say yeah this this person this child was a victim on this report. They might be on another report and they’re not a victim but on this report that they’re a victim. And then I added the variable called state foster care ID (StFCID), that’s simply the AFCARS ID with the state two character code that you know our familiar state postal code like CA for California, capital NY for New York is added to the beginning since it’s text it’s just concatenated to the beginning to make a 14 character code. And the reason I did that was because it’s about linking the tables together you can link the Child File to the Foster Care file or to the NYTD Outcomes file or to the NYTD Services file or to the AFCARS adoption file. So there’re like a family of data sets all sponsored by the Children’s Bureau and they are related on one variable called it’s called the problem is it’s called different it has different names in the different data sets. So I made this variable so that it for two reasons it would make it instead of having to link on two variables, state and foster care ID, or AFCARS ID it’s just one variable that you have to link on, and secondly it has this same name in all the data sets so that you can easily spot it and not have to remember what that variable is called in the different data sets. So also when Kristen described the AFCARS ID, the text file, the entire Child File as an ASCII text file when it comes from us it’s not ASCII because of this particular variable. All the other variables have ASCII values. ASCII is like the first 128 characters that are on your keyboard but there are characters in the AFCARS ID that are beyond the keyboard or ASCII series. So it has to be coded in the different form we call UTF-8 and so that it can handle those oddball characters. A lot of people who use the data have asked questions to our website or to our help email address asking saying there’s something wrong with these IDs they have these weird characters and they don’t look right and there must be some error but those are not errors those are valid encryption characters. When the states encrypt the IDs they do it by character substitution and different states will have different pool of characters that they’re using to do the substitution. And some of them use characters beyond the ASCII range. So you need a different coding system to be able to display those on your screen. Anyway so that’s the variables that we add. 
Then we also do some confidentiality protections we actually take data out of the Child File in the interest of protecting confidentiality. The date of birth is not in the Child File nor is the county of residence of the child. Worker and supervisor IDs are not in this Child File when you get it from us and the incident date is also not in there. There are two actually two counties county IDs in the Child File, one of which you don’t see which is the county of residence but the county that is if the county is identified in the Child File, that’s the county of the agency that was investigating that case it’s not the county of the child necessarily I mean like often of course they’ll happen to be the same county but there are lots of places where a particular county agency takes care of several counties or for some reason or other that child was in that county when the alleged event occurred and they will be investigated by that county. So also for case when the when the case is a maltreatment death, we suppress the state, the county and all of the IDs all the perpetrator ID, the child ID are not in there, they’re just all X’s. We also suppress the county identifier now this is the county agency county of the agency and if they have fewer than 1000 cases then they are masked in the Child File. We set the report date to either the eighth or the 23rd so it’s like within two weeks of of what it was when in the original but we take some of that information out so that people won’t be able to get the exact report date. All the other dates are recorded to match the time span since the report date. Those are the confidentiality protections that we do in the Child File. Then the package that when you order data from NDACAN you’ll get a package of of a documents you’ll get documentation including the codebook and the user guide which you should have you know become very familiar with when you’re investigating when you’re doing your research the codebook and the user guide. There’s also the NCANDS guidelines which are produced by Walter R McDonald and they have a lot of a lot of useful information in there about the data. There are also state mapping forms which show which make a connection between the variables and the variable names on the state side which don’t won’t necessarily be the exactly the same variable names that you we have in the Child File. So they they have a mapping that they show how they what it’s called and how it’s defined on the state side and and how that maps into what variable that maps to in the Child File. Then we also include the Child Maltreatment Report for the year of the data and that’s why we release the data as soon as the Child maltreatment report is available in January. So the the data is released by us more or less simultaneously or released within a couple days after we hear that the child maltreatment report has been published. The data is in various formats: SPSS, it’s also in a SAS file, a STATA file, and then a tab-delimited text format and which is UTF-8 encoded. We also have syntax in SPSS, SAS and Stata for reading the tab delimited text file. That may seem redundant since you’re already getting the SPSS, SAS, Stata in their original form but it you can look at those syntax files if you had a different if you were using a different statistical program and see how the data are defined cuz it’ll show in those text syntax files the values, the variable labels and value labels for each of the variables so you can edit the those syntax files to match whatever whatever application you are using for your for your to do your statistics. Okay that’s my presentation and we’re going to open it up for questions and Erin will will moderate. 
[Erin McCauley] yes thank you very much. Big shout out to all three of our presenters that was a very in depth coverage which I think as a data user can be really helpful cuz I think we get really focused on the specific variables that we use based on our interests. And so hearing an overview of those variables and how they connect and how they were collected has been extremely helpful. So if anyone has any questions you can put it into the chat box and then we can all read it and I will read it aloud and then a direct it to which ever one of our presenters who would like to. So let’s just give people a minute to type. 
okay so we have a question. Do you still have to have an IRB to request the Child File given the confidentiality protections? The data don’t appear any more sensitive than the public AFCARS file. 
[Michael] Yes the answer is yes you do have to have an IRB in spite of the fact that what you’re saying there that it’s doesn’t seem to be any more sensitive than the AFCARS file which you don’t have to have an IRB for. 
[Erin] yeah but you can use the information that Michael’s giving you about how we do the confidentiality changes to to make it publicly accessible and then it kind of different IRB’s will have different procedures but often it’ll be kind of a quicker review because of the nature of the data. 
[Michael] I think it’s more that the the data on maltreatment is more kind of controversial then data about foster care so it’s that’s that’s I think one of the reasons they still require an IRB. 
[Erin] So we have one question. Is there less research conducted on rural counties since they are more likely to have fewer than 1000 kids? 
[Michael]I would say that’s unfortunately true yeah, the suppression of the county identifier is a makes it harder to do the research for those smaller, rural counties. 
[Erin] definitely and it’s just a trade-off between protecting the participants and their confidentiality and then also being able to allow as much research to be done as possible and unfortunately I think kind of the research drawback of that is that we will have have less research on those rural counties but we’ll also be able to protect the people there who were involved in the system. We have another question saying: would you please repeat the comment for parental substance use or abuse, I’m sorry. 
[Madonna] I’m not sure what the question is asking this is Madonna. So during 2016 the Comprehensive Addiction Recovery Act of 2016 amended CAPTA to request additional data elements be collected through NCANDS and that was the number of infants with prenatal substance exposure, the number of those infants who also had a plan of safe care, and the number of those infants who had a referral to appropriate services which are services are designed either for the prenatal substance exposed infants or their caregivers. So those are additional data elements that were added to NCANDS for with the 2018 date. In addition to those we also have caregiver risk factors to identify caregivers with drug abuse or alcohol abuse. So I’m not sure what the question was pertaining to but there are all of those things that we put in the NCANDS Child File and I would suggest that you go to the NCANDS codebook on the Children’s Bureau website to if you have any additional questions about what is included in that or look at the child maltreatment report when it is released in January, or look at prior editions of the report on the CB website. 
[Erin] yes thank you Madonna that’s that’s great that’s exactly what I would recommend. The next question is: is it possible to obtain the data without any modifications made for confidentiality through special data request? 
[Michael] well you don’t get data through a special data request, you get the results of an analysis, usually table. We have in the planning stages a micro-data project where that could be available the user would sign into a special server where the data can’t be moved from that server and they would be able to the analysis on some of these suppressed the things that we suppressed for confidentiality. But that’s that’s still in the works and I don’t I don’t any time line when that might be available but that would be that’s that’s a that’s a service that we really would like to be able to provide but it’s just a matter of jumping through the hoops of making sure that we can do it and still satisfy the confidentiality requirements. We’ve got a plan and we’re trying to work it out with with Walter R McDonald and the children’s Bureau to see if it can if we can do that. 
[Erin] and we have another question saying: is the child ID linked with other data sources, such as data collected from the Department of Ed? 
[Michael] No. 
[Kristen] Yeah at this time the child ID that is submitted to NCANDS is an encrypted version of the state ID. That the states use. Now that ID is a state-level may be the state may be linking it to other data sources within the state, but at the federal level once the Child File is submitted to us, the child isn’t linked to any other data sources including technically the AFCARS file, that is why we collect the AFCARS ID in the Child File in order to link the Child File to the AFCARS file.
[Erin] yes thank you for that. A and then we have another similar question to the one before about the micro-data so were hoping to get something along that lines off the ground at some point but right now we don’t have that available. Can you talk about any limitations of using this data especially due to inconsistency in data collection across states?
[Michael] wait a minute the question about ZIP Code level, the ZIP Code is not a variable in the Child File and probably never will be but I mean who knows about the future but it’s not. We only have county identifiers not ZIP Code.
 [Erin] yes and there’s a Michael has made some really fantastic user guide’s and codebooks so if anyone is curious about a specific variable they can always hop on our website and get those without downloading the data already just to see what’s in there. And Michael really has done a great job over viewing that and that’s the first place I look whenever I’m starting a research project with this data. So how about to the limitations of using this data especially due to inconsistencies in data collection across states.
[Madonna] so this is Madonna the purpose of NCANDS is to provide national data that can be I don’t like to use the word compared, but so that you can do state comparisons across the United States. That being said there are of course different state variations and CAPTA allows for this. CAPTA recognizes that states have their own autonomy and allows for variation. So yes while there are some limitations that you should be aware of this is where I’m going to point to what I consider to be the most important resource of the child maltreatment report and that is Appendix D, state commentary. State commentary is an area for the states to provide additional context as to why their data looks the way that it does. A lot of times states will put in information about state policies, state limitations to collecting and reporting, and state legislation. So if you have a question about why a state’s data looks the way it does, I highly recommend looking at appendix D. If you’re looking at trend data across years you might have to look at the appendix for several years but this is the best place to go to resolve any questions that you might have about reporting. Also in appendix D we provide contact information for the state so that you can feel free to contact the states individually and ask your question.
[Erin] yeah, that’s that’s such a great research and I think you just in order to get the national data there’s always a drawback in this case the drawback is that there’s inconsistencies in the data collection. But you do have really high sample in a national counts so. The next question is: in the future would the archive reconsider including the incident date and the archive NCANDS file? This incident date is required if one wants to analyze maltreatment and care rates over time across states?
[Michael] well we don’t make the decisions about what what variables we can include in the to make available so that’s at a higher level than us.
[Erin] thanks Michael. Our next question is I’m wondering how the parental substance abuse was determined to be the one of the maltreatment reasons?
[Kristen] so I I’m going to kind of make an assumption about what the question is. So if we’re talking about excuse me if we’re talking about infants with prenatal substance exposure based on guidance from the children’s Bureau the states understand that the count of those infants identified in the Child File as infants with prenatal substance exposure must sort of fit into the following categories. They are under one year of age, they have a report source of medical personnel, this includes hospitals, social workers who are acting under the guidance or direction of a medical personnel, and they have a child risk factor of drug or alcohol use. That establishes an infant with prenatal substance exposure. Again this is the first year of data collection so these analyses will be published in the next child maltreatment report. It’s not necessarily a maltreatment many states don’t consider this they don’t consider it as a maltreatment type it’s not listed as a maltreatment type in the Child File. It is simply identifying infants with prenatal substance exposure. They can be victims or nonvictims their maltreatment types don’t have to be linked to the substance exposure. 
[Madonna] So the law was enacted with that amended CAPTA is requiring the state to refer to these children to CPS without regard to whether or not it is considered abuse or neglect. So when a child is is born and if a doctor thinks that there is some substance exposure they must refer the child to CPS regardless of whether or not CPS might consider it to be abuse or neglect. And without regard to whether or not the drug or alcohol is legal or illegal. That doesn’t matter either. The child is just required to be referred to CPS.
[Erin] great so hopefully that clarified that. Our next question is about kind of the access to the data, and it says to use NCANDS for a dissertation a faculty member has to apply to obtain the data on behalf of a student. The student is designated as a “staff member” for the project. Is this form available online?
[Michael] yes the form is available online. You go to the our website and look at the the data sets and choose the data set that you are interested in and click on that and then it goes to a page like say it was Child File. And then it would have “apply for this data” as a click and then you get to a fillable I think it’s a fillable PDF that you would fill and send to us.
[Erin] great thank you. Then we have a question that some data was previously released to researchers understand your working to change this but when was this policy previously in place changed to disallow this? So we never had a policy, we just did a pilot of micro data access so there was not a policy change we just piloted it. The next question is there some children who are in NCANDS who have what seemed to be like obscene numbers of records assigned to them in one case over 1000? Are there any suggestions for thinking about how to approach the data i.e. data carpentry, data reduction etc. when there are so many cases keeping in mind that research questions also drive this concern?
[Michael] I don’t know if Kristen has any information about why those why that particular phenomenon occurs. I know it does but I think what I’ve been told was that the states sometimes have a a placeholder ID that they use and then they don’t delete it when they get it they change it to our real they create a new I guess they create a new record with the real ID and those other IDs are just sitting there and they accumulate lots of cases. Cuz if you look at them they’re not the same people. You know they don’t have the same demographics when you have those IDs that have tons of instances. Do you have anything about that Kristen?

[Kristen] so yeah so this um you know it’s an ongoing validation issue with states and we try to part of the NCANDS project is to provide ongoing technical assistance to states to make improvements to their data files. We have in the past had issues with dummy IDs and the issue is that if you’re looking over multiple years whether a state has resubmitted those files to correct for that or whether they could even resubmit to correct for that. We also have issues of states that change systems and often times updating or upgrading their their system their SACWIS [inaudible] whatever can cause issues with IDs and linking IDs to prior years. So we work we try to work very closely with the states to avoid those kinds of issues but it doesn’t always happen at the outset. And then of course you end up with at least one data file that does have some data issues. The I’m trying to think suggestions for thinking about how to approach the data when there are so many cases I think that the first step would be again going back to commentary for that particular data year or years to that particular state To either find information in the commentary or contact the state directly. This is like I said sort of an ongoing technical assistance improvement program if you well I guess and we try to collect we try to encourage the states to include that information in that commentary as well. 
[Erin] Yes that’s great such a helpful perspective and that message that it’s kind of an ongoing process and were also helping states to better collect their data is is super important. For people who are interested in IRB applications who might not have an IRB, Andres gave us a little information in the chat about that so if that applies to you I suggest reading it. We have another question saying state mapping forms provided in the documentation for any given Child File here do not cover all states and forms for some states reflect information collected in prior years. Is there a way to determine which variables were reported by each state in each year aside from going through all the state mapping forms for each year?
[Michael] I don’t think there’s a way without doing that like that’s what I would have to do.
[Erin] yeah that’s what I would do as well. Go ahead Kristen.
 [Kristen] I was just going to make a a mention about the mapping form. So we want to go back to that statement that Madonna made way back at the beginning of the presentation about how NCANDS is a voluntary system.
 [Kristen] so we strongly encourage states to send in updated mapping forms every year. We work closely with states to update specific mapping forms especially as issues arise during technical assistance or information is provided to us about a state’s you know a change in policy, programming, data collection, anything that might affect how they are mapping state variables NCANDS variables. We provide ongoing technical assistance in the form of technical bulletins, and briefs, and webinars to encourage them to make sure that their mapping forms are up to date and that they are using correct definitions and instructions. But we cannot we do not cannot require them to submit mapping forms to submit a complete set every year or even across years. So again it goes back to the commentary for additional information and contacting the states directly when a researcher finds a gap in a mapping form.
[Erin] great thank you very much yeah. Always remember to important this is voluntary and that the WRM is out there helping states through that whole process with really in-depth technical assistance we really wouldn’t have any data if it weren’t for them. So thank you very much.
 [Michael] can everybody see these questions?
[Erin] yes yeah the chat box is open to everyone
[Michael] well unimportant one is from Andres who’s our on our team. He says "researchers at an organization without an IRB may apply to receive in NCANDS Child File from NDACAN and the application will be reviewed by the NDACAN director. So it goes on to say there is a space on the application for restricted data sets form for the researcher to describe their organization and their organizations data protection policies, infrastructure, existing certifications in handling other sensitive data.” So the I guess those would be some of the factors that would be taken into consideration when approving a application for data from an organization that doesn’t have an IRB.
[Erin] yes thank you for highlighting that Michael. You know our the archive does a lot to support users we try not to just be a place where you can come download data. We really assist in any way possible and that includes if you’re in an organization where you don’t have an IRB but you still want access to the data. And it seems like that is the end of our questions. So before we had out i just want to highlight that next week Michael’s coming back and then we’ll also have Tammy White and she and Michael will be talking about the other two data sets in our administrative cluster. The a foster care file and then the NYTD files which are the youth who are identified as aging out of foster care without finding a permanent placement and it follows them after release. But  we’ll hear more about that next week and it will be Wednesday from 12 to 1 just like this week. Thank you everyone so much for joining us it was a really great session and just thank you again to our presenters we’re so lucky to have you guys all here all three of you giving different perspectives and really helping us understand where these data come from and what we can do with them. So with that I’ll say thank you to everyone and hopefully we’ll see you next week.
[Voiceover] The National Data Archive on Child Abuse and Neglect is a project of the Bronfenbrenner Center for Translational Research at Cornell University. Funding for NDACAN is provided by the Children’s Bureau.