After the Webinar: Using Murder Data to Help Justice Professionals Work Cases and Solve Crimes. Q&A with the Presenters

Webinar presenters Eric Witzig and Thomas Hargrove answered a number of your questions after his presentation, Using Murder Data to Help Justice Professionals Work Cases and Solve Crimes.  Here are just a few of their responses.


Audience Question: In the UCR data for your site, do you adjust for the number of months that agencies report? The UCR data is full of agencies recording partial data. 

Thomas Hargrove: We never try to make an adjustment for missing data. That does happen. Right now, we don’t have a procedure for indicating that there are missing data. We might be able to program that in the future. The problem is it’s hard to know why we don’t know about records that were not reported. Sometimes, the police agency didn’t make a report because they had no qualifying case. Sometimes they didn’t make a report because they were having a technical problem or that they fired the guy who runs the statistical reporting project. Who knows? It’s hard to understand why data were not reported whether it was because there was no data or there was technical or other problem involved. Unfortunately, the way that the FBI reports data is whether it’s missing or whether it was a lack of murder. Those events are recorded as zero. zero is the number. It would be wonderful if the UCR reported something other than 0 perhaps a blank indicating data are missing. We don’t get that. There might be ways to identify when cases are missing. Right now in our presentation, we’re not doing that.



Audience Question: Does your data suggest that are homicides in high crime areas, in larger impoverished cities tend to go unreported compared to a small town in high profiled cases? 

Thomas Hargrove:  No, actually we find that major cities tend to do a better job reporting. We don’t have very many complaints even among the most challenged of police departments. What we do find is police are doing a very good job of reporting the occurrence of murder. They sometimes make mistakes to their detriment on whether or not the murder was cleared. In fact, we’ve had several disappointing conversations with police departments. They have been reporting data that make them look far worse than their actual performance might indicate. That’s unfortunate. There’s some misunderstanding under the reporting rule. When a case is cleared, that clearance that should be reported whenever the clearance was made. Some police departments think that they should only report a clearance if the clearance occurred the same year as the occurrence of the murder. That’s not right. If the JonBenet Ramsey murder was solved today, the police in Colorado should add one for the clearance account for this year. It doesn’t matter when the murder occurred. There can’t be some misunderstanding. Again, we recommend the police to download a copy or uniform crime reporting handbook which you can get at our website.



Audience Question: Is there any way for the project to know if the homicide incident was not reported via the FBI UCR program? 

Eric Witzig: There are a couple of ways to do that. We have done a couple of things. One, we have compared and contrasted maps, data from the UCR with data from the Centers for Disease Control. CDC collects data on all death certificates in the United States including homicides. We have discovered that CDC has more homicides recorded in their database than recorded in the UCR. Yes, there is a gap there. Moreover, recently we have been contacted by members of the media in the western state who have been concerned about the homicide of Native American women and the disappearances of Native American Women from reservations and other places in their states. We have done some spot checking on these and have found that they too have not been reported as homicides. About half of them are not reported as homicides. So the short answer to your question is yes, we have discovered that there are cases that are not reported in the UCR that are in fact bona fide homicides.

Thomas Hargrove: If you go to our website,, there are news events on the home page. You can look at the second news item. It is about this phenomenon that about half of all Native American murders were not reported to the FBI. We are discovering aspects about this problem daily. It turns out I guess that I should be happy to tell you all. It turns out that amongst failure to report important crime data are, in fact, federal law enforcement, the FBI, the Bureau of Indian Affairs, The US Park Police and all of the investigative agencies with the Department of Defense has not been reporting data to the Uniform Crime Report. This seems to violate an Act of Congress passed in 1988 which Congress instructed all law enforcement in Federal Government to report to the FBI and that the Justice Department shall make those data available to the general public, everyone who uses the uniform crime report which would include us. We don’t know where we’re going with this. We’ve been in contact with members of Congress and we sent letters to all of the Bureau heads including Christopher Wray the director of the FBI. It’s possible that we are going to court in which where it’s possible we’ll have to bring freedom of information act case against all law enforcement in the Federal Government. It is a major reporting flaw that becomes obvious when you look at specific types of murders like Indian murders because they tend to be more than others investigated by federal authorities. We do have a method of finding where the unreported bodies are buried literally. We are reviewing that. This is a good thing because, I think, we are about to dramatically improve the quality of the data but it’s going to take a while. We are gearing up for a serious conversation with the entire Federal Government.



Audience Question: To that point, I wanted to let you know Kyle also asks the question does this data include homicides on tribal lands or military bases? You certainly just referenced the tribal land aspect of Kyle’s question. 

Thomas Hargrove: Yeah. The missing data are also from military cases. We are actually a little concerned that there might be a series of connected murders in and near Fort Leonard Wood Missouri. We’ve had trouble investigating those now we think we know why. It’s because records by military police have not gone into the Uniform Crime Report. We may have considerably more knowledge in a few months or maybe a year but we are convinced that we will get these data eventually because we have an Act of Congress behind us in this request. There are no military police investigated homicide in our data.



Audience Question: Is there any possibility that this algorithm in your database might become applicable to us? 

Eric Witzig: I think the answer to that question is – the short answer is yes. The somewhat longer answer is that each nation needs to create or have if they have not done so already its own UCR-like and SHR-like database of homicide. Once that database is present, then the algorithm can be used to search it. The short answer is yes.

Thomas Hargrove: We do have conversations with people around the world and also a lot of people would like to see an algorithm like ours to fly to their own record. The key is whether or not other nations have something similar to the supplemental homicide report. So far, I have to tell you the answer is no. That the United States is unique among all of the world in that we made available a public record of all the nation’s homicides. It’s not a complete record as we might wish but when you have complete reporting, when you count something carefully, you can do magical things such as detecting serial murder. I believe that there will be an attempt, in fact, an attempt is underway in Canada to try to develop an equivalent database to our supplemental homicide report. If they are successful and they have  (indiscernible/audio issue 1:03:42) possible then we possibly can help establish a MAP Canada where we would happily give them our algorithm or rewrite the algorithm to work with their data. We’re happy to work with folks around the world. We do have that conversation around the world.



Audience Question: Approximately how many serial killers do your data indicate might be active in the US at the current time?

Thomas Hargrove: Fools rush in where wise men fear to prod.

Eric Witzig: That’s a complicated answer. It involves a couple of different things. Let me look at it and I need a couple of seconds to get my head together. As you probably know the United States best database of known serial killers and their victims has been maintained by Dr. Michael Aamodt from Radford University. I’ve spot checked it from time to time. It’s pretty comprehensive and pretty accurate. The MAP algorithm is going to look for what it believes might be serially related offenses. It can discover those that are known to law enforcement such as Mr. Ridgway up in the Pacific Northwest, a case that was investigated by now Congressman, then Sheriff’s deputy David Reichert. Unless until law enforcement conclusively states that yes, these states are conclusively linked to each to the other then we won’t really know, won’t we?

Thomas Hargrove: As to how many undetected serial killers there could be, only the devil knows that. I believe the answer is in the hundreds perhaps even up to two thousand. That’s a hard thing to know. The algorithm is signaling a lot. A lot more than anyone would like to believe.



Audience Question: Regarding the FBI not reporting data, don’t most cases start at the local or state level before the FBI becomes involved? If so, wouldn’t data of the crime be originally reported by the local jurisdictions? 

Eric Witzig: That’s a great question and one that we have been looking at. Yes, most homicides in the US is the province of State and local law enforcement because they have “Original Jurisdiction.” However, there are certain offenses for which the FBI has original jurisdiction. Depending on its treaty with Native American Nations in the United States, either Indian police or the FBI will have original jurisdiction with the homicide on an Indian reservation. There are other instances where the FBI has original jurisdiction such as assault upon the President of the United States and other highly placed government officials. Ordinarily, state and local law enforcement has original jurisdiction and should be reporting, the FBI can be requested by the state and local law enforcement to assist in an investigation and the responsibility of, my humble opinion, would rely upon the local entity to report same.

Thomas Hargrove: We are aware of individual murders that are primarily investigated by the FBI that were not reported to the Uniform Crime Report. If there is anyone out there that would like to talk to us about this issue, we would love to talk to you.



Audience Question: What is the R stats program or where can it be found? She said she is familiar with the SPSS but have never heard of this particular program. 

Thomas Hargrove: R is an open source system. She can download it for free from quite a few places. It does have a bit of a learning curve. I think SPSS is easier to learn. By the way, there is an open source version of that SPSS developed like a new software called PFTT. That is actually is what we use. It works great. You can get an open source version of SPSS and you can get R for free. We recommend that. I’m told that R is becoming the common software in the statistical world. She’s welcome to download. Call me if you like some guidance on how to do that.



Audience Question: If the results of data analysis suggest the possibility of a serial killer, should we approach a law enforcement agency and how do you suggest that they approach this? 

Eric Witzig: We’re happy to take the question. We are pleased to do that and pleased to pass along what we know for the aid of state and local law enforcement. With regard to – go ahead, Thomas.

Thomas Hargrove: We actually have rules of engagement, what will we do in our algorithm indicate that there is a possible series. We have established that we will not go public with that concern until we’ve looked into it in more detail. That includes putting names and narratives of the victims and the crimes. When you do that, when you go through a newspaper account or statements put out by local police when you do that exercise, you will quickly come away with a yes or no, is this likely to be a series?  That is our procedure. We don’t go to the police unless we first use public sources to determine the names and narratives of these crimes. When you do that, you’ll know to your guts whether or not that could be a connected series.

Eric Witzig: When we look at this, as I mentioned earlier, based on the data points that the Supplementary Homicide Reports collect, if you have the experience and the knowledge working with serial homicides, you can rather quickly tell and point to those offenses which may have been committed by the same offender because of the pattern that the offender used to commit the offense. I would not recommend this for laypeople. I would suggest that there is a knowledge curve that needs to be applied. When we call state and local law enforcement to say hey you might want to take a look at this, we have a number of data points, information that we can tell them to suggest why this series may, in fact, be part of a series. It’s up to them to say yes or no.

Thomas Hargrove: If you go our website, at the homepage and look at the first news item about the Chicago series. At the end of that report, there’s a click here and you can download the report that we prepared for the Chicago city council which explains our reasoning for why we are certain that there are serial killers active in Chicago. That report was, in the end, resulted in the creation of a task force that is currently at work.

Eric Witzig: In other words, to make the answer a little bit shorter and maybe a little bit clearer, once the algorithm goes on red alert, we need to look at the individual cases and hopefully possibly collect more data from newspaper and other reporting to flesh out the data a little bit to determine whether or not it is part of the series before we number three, make contact with law enforcement.


Click Here to Watch a Recording of Using Murder Data to Help Justice Professionals Work Cases and Solve Crimes.



Additional Resources
1 year ago
Using Murder Data to Help Justice Professionals Work Cases & Solve Crimes
The Murder Accountability Project (MAP) is a non-profit initiative from experts in the field of inve […]
1 year ago
Crime Scene Responsibilities for the First Responder
Crime scene investigation has got to be the most exciting, if not glamorized, part of policing – a […]
2 years ago
Implementing Body Worn Cameras: An Interview with Michael White
Rolling out a body worn camera program for your agency is more than just handing out the equipment, […]
3 years ago
Solving Gun Crimes Infographic
Feel free to download and share this infographic!