by Patrick Siewert, Principal Consultant, Pro Digital Forensic Consulting
Since introducing our private sector clients to the impact that cellular call detail records (CDR) analysis & mapping can have on their cases, we’ve had a lot of robust discussions with litigators and clients about the veracity and value of this evidence. CDR analysis has been used for decades in law enforcement to help prove or disprove the approximate location of criminal defendants in major crimes. Only in the past several years have civil litigators and insurance companies also been introduced to the value that this evidence can have on their cases and/or claims investigations. In the time we’ve been conducting CDR analysis, we’ve worked on varying types of cases from criminal prosecution for smaller prosecutors’ offices to domestic litigation to help prove/disprove cohabitation to high-dollar insurance claims to help determine if the claim and associated statements made under oath are verifiable with regard to location. This specialty offshoot of digital forensics requires constant knowledge updating with regard to carrier practices and specialized training and tools to be able to perform these analyses effectively.
However, mainly among the Criminal Defense Bar, the notion has been put forth that CDR analysis may be “junk science” and therefore potentially unreliable as evidence in legal proceedings. One high-profile case in which CDR analysis was used to obtain a conviction was the case of State v. Adnan Syed, chronicled in the Serial Podcast. However, as we’ve seen more recent developments in that case unfold, the “junk science” claim doesn’t necessarily lie with the practice, rather with the potential practitioner. Indeed, even in computer forensics, certain vendors of forensic tools like to claim their tool has been “validated in court”, when in reality it is the examiner and their competence that needs to be validated in court. The tool (or in this case, the cellular records) is/are just a dataset that needs to be analyzed competently to be introduced as evidence in a legal proceeding.
Toward the end of establishing that CDR analysis is not “junk science”, here are three salient points that will help debunk the myth that these records and their associated analysis are not worthy of evidentiary status.
Reason #1: Cellular Records Are “Pure” Evidence
What do we mean by “pure” evidence? Consider for a moment other types of digital evidence that are analyzed for use in court, such as the cell phone itself or a computer system. These items are generally affected by the user to a great degree and therefore can be open to some scrutiny about the weight and value they hold. Cellular Records are only available via court order or search warrant to the cellular provider. A Verizon Wireless customer cannot call customer support and ask for their cellular call detail records with historical cell site data. The provider will not provide this data this absent legal process. This means the user has very limited (if any) ability to manipulate the data, which makes the evidence about as pure as it gets.
Furthermore, the record-keeper has no vested interest in altering the evidence. In fact, they have every reason to maintain better, more accurate records! It is a fact within the cellular industry that CDRs were never meant to be used as evidence in legal proceedings. CDRs are kept by cellular providers so they can log and analyze their own networks for efficiency and to increase overall customer experience on the network. Simply put, the records are kept for customer service purposes and cellular companies don’t make money by having poor customer service. It is a fortunate byproduct that these records may be obtained via legal process and analyzed for potential use in legal proceedings. This is why cellular providers don’t maintain these records indefinitely, as detailed in our 2017 article Cellular Provider Record Retention Periods.
Name another type of digital evidence that the user never touches and to which they generally don’t have access!
Reason #2: Automated Tools Have Greatly Decreased The Human Error Factor
Back in 2001 when the incident detailed in season 1 of the Serial Podcast occurred there were few, if any, automated tools with which to conduct CDR analysis. In modern casework, we have many options for automated tools analysis, including CellHawk, ZetX, CASTviz, Map Link and Pen Link, as well as some others. Use of automated tools can save time and greatly reduce error, but they come with a few warnings:
- Not all tools are created equally. If you’re using a tool that is free [to law enforcement], you’re generally getting what you pay for.
- Don’t rely on the tool to do all of the work. Automated tools are great, but they cannot tell you if someone likely shut their phone off or sent a call to voicemail or left their phone in one location while committing an offense somewhere else. Only manual analysis of the data and the behavior of the user can help verify these conclusions.
- VALIDATE! If an automated tool is telling you something, make sure to always refer back to the original record for validation. If an automated tool is citing a GPS coordinate for location, make sure you validate that there is actually a cell site at that location.
Reason #3: Trained, Experienced Analysts Don’t Deal in “Junk Science”
One of the traps digital forensic examiners of all ilks are susceptible to is the drawing of conclusions not based on fact. While it’s true that a trained, experienced professional may reach conclusions based upon device activity, those same conclusions have to be rooted in some facts at some point. The trap that sometimes rears its ugly head is when we reach conclusions that are either outside of our expertise or are not supported by the data.
There are several traps documented in litigation over the course of the life of CDR analysis in legal proceedings that have led to the claim of “junk science”. Probably the biggest of these (and the one cited in the article linked above and again here) are conclusions about cell site range. As analysts, we are not cellular engineers and we cannot be engaged in speculation or discussion about the “range” of a particular cell site. This is why in most cases we approximate location of the target device in the investigation and do not get entwined in discussions about cell site range. Even if we were fortunate enough to have propagation maps from the cellular provider which detail the effective/optimal range of a cell site, we still won’t draw conclusions about range. It is not within the expertise of most analysts to discuss range. That is for a cellular engineer to conclude, not an analyst of cellular records.
There are behaviors and activity that the records can tell us about, however. A trained analyst can usually tell of the phone was off, or if a call was sent straight to voicemail, or if the phone was left in one location for a prolonged period.
At the heart of the records is user behavior. Is there a pattern of behavior that is not adhered to during the time of the alleged incident? Is there link analysis that can be done to confirm likely associates or accomplices? If there are alleged accomplices, does normal text or call activity cease with these persons during the time frame of the incident?
All of these items and more can help lead a trained, experienced analyst to conclusions with a reasonable degree of certainty, but with most of these items, we require a larger dataset to compare the behavior at the time of an incident with behavior at other times. An analyst cannot identify these behaviors with 24 or 48 hours’ worth of records. This is not enough data from which to draw conclusions about behavior. This is also why we highly advise obtaining at least 30 days of records on either end of the incident, preferably more. More data is better when it comes to CDR analysis.
The ultimate test of whether or not the conclusions based upon trained, experienced analysis of the records is “junk science” lies with the competencies of the analyst. One who draws conclusions not based on facts is what leads to an otherwise valid form of data analysis to be dubbed “junk science”.
Wrapping It Up
In any forensic discipline, there is the possibility for human error or oversight. We’re not infallible, after all, and we can’t be expected to be perfect all the time. But CDR analysis is the only one in which the term “junk science” has been bandied about quite a bit. Deeper inspection of the issues involved in each case where this claim has been made can provide lessons for current and future analysts to read and take heed. It’s when our conclusions span beyond the breadth of our expertise and what the data tells us that we get into trouble. Ultimately, everyone wants to see justice done. If we can use CDR analysis successfully in litigation without reaching past our ability into conclusions that are open to extreme scrutiny, justice will be served.
About Patrick Siewert
Patrick Siewert is the Principal Consultant of Pro Digital Forensic Consulting (www.ProDigital4n6.com), based in Richmond, Virginia. In 15 years of law enforcement, he investigated hundreds of high-tech crimes, incorporating digital forensics into the investigations, and was responsible for investigating some of the highest jury and plea bargain child exploitation investigations in Virginia court history. Patrick is a graduate of SCERS, BCERT, the Reid School of Interview & Interrogation and multiple online investigation schools (among others). He continues to hone his digital forensic expertise in the private sector while growing his consulting & investigation business marketed toward litigators, professional investigators and corporations, while keeping in touch with the public safety community as a Law Enforcement Instructor.