15

For homicide detectives, the clock starts ticking the moment they are called. Their chance of solving a murder is cut in half if they don't get a lead within the first 48 hours.

This was the lead in message for the "Investigation Discovery" channel program The First 48.

Is this true? If so, what are actual odds of solving a murder before and after this time limit? Is this time limit started at the time of the crime or the time the crime is discovered? Is there a more fine grained analysis of the odds of a murder being solved given a timeline of when the first clues are found?

N.B.: The quote specifies having a lead, not being solved in 48 hours. Some parts of answers seem to bias toward being solved or solvable within that period.

ventsyv
  • 7,136
  • 2
  • 27
  • 44
user53816
  • 448
  • 1
  • 3
  • 13
  • 7
    I've always wondered if investigators just kind of give up after this timeframe, thus making it self-fulfilling –  Sep 07 '20 at 00:42
  • 9
    I think there are cases that are easy to solve, and cases that are hard to solve. If you spent 48 hours searching for clues and didn’t find any, the case is likely a harder one. But if the police was called only after 48 hours, it may still be an easy case. Of course at some point evidence starts deteriorating. Rain wiping blood away, out-of-town witnesses returning home etc. – gnasher729 Sep 07 '20 at 09:09
  • 3
    Strangely, after some years chances go up. Friends of suspects are not friends anymore, people afraid to come forward are not afraid anymore, and so on. – gnasher729 Sep 07 '20 at 09:11
  • 3
    @gnasher729 that certainly seems reasonable. Is there any where you can look for data to back up your theory? – user53816 Sep 08 '20 at 01:44
  • 3
    @gnasher729 I agree, the first example that comes up to illustrate is to replace "solving a murder" with "finding my car keys." Most of the time you'll find them within two days, the rest of the time they're likely in a sewer drain. – Cristobol Polychronopolis Sep 09 '20 at 20:49
  • It needs to be noted that the probability of a solution will probably vary widely by location. Even ignoring differences in the skill of the investigators, there will undoubtedly be a significant difference in the case of a murder committed in a suburban middle-class neighborhood vs one committed in a "slum". And there would also be differences for rural scenarios. – Daniel R Hicks Sep 10 '20 at 00:46
  • @DanielRHicks it should be fairly easy to control for those variables. – phoog Sep 15 '20 at 14:48
  • @phoog - But why bother? What benefits are there to knowing this with more accuracy? – Daniel R Hicks Sep 15 '20 at 17:17
  • @DanielRHicks outside the context of this claim, I don't see much benefit. Perhaps that suggests that the claim is not notable. – phoog Sep 15 '20 at 17:30
  • @DanielRHicks If accurate, than placing more resources on an investigation during that period could provide a better probable outcome. – user53816 Sep 16 '20 at 18:19
  • @psaxton - So could improved social programs. – Daniel R Hicks Sep 16 '20 at 20:18
  • @DanielRHicks the claim has nothing to do with social programs. I'm not sure you grasp the point of this stack. If you have nothing further to contribute to refining the question, please refrain from further comments. – user53816 Sep 17 '20 at 15:11
  • @psaxton - My main point is that the inherent variability of the number is so large that the difference between 2 days and, say, 10 is insignificant. – Daniel R Hicks Sep 17 '20 at 20:44
  • @DanielRHicks that could be a perfect answer. Provide a link to the studies or data with a layman description how/ why your sources prove it and I will accept it as the answer. – user53816 Sep 18 '20 at 13:44
  • Is the claim that "Leads must be found within 48 hours or it's going to get a lot harder" or is it just that "Most leads leading to a solution usually get found within 48 hours"? – komodosp Sep 23 '20 at 11:31
  • What does 'solved' even mean? 'Results in an arrest and conviction'? A sound conviction? Even if that conviction is unsound, overturned on appeal, convicts the wrong person? Or a conviction for a lesser offense? (e.g. assault)? IIUC as far as gang-related homicides in the US, convictions often depend on informants, unreliable eyewitnesses, forensic errors, plea bargains, sentence reduction, jailhouse snitches, etc. Just because the defendants don't have the money to appeal or challenge the prosecution. Compare the conviction rates when defendant has public defender vs private attorney. – smci Dec 10 '20 at 05:05
  • For example was the killing of Trayvon Martin 'solved' or not? Or [the 21-year saga of eccentric NY heir Robert Durst for allegedly killing Susan Berman in 2000](https://en.wikipedia.org/wiki/Robert_Durst). Durst still hasn't gone on trial, and he's almost surely spent more on criminal lawyers than most of us earn in our lifetimes. And he's sure to appeal. Is that considered solved? – smci Dec 10 '20 at 05:13
  • @smci I cannot provide the definition of "solved" for the quote, it is not my quote. Personally I would consider the killing of Trayvon Martin solved: George Zimmerman shot him and was found "Not Guilty" of second degree murder. The identity of the killer is not in question even though he was not convicted. I haven't even passing knowledge of Robert Durst to weigh an opinion. – user53816 Dec 11 '20 at 20:09

2 Answers2

6

The best research I'm aware of is a Washington Post investigation of 8,000 murder arrests. Their data journalists found that half of the arrests happened in 10 days or fewer, not 48 hours.

A Washington Post examination of 8,000 homicide arrests across 25 major U.S. cities since 2007 found that in half of the cases, an arrest was made in 10 days or fewer.

The analysis underscores what police leaders and homicide experts have said about the passage of time working against detectives. But it also dispels the notion of a “48-hour rule” that most cases, if solved, are wrapped up in two days. Only 30 percent of the cases led to an arrest within that time frame, the analysis found.

So the frequently given 48-hour rule of thumb doesn't seem very accurate, at least for homicide cases. However, a graphic further down in the Post article does show that the likelihood of an arrest declines steeply within just a few days.

Of course 'finding a lead' is an ill defined term and not necessarily the same as arresting a suspect. The Washington Post piece quotes a police sergeant who argues that 48 hours for 'identifying a suspect' isn't far off:

Evidence collection and lab tests often delay an arrest beyond a couple days. “It’s probably more accurate to say that you had a suspect identified in the first 48 hours,” said Sgt. Greg Van Heyst, who supervises the Tampa Police Department homicide unit.

Ultimately one would need to come up with a clear set of criteria for what constitutes a 'lead' to accurately confirm or dispel the claim.

waltzfordebs
  • 731
  • 4
  • 10
  • 4
    I would think a "lead" is a person suspected to be either a witness or the perpetrator. That seems how the term is typically used. –  Sep 09 '20 at 14:28
  • @fredsbend that's one possible definition. It seems difficult to systematically analyse from police reports though. Witnesses in particular can include a number of people who become part of the investigation files because they _might_ have relevant information and later turn out to be totally irrelevant – waltzfordebs Sep 09 '20 at 19:39
  • Note that this is ignoring the part of the statement "...are cut in half...". It doesn't say it suddenly becomes impossible. Only that it becomes harder. – Keith Morrison Sep 10 '20 at 20:21
  • I'd listen to @KeithMorrison if I were talking about solving crimes. – Ink blot Sep 13 '20 at 22:23
  • The median time to an arrest is not obviously related to the probability of solving a murder case given the time at which the first lead is identified (whatever "solving" means, and, for that matter, "lead"). – phoog Sep 15 '20 at 14:22
  • I'm accepting this answer as it at least provides a link to a WaPo article and the question has been open for nearly a month. If another answer is posted with better data I may change the accepted answer. – user53816 Oct 02 '20 at 15:39
2

This question does not have a country tag, although there seems to be an implicit US assumption. In order for the claim to be feasible there needs to be a significant proportion of murders that go unsolved.

This seems to be the case in the US, where around 40% of murders are unsolved (source, not super reliable but they claim to quote the FBI which should know about this). However, this is not the case in for example Germany where less than 10% of murder cases go unsolved (source in German).

Hence unless you believe that the German police solves almost all murders in less than 48 hours, their chances of solving a murder after 48 hours must still be well above 50% to get to over 90% total solved cases, so they chances to solve the case can't be halved. I believe the proportion of solved cases is approximately similar to Germany in most other EU countries but haven't looked for any sources.

quarague
  • 1,474
  • 9
  • 13
  • 3
    Why shouldn't we believe that German police have leads in the first 48 hours for most cases? – Oddthinking Sep 10 '20 at 00:02
  • You are correct the program is US based. I didn't post a location tag because the claim doesn't specify a location. I considered adding an applicable region to the large list of questions I've already asked, but don't want to expand the scope of the question any further. – user53816 Sep 10 '20 at 01:50
  • If someone can provide a data set confirming or confounding the claim, maybe a region for where the claim is true could be extracted. – user53816 Sep 10 '20 at 01:52
  • "there needs to be a significant proportion of murders that go unsolved": That doesn't follow at all. For example, if the proportion of cases in which a lead is found in under 48 hours is very high then the proportion of cases solved overall will be very close to the probability of solving such cases. If that probability is very high then the proportion of unsolved cases will be small even if the probability of solving other cases is *less* than half as great. Granted, those two figures probably aren't that high, but there isn't anything given in the question to support such an assumption. – phoog Sep 15 '20 at 14:36
  • The claim isn't about whether the case is solved within 48 hours but whether it is solved eventually considering whether the first lead is developed within 48 hours. So it is not necessary to believe that German police solve almost all murders within 48 hours to accept the claim as stated. – phoog Sep 15 '20 at 14:45
  • @phoog I guess you are right. I depends on exactly what counts as a lead, which is not really well defined. If a married woman is murdered, her husband is usually the first suspect. Does that already count as a lead? Only if later on it turns out to be true? – quarague Sep 16 '20 at 06:40
  • I absolutely agree. The claim is not defined precisely enough. In addition to a definition of "lead," we lack a definition of "solve." For example, suppose the police settle on a suspect who has died, or evades justice as a fugitive, or is otherwise unable to be tried in court. Is that considered "solved"? What about wrongful convictions? What about partially solved cases? And, as your comment suggests, what about cases where the only leads arising in the first 48 hours are false? – phoog Sep 16 '20 at 14:54
  • @Oddthinking Conversely, why should we believe it? – CGCampbell Sep 18 '20 at 13:04
  • 1
    @CGCampbell: I am not the one making the claim here. My intuition is that most murders are domestic in nature, and the most likely killer is immediately obvious to everyone. – Oddthinking Sep 18 '20 at 13:36