History of Work on Domestic Violence

View Segments Segment :

  • Citations
  • Add to My List
  • Embed
  • Link
  • Help
  • Citations
  • Add to My List
  • Embed
  • Link
  • Help
Successfully saved clip
Find all your clips in My Lists
Failed to save clip
  • Transcript
  • Transcript

    Auto-Scroll: ONOFF 
    • 00:06

      [History of Work on Domestic Violence]

    • 00:11

      LARRY SHERMAN: Hello my name is Lawrence Sherman.[Professor Larry Sherman, Director of Criminology,University of Cambridge] I'm directorof the Institute of Criminology at the University of Cambridgeand a professor of criminology at the University of Maryland.For over three decades, I've been doing experimentswith police departments, testing different waysto try to protect people from becoming victims of crime,

    • 00:32

      LARRY SHERMAN [continued]: to try to reduce the harm against victims of crimewho are attacked by criminals in various ways,and generally to improve the effectiveness of policeagencies through scientific knowledge--not about blood stains, but actually about predictingand preventing crime.In this case study, we will look at a live research

    • 00:55

      LARRY SHERMAN [continued]: that has probably provoked more controversy than anyof my other experiments and that is how the police shoulddeal with domestic violence.[Police Response to Low Harm / High FrequencyCases of Domestic Violence]In this case study, we'll begin with the first experimentI did in collaboration with police on their response

    • 01:18

      LARRY SHERMAN [continued]: to the low-harm, high frequency kinds of callsthe police receive millions of times a year around the world.We'll talk about the reaction to that study, whichled to a very large increase in peoplegetting arrested in the United States for domestic violence.And then talk about how that experiment was repeated

    • 01:41

      LARRY SHERMAN [continued]: and actually found different results that were notas welcome and did not lead to as much changebecause people denied the methodology beingsupportive of the results and came up with other reasonsto ignore the research.And then, the third part of this case study,we'll look at the long-term follow

    • 02:01

      LARRY SHERMAN [continued]: ups of some of those further studiesto show how the discovery that doing whatis thought to be best for the victimcan actually kill them and double their death rate.Even that met with resistance and the kindof political reaction that doesn't reallygrapple with the science.We'll conclude by comparing my experience

    • 02:24

      LARRY SHERMAN [continued]: with domestic violence research to the experience of doctorswho promoted the idea that their colleagues should washtheir hands when they treat one patient after they've treatedanother patient and not spread what turned out to become germseven before germ theory had been invented.

    • 02:44

      LARRY SHERMAN [continued]: And the difficulty that we have, even today, in getting researchto be widely accepted in changingthe practice among doctors, just as it is widelyaccepted research being ignored in relationto domestic violence.So let's start in 1981 when on Saint Patrick'sDay in Minneapolis, Minnesota, 35 constables of that police

    • 03:09

      LARRY SHERMAN [continued]: department, with the unanimous endorsement of the elected citycouncil of Minneapolis, went out on the streetto randomly assign arrest.What do I mean by randomly assign?I mean that there was a lottery methodto decide who would be arrested and who wouldn't, if thereis sufficient evidence that somebody had slappedor kicked or punched an intimate partner

    • 03:32

      LARRY SHERMAN [continued]: with whom they were residing or had resided.Now, the question of whether theyshould be arrested at that time was really quite controversialbecause the police were not making arrests.So the new idea was to try arrestcompared to just calming things down and leaving the scene.That's not widely understood these days.And one reason the city council approved the random assignment

    • 03:56

      LARRY SHERMAN [continued]: of arrest was to test it out to see if it wasgoing to be safe and effective.And actually what happened is that the arrest option hadthe lowest rate of repeat offending over the next sixmonths.It was only 10% compared to sending the suspect awayfrom the home for the night where the repeat offending ratewas 24%.

    • 04:17

      LARRY SHERMAN [continued]: This finding made headlines in over 300 newspapersaround the United States.It was reported in The London Times,in the Sydney Morning Herald, all across the English speakingworld.The startling news came out that if you only arrested,not even prosecuting successfully,if you arrest these abusers of their intimate partners

    • 04:40

      LARRY SHERMAN [continued]: then you would be protecting the victimsfrom further abuse in a substantial numberof these cases.Now that was actually a very promising finding thatneeded to have more research.In science, we rarely accept the result

    • 05:00

      LARRY SHERMAN [continued]: of one experiment, especially whenif we're looking at different conditions, different kindsof people, different kinds of places,you could get different results.And the United States Department of Justiceaccepted that and through their National Institute of Justicefunded six more experiments in policing domestic violence thatessentially repeated the Minneapolis design.

    • 05:23

      LARRY SHERMAN [continued]: And what those experiments found was something quite different.There were larger samples, we could look at more factors,and, in fact, the evidence from three of those experiments veryclearly showed that arrests do deter men who have jobs.But when it comes to men who don't

    • 05:43

      LARRY SHERMAN [continued]: have jobs, unemployed men, arrest actuallymakes them more violent.And even in neighborhoods of low unemployment,arrest was effective.But in neighborhoods of high unemployment,arrest simply caused more violenceagainst women and other victims of domestic violence.In this graph from Miami, Florida,

    • 06:07

      LARRY SHERMAN [continued]: where all of the partners had to be marriedunder state law for the police to make an arrestwithout having seen the offense, even though itwas a misdemeanor, they were giventhat power for married couples.But if the male or the offender was unemployed in the Miamiexperiment, the effective arrest was to more than double

    • 06:31

      LARRY SHERMAN [continued]: the risk of repeat offending compared to not being arrested.This is in stark contrast to the offenderswho had a job where, if they were arrested,it cut their repeat offending rate in half.And then when you go to three different cities, the Milwaukeeexperiment where most of the suspects were African American,

    • 06:56

      LARRY SHERMAN [continued]: you see the same pattern as we saw in Miamiwhere most of the suspects were of Hispanic origin.And in Omaha, Nebraska, where most of the suspectswere white working-class, we had, again, the same pattern--arrest increases violence among unemployed suspects, arrest

    • 07:16

      LARRY SHERMAN [continued]: reduces violence among those people who have a job.Well, this was not very welcome news to policymakers.The idea of making arrest decisions conditionedon people having a job seemed through a gut instinctto be somehow wrong.And the idea of punishing people because they had a job

    • 07:36

      LARRY SHERMAN [continued]: made it even worse, in a way.So essentially what happened with these resultswas that they were ignored because 28 states had alreadymade it possible, and indeed mandatory,for the police to make arrests in these caseswhere they had the evidence sufficient to make the arrest.And those laws were not repealed despite this new research

    • 07:60

      LARRY SHERMAN [continued]: finding from three different cities.And part of the problem is that the over-generalizationfrom minor struggles between partners to homicidewas accepted as the best reason to make the arrest,that you'd somehow be preventing homicide.

    • 08:21

      LARRY SHERMAN [continued]: But that was a hypothesis that had neverbeen tested, at least not until 2014when Heather Harris at the University of California,Berkeley and all I did a 23 year follow-up of the Milwaukeecases.Over 1,100 victims were tracked through the death recordsof Wisconsin where we were able to see that if the offender had

    • 08:44

      LARRY SHERMAN [continued]: been handcuffed by random assignment in frontof neighbors, very often put in jail for two to 72 hours,and with 95% of them never being prosecuted,that the long-term effect of that on the victimswas actually to increase their mortality.That is, with the police in this case

    • 09:06

      LARRY SHERMAN [continued]: doing random assignment, by phoning a police headquartersand getting the way that we deal with the caseto be assigned by equal probability,there's no difference between these victimsexcept what the police did.And what the police did had no effecton whether the victim would be murdered,because only three out of 1,135 victims were murdered.

    • 09:30

      LARRY SHERMAN [continued]: And those three-- fortunately the only three--those three had no difference in themabout whether their partner had been arrested or simply givena warning.If we look at the overall pattern of death,9.28% of the victims whose partners had been arrestedhad died after 23 years.

    • 09:52

      LARRY SHERMAN [continued]: They were an average age of 30 at the timethe police experiment was done.And in that time period, they had become 64%more likely to die than those victims whose partners hadsimply been given a warning.If we look at whether this was somethingthat changed over time, we see the timeto death for victims analysis showed

    • 10:13

      LARRY SHERMAN [continued]: it was always more deadly to the victimsto have their partners arrested than to have the offender bewarned.And if we look at the heart disease deaths,we get some insight as to what might be killing them,because there was a 134% higher death rate from heart diseaseif the offender had been arrested

    • 10:34

      LARRY SHERMAN [continued]: than if the offender had been warned.But the biggest difference was by race.We see that for the black victims, if their suspect wasarrested, they were twice as likely to die-- 9.8% comparedto 5% of the victims died.When the suspect was warned, they onlyhad the 5% death rate, but the arrest

    • 10:54

      LARRY SHERMAN [continued]: led to almost a 10% death rate.For white victims, the differencewas in the same direction but it was much smaller.So in some sense, this is very much a race-specific effect tohighly concentrated African American poverty areasin one of the most highly segregated cities in the UnitedStates, which was Milwaukee at that time.

    • 11:16

      LARRY SHERMAN [continued]: If we take it further and look at whether victims themselveshad a job, what we see is a terrible tragedyof employed victims whose partnersare arrested having the highest deathrate of any of these subgroups-- four times as much likelihoodof their dying if their partners hadbeen arrested than if their partners were only warned.

    • 11:37

      LARRY SHERMAN [continued]: The big question is why this is true.And we don't really have an answerexcept that we can identify the cause as the police decisionto arrest, which is now something the police haveno option to do in the state of Wisconsin,because the state law requires them to make arrestsin these circumstances.And the state law was not reviewed in response

    • 11:59

      LARRY SHERMAN [continued]: to these findings, which came out in early 2014.[Politically Popular Research and Public Policy]The reason the laws were not changed in responseto the new research evidence on victims dying after arrestsof their partners for domestic violenceis similar to the reason that the medical practices

    • 12:23

      LARRY SHERMAN [continued]: in this country still don't align with the research showingthat doctors should wash their hands when they gofrom one patient to the next.In a recent estimate is that 100,000 people a yearin the United States die from infections that are causedby poor hand washing practices.And the ability to change that with research is just

    • 12:45

      LARRY SHERMAN [continued]: as limited as the ability to change domestic violencepolicy based on research, which this case study has told youabout in which one experiment produced a politically popularresult. Further experiments producedpolitically unpopular results.And what we see is that the law is stilldriven by the politically popular evidence, which

    • 13:07

      LARRY SHERMAN [continued]: is frequently cited, and not the politically unpopular evidence,which is much stronger and which gets ignored.A society that wants to make policyon popularity and emotions is notgoing to be as well serving its victims of domestic violenceas a society that is willing to face the facts.[Conclusion]

    • 13:29

      LARRY SHERMAN [continued]: In this case study, students shouldhave learned that the Minneapolis domestic violenceexperiment showed that arrests reduced review offendingagainst domestic violence victims,and that led to laws being changed all over the US.But it also led to more experiments,which failed to confirm my initial finding in Minneapolis

    • 13:50

      LARRY SHERMAN [continued]: and show that it was a much more complicated picture in whicharrest was often backfiring.And therefore, the laws requiring the policeto make arrests in every case should be at least reviewed,if not repealed.And that hasn't happened.Those are the things that studentsshould learn about how research has developed

    • 14:12

      LARRY SHERMAN [continued]: in relation to public policy and howthat is no different from many other examples of evenmedical research, which gets ignored by the people whoshould change their practices based on the research evidence.[Further Reading]If you're interested in reading more

    • 14:33

      LARRY SHERMAN [continued]: about this area of research, you canread my book called Policing Domestic Violence-- Experimentsand Dilemmas, published in 1992.It's available on Google Books.You can also read the American Sociological Review articlein 1984 on the original Minneapolis experimentby Lawrence Sherman and Richard Berk.You can read the 1992 replication in Milwaukee

    • 14:54

      LARRY SHERMAN [continued]: by Lawrence Sherman and Douglas Smithin the American Sociological Review.And in the Journal of Experimental Criminologyin 2014, you can read the 23 year follow upthat Heather Harris and I did showingthat the victims whose partners were arrestedwere more likely to die.

History of Work on Domestic Violence

View Segments Segment :


Professor Larry Sherman discusses his research on domestic violence and the effect that arresting the perpetrator has on repeat offending. A study done in Minneapolis in 1981 showed that arresting domestic abusers significantly reduced the likelihood of repeat offense. But further research revealed contradictory results, but these findings have been ignored by lawmakers.

SAGE Video Cases
History of Work on Domestic Violence

Professor Larry Sherman discusses his research on domestic violence and the effect that arresting the perpetrator has on repeat offending. A study done in Minneapolis in 1981 showed that arresting domestic abusers significantly reduced the likelihood of repeat offense. But further research revealed contradictory results, but these findings have been ignored by lawmakers.

Copy and paste the following HTML into your website

Back to Top