MENU

Reason.com

Free Minds & Free Markets

VOLOKH CONSPIRACY

Mostly law professors, blogging on whatever we please since 2002 · Hosted by The Washington Post, 2014-2017 · Hosted by Reason 2017 · Sometimes contrarian · Often libertarian · Always independent

Why the Supreme Court Reviews So Many Qualified Immunity Cases

Here's a theory, at least.

My friend and co-blogger Will Baude asks an interesting question: What explains the Supreme Court's continued interest in reviewing seemingly fact-bound lower court decisions on qualified immunity? Will writes:

I have criticized the Court's qualified immunity doctrine at length, but I do understand that one might disagree, especially if one believes in evolving judge-made law (see this draft response from Hillel Levin and Mike Wells) or might think the issue so settled by stare decisis that my critiques are merely academic. Still, it is worth noting that the Court treats qualified immunity not just as ordinary settled law, but as an area of law so important that it is worth deciding a series of factbound cases that would never earn the Court's attention if they involved a different legal issue. Moreover, the Court seems uninterested or unable to find such cases where a lower court wrongly denied relief to a person whose constitutional rights were violated.

I remain unconvinced that this special legal treatment has a good legal basis.

I've wondered about this, too, and I wanted to offer a possible explanation. I apologize that my explanation is long, tentative, and perhaps a bit rambling. I hope there's at least the kernel of a useful idea in here, but of course you'll be the judge of that.

First, some background. It seems to me that the Supreme Court uses qualified immunity to create what Meir Dan-Cohen famously called "acoustic separation" between conduct rules and decision rules. Here's the idea, at least as applied to criminal investigations. Ex ante, there may be some ideal rule that you would want the police to follow. That's the conduct rule, in this case the constitutional standard. But ex post, when the police don't follow the rules, the less-than-ideal remedies available to deter misconduct may support a less restrictive rule for when that remedy is actually imposed for violations. That's the decision rule.

Of course, we can debate whether remedies really do cause harms that justify a separate decision rule. But the cases suggest that the Justices believe they do. Less-than-ideal penalties may incentivize shirking, the Court tells us, when an officer faces personal liability. And in the criminal setting, evidence lost to the exclusionary rule imposes the public harm of leaving crimes without successful prosecutions.

Under this set of assumptions, the ideal system might be one with so-called "acoustic separation" between conduct rules and decision rules. In plain English, the best system may be one in which the police think the conduct rule governs and aim to follow it, but in which the courts actually only impose liability ex post under more limited decision rules.

I suspect that's what the Justices are trying to get from qualified immunity doctrine. Constitutional doctrine is the conduct rule, and qualified immunity is the decision rule. Qualified immunity creates a buffer between the conduct rule and actual liability. Because qualified immunity uses a very lawyerly deference standard -- whether an officer violated clearly established law that a reasonable person would have known -- an officer in the field can't really know where the line is and can just aim for the conduct rule. Ideally, at least, the lawyerly and vague nature of the qualified immunity standard creates the needed acoustic separation. But because the standard is quite deferential under the Supreme Court's approach, remedies are imposed ex post only in pretty rare and extreme cases. If the system works as the Justices imagine -- a huge "if," obviously, but run with it for now -- you may end up with the needed deterrence (from the conduct rule announced ex ante) with less of other harms imposed (from the decision rule applied ex post).

Now you're wondering: How does any of this explain the Supreme Court's interest in reviewing so many qualified immunity cases? Well, I'm not totally sure. But here's a theory for how it might explain it: Because the buffer of qualified immunity is the same general standard in every context, misapplying the buffer in one specific case can effectively alter decision rules in a way that creates the functional equivalent of a circuit split. Put another way, misapplying qualified immunity doctrine might look like a fact-bound case but operate like an important circuit split. And that might plausibly explain Supreme Court review in what otherwise looks like fact-bound cases.

Here's my thinking. The conduct rules of constitutional law are usually pretty specific rules that apply general constitutional principles. In the Fourth Amendment setting, I have called this the difference between the principles layer of the doctrine and the application layer of the doctrine. The principles layer might be that a search occurs when the government violates a reasonable expectation of privacy (a general principle). The application layer might be that using a thermal imaging device directed at a home is a search or that using a drug sniffing dog during a traffic stop is not a search (applications of the general principle to create specific rules).

Circuit splits that justify Supreme Court intervention are almost always at the application layer of the doctrine. That is, lower courts will split on the correct rule that emerges from the general principle in a very specific situation. It's not like circuits will disagree about whether to follow the reasonable expectation of privacy test. Lower courts instead disagree on a specific application of the broader principles, such as whether surreptitiously installing a GPS device on a car is a search or whether collecting historical cell-site records from a cell phone provider is a search.

Qualified immunity works differently. It's one general and exceedingly fact-specific standard -- whether an officer violated clearly established law that a reasonable person would have known -- that isn't subject to neat rules. It's a particular thickness of buffer on top of the constitutional conduct rule, but it's not of a thickness that is reducible to a bright line. You have to read some qualified immunity cases to get a feel for how thick a buffer it is beyond the constitutional standard. And that thickness applies all around in every context. The thickness of the buffer that is qualified immunity -- that separates the conduct rule from the decision rule -- is the same general standard regardless of what underling conduct rule at the application layer is being considered.

The takeaway is that misapplying the decision rule of qualified immunity could have quite broad effects. What looks like a fact-specific case could have much broader implications. First, a decision applying qualified immunity in a particular way is not just reaching a fact-bound application: It is effectively creating a rule about the thickness of the qualified immunity buffer. And second, because the thickness of the qualified immunity buffer is understood to be the same in all contexts, a precedent that wrongly narrows the thickness of the buffer in one context could apply in all contexts. That makes a decision applying qualified immunity to a particular set of facts have potentially much greater significance.

Perhaps an example with some numbers might prove the point. Say the Supreme Court regulates different law enforcement practices on a scale of 1 to 10, with 1 meaning no limits imposed on the government and 10 meaning the imposition of very strict limits on the government. (These are totally artificial numbers, obviously. Consider this just a proof of concept, not a hypothetical based on empirical evidence). Say the Court concludes that the conduct rule of an arrest should be set at a 6.0, wiretapping should be an 8.0, a Terry stop should be a 3.0, and an excessive use of force should be a 4.0. The Court further wants the buffer of qualified immunity that creates the best decision rule to be a 2.0. That is, for each government practice, the actual point of imposing a remedy should come when the government's violation was two points on the scale less than the conduct rule ideal point. A remedy for an illegal arrest should come when the government was below a 4.0; a remedy for an unlawfully excessive use of force should be imposed only when the government dropped below 2.0, and the like.

Now imagine the Ninth Circuit has a case on excessive force. Let's say a plaintiff sues a police officer claiming that the officer's use of force against the plaintiff was unconstitutional because, he claims, it was below the permitted 4.0 standard of the constitutional rule. Let's also say the Ninth Circuit correctly rules that the facts of the use of force, based on the totality of the circumstances, measure a 3.0. It therefore holds that the officer's use of force was excessive and violated the Fourth Amendment. So far, so good.

Now we get to the important part. Say the Ninth Circuit misapplies the qualified immunity standard. It applies that general standard in a particularly narrow way as if the buffer of qualified immunity were only 0.5 points thick instead of 2.0 points thick. As a result, the Ninth Circuit holds the officer personally liable. First, the the officer violated the constiution, as the facts were a 3.0 which is below the 4.0 constitutional conduct rule. And second, the buffer wasn't big enough to avoid liability under qualified immunity because the buffer was applied to be only 0.5. Because the facts were a 3.0, and the buffer was only 0.5, the officer acted below the threshold of liability and is held personally liable.

As I said, this is obviously a totally artificial and highly stylized hypothetical. But I hope it shows why misapplying the qualified immunity standard might have far greater systematic implications than merely misapplying the Fourth Amendment conduct rule. On one hand, whether the court properly applied the conduct rule has no far-reaching implications. In this case, the Ninth Circuit held that the use of force on the facts of this case was a 3.0, which was below the constitutional threshold. That happened to be correct in my hypothetical. But whether it was right or wrong, in a future case that holding would have only limited significance. The constititutional conduct rule for an excessive use of force is a 4.0 no matter how the Ninth Circuit rules in this particular case. A very fact-specific application of the Fourth Amendment ruling that a particular set of facts was a 3.0 isn't likely to matter very much.

On other other hand, whether the Ninth Circuit properly applied the qualified immunity standard could have more far reaching implications akin to creating a circuit split. Later precedents could look to that Ninth Circuit decision for guidance on the thickness of the qualified immunity buffer. If this case treated the buffer as a 0.5 instead of the 2.0 buffer that the Supreme Court intended, future lower courts can follow the circuit decision treating the buffer as only a 0.5. The practical effect would be to shift the decision rule for an excessive use of force from a 2.0 (the 4.0 conduct rule minus the 2.0 buffer) to a 3.5 (the 4.0 conduct rule minus a 0.5 buffer). What looked like a fact-bound ruling on qualified immunity would have the effect of a precedential rule on the thickness of the buffer between the conduct rule and decision rule. It's the functional equivalent of rejecting the Supreme Court's intended decision rule.

Further, because the thickness of the qualified immunity buffer is presumed to be the same in all contexts-- remember, it's just one general standard -- a ruling treating the buffer as only a 0.5 in an excessive force case provides a precedent for moving the decision rule in every case. The decision rule for an arrest could move from 4.0 to 5.5, for wiretapping from 6.0 to 7.5, for a Terry stop from 1.0 to 2.5, etc. A ruling on the decision rule could have really far-reaching impact because the decision rule, unlike the conduct rules, are assumed to be universal.

Now you might reasonably believe that the Supreme Court's approach is wrong and see a narrow application of the qualified immunnity doctrine as a good thing. You might conclude the existing conduct rules should be at higher numbers, and moving the decision rules might induce the police to act as if they are. Or maybe you think the gap between the conduct rule and decision rule should be thinner, such that slimming it down is for the public good. There are important and signficant arguments for both of these positions. But the key for us is that such judgements are about resetting the conduct rules and decision rules of criminal procedure law, not just making fact-bound judgements in a particular case. If you're a Supreme Court Justice who thinks that the existing conduct rules and decision rules are more or less correct, you might see lower-court efforts to thwart those rules as something deserving your pretty close attention.

Anyway, that's a possible explanation for the Supreme Court's continued interest in qualified immunity cases. I'm pretty tentative on the explanation myself, so I don't claim to be certain that I'm right. But I thought the argument worth developing in case others find it interesting.

Alternatively, maybe the Supreme Court just really likes the police.

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of Reason.com or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  • Stephen Lathrop||

    I can't figure out whether I'm too thick to understand this, or not thick enough.

  • Orin Kerr||

    Probably the latter. Fraud!!!

  • David Nieporent||

    Are you asking for our opinions?

  • Brett Bellmore||

    It sounds to me like you're saying that the Supreme court wants to have a very lax standard, "Police can get away with almost anything.", to avoid the costs of punishing police, while having people think the standard is much stricter, "Police will be punished for wrongful conduct!" in the hope police will be deterred by it. And that they're taking a lot of QI cases because maintaining this disconnect (Your 'acoustic separator'.) between applied and perceived rules requires a lot of work. They're trying to have their cake, and eat it, too.

    Doesn't make much sense, because the court rulings on QI are public, the police are aware of them, as they have a very strong incentive to be aware of what the real rules they're operating under are. Deterrence without actual punishment isn't a real thing. Police are responding to the real system, not the rhetorical system. You eat your cake, you don't have it anymore.

    And the Justices, (Very ironic title in this context!) are not idiots, they must know that.

    Cont.

  • Brett Bellmore||

    I think there's another way to describe the situation. The Supreme court wants a system under which police can do almost anything without being punished. At the same time they don't want to admit that they want the police to be able to abuse the public without consequence, because they're aware that the public would (IS!) be outraged by it.

    So they propound rules which ought to result in police being punished for misconduct, and then keep not actually applying those rules in individual cases. But still announce the rules anyway.

    And some lower courts take the rules the Supreme court is explicitly propounding, and try to actually apply them, instead of the actual rules the Court means for them to apply, but doesn't dare publicly announce.

    So the Court takes the appeal to spare the cop, and then propounds the fake rule yet again.

    And all levels, the individual cops, the lower courts, the Supreme court, are aware of what is going on here. The Court is lying about the actual rule to try to avoid public outrage, and some of the lower courts are taking the lie seriously and daring the Court to publicly not make it real.

    It's a covert circuit split that the Court can't resolve because it's preferred resolution would be massively unpopular.

  • nonzenze||

    That's not quite fair, In the abstract, I think it's correct to say that in ideal world:

    (1) Each officer earnestly and truly believes that he will be liable if he violates the law.

    (2) When an officer violates the law despite (1), that it is actually judged leniently.

    In such a world, you'd have all the deterrence and none of the side effects of remedies.

    Now, in my opinion you cannot have both of them because human behavior isn't driven by what is said but rather by the actual incentives in place. That is, you cannot maintain deterrence without applying the penalties or else behavior will drift until it settles in balance with the actual deterrence.

    In other words, I assert that conduct equilibrates to the level dictated by the decision rule. This is a legal version of TANSTAAFL :-)

  • Martinned||

    Why would we want to judge police officers particularly leniently? If anyone should know the law it should be them. Lead by example, cleaner than clean, let he who is without sin cast the first stone, etc.

  • Lee Moore||

    The police have a job to arrrest unauthoried stone casters. If they don't get the benefit of the doubt, they'll just sit on their hands. So while "cleaner than clean" is a perfectly good reason to throw the book at a police officer who's broken the law, clearly, plainly, unarguably and unreasonably (ie he should have known) you certainly don't want police officers to apply a "beyond a reasonable doubt" standard in going about their policing interventions. Their job (and the job of soldiers) inevitably includes a lot of fog and uncertainty. If you don't cut them a lot of slack they'll just stay inside drinking coffee and munching on donuts. Go count the bodies in Chicago.

    The bigger mystery is why prosecutors get the same, or even more, slack. They work in nice warm offices, with time to think things through. They're not out in the kitchen knife waving community risking their necks.

    The trouble with lawyers is that they imagine everybody else has a couple of years, and multiple redrafts, to get their act together.

  • Martinned||

    I think you got that "beyond a reasonable doubt" standard backwards. The police can and should get the benefit of the doubt if they're on trial. And so should everybody else.

    And let's please not compare cops to soldiers. It's bad enough that many US police forces are armed as if they're patrolling downtown Basra. Let's not adjust the law to that assumption as well...

  • nonzenze||

    I think you took the soldier comment in bad faith. The commonality wasn't in militarization, it was in the condition of working under irreducible uncertainty.

  • Lee Moore||

    I think you got that "beyond a reasonable doubt" standard backwards.

    I don't believe I did. The point is whether we think it would be desirable if police officers typically asked themselves "Am I convinced beyond a reasonable doubt that I could not be sued, or subjected to disciplinary action potentially costing me my job and my pension, if I intervened here." This standard would, IMHO, relieve the sufferings of people unjustly interfered with by the police by roughly one hundredth of the corresponding increase the sufferings of crime victims. Which would be a poor bargain.

    The police can and should get the benefit of the doubt if they're on trial.

    The law has other slings and arrows besides criminal trial. Qualified immunity is a defense in a civil action.

  • FlameCCT||

    It is unfortunate that most gangs are as well armed as the insurgents in Raqqa, Basra, etc.

  • SykesFive||

    The claims against police officers are usually civil in nature and don't implicate the "beyond a reasonable doubt" standard.

  • RoyMo||

    Soldiers seem to be judged more harshly than police. They are subject to discipline and court martial, they don't have qualified immunity and have higher standards. For example there is no equivalent of article 99 for police officers, and no criminal punishments for dereliction of duty and cowardice.

  • NToJ||

    "If you don't cut them a lot of slack they'll just stay inside drinking coffee and munching on donuts."

    Then we just fire them? I don't think America's unique qualified immunity doctrine is the only way to police crimes in the world.

  • TwelveInchPianist||

    "Then we just fire them?"

    That's not much of a disincentive either, given that they'll just be re-instated with back pay a few months later.

  • David Nieporent||

    The police have a job to arrrest unauthoried stone casters. If they don't get the benefit of the doubt, they'll just sit on their hands.

    This overdeterrence argument doesn't apply to any other profession -- not doctors, lawyers, accountants, drivers, pilots, plumbers, etc. Why are LEOs so unique that they won't do their job if they're not allowed to be negligent?

    Go count the bodies in Chicago.

    Yes; we need to restore QI in Chicago so that police will do their jobs there.

  • Lee Moore||

    This overdeterrence argument doesn't apply to any other profession -- not doctors, lawyers, accountants, drivers, pilots, plumbers, etc. Why are LEOs so unique that they won't do their job if they're not allowed to be negligent?

    That's odd. I always thought the sort of lawyers who could do you serious damage - judges and prosecutors - did get immunity. Must have been abolished when I wasn't paying attention. Even so, as nonzenze elegantly put it, police officers and soldiers operate in conditions of "irreducible uncertainty" - and often in conditions requiring snap decisions, in the presence of deadly weapons. So even if judges and prosecutors have recently lost their immunity, there are still objective reasons for police officers and soldiers to get a wide margin of error.

    Yes; we need to restore QI in Chicago so that police will do their jobs there.

    You don't need to abolish QI to get the police to sit on ther hands. There are other ways. You can enter into a consent decree undertaking to ensure that the police sit on their hands. You could stop paying the police. There's all sorts of ways if you use your imagination and if sitting on their hands is what you want the police to do. But however you achieve it, you're going to reap a rich harvest of robberies, rapes and corpses.

  • David Nieporent||

    Even so, as nonzenze elegantly put it, police officers and soldiers operate in conditions of "irreducible uncertainty" - and often in conditions requiring snap decisions,

    So do surgeons. But they don't get immunity.

    (Remember, immunity is not simply a "wide margin of error." The issue is not whether a particular police act should be found to be reasonable or not. The issue is whether that question should even be asked.)

  • nonzenze||

    Martinned, if you accept point #1 (and note, that I don't!) then the police have already acted as if they earnestly and truly believe they will be held personally liable. So there is no further goal to be accomplished.

    Meanwhile, remedies are actively harmful along some dimensions: if you exclude unlawfully obtained evidence then a criminal goes free to victimize another innocent. If you award damages, then that money comes from somewhere where it could have been better spent. If you disincentivize police from policing, you would get less policing that the democratic process would have selected without the disincentive.

    You wrote in a previous thread that there are only tradeoffs. This is putatively a way to cheat the tradeoff -- by having both belief in a strict system and the facts of a lenient one. I don't believe you can do it, but I understand the theoretical basis for believing it would be better to cheat it.

  • Martinned||

    For the record, I think the exclusionary rule is wrong. I think the consequence of having unlawfully obtained evidence should be a discount on the ultimate sentence. (Unless it's really serious, in which case the case should be thrown out.)

    And in general, I don't think the distinction between conduct rules and decision rules. Ultimately, the cops will be advised on what not to do based on the decision rules and nothing else.

    (Then again, I've been re-watching some old NYPD Blue, so I might just be a bit more pessimistic about the police than usual. Great show, but wall-to-wall rights violations in every episode.)

  • Brett Bellmore||

    The exclusionary rule was adopted, so I understand, because the judiciary had to give some effect to the 4th and 5th amendments, but didn't see punishing police for violating them as a realistic option. So, you try to make the victim whole by barring use of the unconstitutionally obtained evidence.

    The problem with it being that this results in people you can see are guilty getting off, which the system really doesn't want to accept.

    "Ultimately, the cops will be advised on what not to do based on the decision rules and nothing else."

    Exactly: There are no conduct "rules", only decision rules. Nobody cares about rules that aren't enforced, and only the decision rule gets enforced.

  • NToJ||

    It seems to me that if you eliminated qualified immunity, you could do away with the exclusionary rule.

  • Joe_JP||

    Part of the point of the exclusionary rule is to avoid judges being complicit in the violation of constitutional rights -- upholding the use of wrongly obtained evidence taints them too and they have an important rule in the process respecting handing down warrants particularly.

  • nonzenze||

    I agree, since it was in fact my point that cops will (not consciously) behave as the decision rules no matter the attempt to set the conduct rule.

    Either way, I think it's clear there are tradeoffs, no?

  • nonzenze||

    To add: I agree that this division is essentially dishonest. I guess I take issue with the assertion that this dishonesty is actually bad. If you could actually make the acoustic separation work (for reasons documented above, I think it's not possible) then the dishonesty would be a fantastic thing.

    This harkens in some way to the idea I read somewhere that some strategies can never work if they are declared openly . . .

  • Brett Bellmore||

    Or just deducible from public information.

  • nonzenze||

    I don't consider that deducible is the material difference. I think it's rather that incentives 'naturally' create the situation then incentivize irrespective of the declared truth.

    The dishonesty is unraveled not by conscious action, but simple by virtue of not being enforced.

    [ Note: there are dozens more examples I thought up this morning of situations it's indisputably better (even morally) if people would believe that they operate in a very strict regime even if that wasn't the case. If the goal of the strictness is to influence people's action then the very fact that they believe accomplishes this! ]

  • Peter Gerdes||

    I think you too are offering an overcomplex explanation. It's not very plausible to think the supreme court is so concerned about bad PR that it's engaged in some complex dodge convince the public of one standard while applying another.

    The best, and simplest, explanation in my view is just that the supreme court announces what they actually think the correct constitutional standard was and then said 'But it doesn't seem fair to punish officers who didn't parse the law the way we did,' especially in cases where there was genuine judicial disagreement. So qualified immunity was born out of a reasonable concern for justice even though it really should have been left to the legislature to amend the law to immunize officers in cases precedent wasn't clear.

    I don't think we need to appeal to a complex theory to explain the current situation. I think some members of the court really do feel passionately about protecting police officers and sympathize with the fact that their job means a momentary bad choice can open them up to personal liability. As long as no one on the court feels equally passionately about pushing in the other direction and the rest follow precedent we get the pattern of grants and reversals that we see.

  • Martinned||

    Sure, but that doesn't answer the original question: Why so many summary reversals?

  • Peter Gerdes||

    Because 4 members of the court care enough about denials of QI they are willing to vote to grant cert in those cases and enough other members of the court look at precedent and say 'Sure, the appellate decision was wrong.'

    Since there aren't 4 justices who are similarly passionately voting to grant cert when the appeals court grants QI we get the imbalance.

  • Peter Gerdes||

    I don't think we need any extra theory other than some members of the court are think its really important that police be protected by QI and see failing to do so as posing a greater threat than merely getting a single verdict wrong.

    I mean one can even see why they might think that. If they believe that doing QI wrong risks undermining police enforcement generally but that getting the law wrong in a particular case just affects those people (though those seem like concerns about consequence that IMO should probably not play such a prominent role).

  • nonzenze||

    Peter, if this were the case then they would strictly enforce the 'two step' marching order of first making a legal finding about whether the right was violated and only then going to the 'clear precedent' prong.

  • Peter Gerdes||

    Hmm, that is a good point.

  • tkamenick||

    " At the same time they don't want to admit that they want the police to be able to abuse the public without consequence, because they're aware that the public would (IS!) be outraged by it."

    That's the rub. If Kerr is right, this is actually what the court is doing. They say that the threshold is X, but in practice enforce a threshold of X+Y. As nonzenze hints at, you can't have both worlds. The rule is what is enforced, not what is publicly said.

  • bernard11||

    I think you are getting at what I see as a logical issue here.

    If I understand this idea correctly, and I am definitely not sure I do, the police are told, "you cannot go below 4.0, say, in this type of situation," but the court will punish them only if they go below 2.0.

    I don't know if that's what's going on, but it sounds remarkably silly if it is. Police officers are not deaf, blind, and illiterate. Once it becomes clear, as it readily will, that 3.0 or 2.5 is really acceptable the so-called "conduct rule" is washed away.

  • Orin Kerr||

    Brett writes: "Doesn't make much sense, because the court rulings on QI are public, the police are aware of them, as they have a very strong incentive to be aware of what the real rules they're operating under are."

    The police are very rarely sued, though, so it's not like they tend to experience those rules. And figuring out what the standards are is pretty tricky for lawyers -- much less police officers -- to learn. Take the question of when an officer can be sued for serving a warrant not based on probable cause. You can teach an officer what probable cause is based on Supreme Court cases. But how do you teach them where the qualified immunity line is for a civil suit based on lack of probable cause? That's considerably harder, I think. My two cents, at least.

  • Brett Bellmore||

    I think the problem here is that you want to believe the Court is doing something complicated but honest, when they're really doing something simple and dishonest.

    There is no conduct rule. There is only the decision rule. Because a rule that doesn't actually get enforced is no rule at all, and only the decision rule is enforced. And the police are aware of this.

    And the Supreme court's decision rule is that cops can get away with basically anything.

    But the public would hate them for openly propounding a rule like that, hate them so much that they might face consequences even on their high and mighty thrones, so they lie about what the rule they're applying is. And count on the lower courts to pick up on what they're doing, and play the same game of pretending that the cops have actual, enforceable limits on what they can do, while in practice treating them as untouchable.

    Only some of the lower court refuse to play along, and try enforcing the lie instead of the Court's actual rule. And the Court can't publicly rebuke them, because it would expose the lie. So the Court just keeps taking the cases and reversing the rulings without clarifying the rule, because there isn't really any way to clarify it, it's a lie.

  • Michael P||

    For what little my lay opinion is worth, I think Brett's model holds together a lot better.

  • Martinned||

    Likewise. (On both parts of the comment.)

  • Orin Kerr||

    Brett, your explanation of how the Justices and the police think conflicts with my experience with the Justices and my conversations with a lot of police officers. But I realize this is the Internet, so

  • BillyG||

    your explanation of how the Justices and the police think conflicts with my experience with the Justices and my conversations with a lot of police officers

    I'd have a lot more faith in your explanation if I saw fewer cases granting QI to outrageous conduct like the one from this short circuit where the cop pointed a loaded gun at the head of a suspect and threatened to murder him.

    Unlike what you said above, the requirement is not whether an officer violated clearly established law that a reasonable person would have known, it is whether the courts have previously previously ruled a behavior by a police officer violates the law. It's pretty clear most of the behavior we've seen in QI cases would get a non-cop arrested, prosecuted, and jailed in short order. With the Iowa woman from the QI granted here, what would have been the ruling on justifiable use of self-defense on behalf of another if the cop were instead a civilian? My beliefe is it would have been murder charges by the DA.

  • Peter Gerdes||

    But I think it's fair to say that we want to discourage certain conduct by average members of the population that we don't want to discourage from cops. In particular, we want to generally encourage cops to get involved in potentially dangerous situations and confront armed individuals who pose a potential danger to others. In contrast, we want to encourage average civilians to avoid such confrontations, leave the scene and call the cops rather than counting on their ability to exercise their right to self-defense should things go south.

    Don't get me wrong. I think the QI precedent is a travesty and the supreme court has no business trying to strike the balance in such cases. That's clearly a legislative responsibility.

    However, I don't think we want exactly the same standard for police use of force as we do for civilian use of force because of the point mentioned above. However, I don't see any benefit to offering immunity in cases where no reasonable person would think that was an acceptable way to act just because it isn't clearly established. But in split second calls that might be reasonable mistakes I think the legislature might reasonably offer a different, more accommodating, standard for police than for normal civilians.

  • Martinned||

    Nobody is proposing having the same standard. The law only imposes liability for violation of constitutional rights, not for everything that would be a tort if done by a member of the public.

  • Peter Gerdes||

    I was responding to what I took to be the implied point of BillyG's last paragraph. The fact that it would get a non-cop jailed isn't really relevant since we might not want the same standard in both cases.

    Personally I think just saying liability for a violation of constitutional rights is a bad law. It really should distinguish between not reading someone their Miranda rights before the Miranda decision and beating them after handcuffing them. However, I think it should have been up to the legislature not the courts to fix that oversight.

  • Rossami||

    I don't think that's fair to say at all. First, yes we do want cops to get involved in potentially dangerous situations. That's why we pay them. That's also why they are afforded a relatively high status in the community. Second, however, we also want strangers to bring their own talents and capabilities to bear on dangerous situations. That's why we have good samaritan laws and afford very high status to those people who put themselves in harms way without being paid for it.

    Remember that on a statistical basis, many "average" members of the population are not at all average. In any given crowd, there will be veterans, paramedics and other off duty and/or retired specialists whose ability to react to the potentially dangerous situation is at least as good as a responding cop's.

    In fact, the only people who want all those "average civilians" to run for cover are the cops themselves. They want to "control" the situation without regard for whether that control will actually lead to the best outcome.

    I agree that QI is a travesty and it's long past time to abolish it. I do not agree that police should get special treatment. Civilians sometimes have to make split-second life-or-death decisions, too. If police should be granted some leeway in those split-second decisions, so should similarly situated civilians. If civilians don't deserve that leeway, neither do the police.

  • Peter Gerdes||

    Do you really believe that police should be subject to the duty to retreat in the same way that any other civilian would be? So if the police show up to a robbery and the suspect has a gun and says "Leave or I'll shot you" they have to just pack up and leave? Surely you wouldn't go that far.

    While you make some pretty good points that maybe we should expect more involvement from and intervention by other civilians I think the idea that civilians should back off from dangerous/unpredictable situations and call in the professionals to deal with it is pretty widely accepted by the public (and therefore informs how we regulate or reach judicial decisions). I suspect people like this rule because it gives them an easy answer to the question of whether they did the right thing

  • David Nieporent||

    Do you really believe that police should be subject to the duty to retreat in the same way that any other civilian would be?

    Yes. Which is to say, there ought not to be a duty to retreat.

  • Peter Gerdes||

    Do you really believe that police should be subject to the duty to retreat in the same way that any other civilian would be? So if the police show up to a robbery and the suspect has a gun and says "Leave or I'll shot you" they have to just pack up and leave? Surely you wouldn't go that far.

    While you make some pretty good points that maybe we should expect more involvement from and intervention by other civilians I think the idea that civilians should back off from dangerous/unpredictable situations and call in the professionals to deal with it is pretty widely accepted by the public (and therefore informs how we regulate or reach judicial decisions). I suspect people like this rule because it gives them an easy answer to the question of whether they did the right thing

  • Sanctimonica||

    I deal with QI on a daily basis in the civil courts. I don't see the the courts using it to excuse outrageous police conduct, at all.

    I see courts routinely deny QI, especially when facts are disputed. There is no place for QI when the evidence creates a genuine issue of material fact whether the officer violated either the "conduct rules" or the "decision rules."

    QI gets a bad rap because its critics look at cases in hindsight. How can the officer get away with shooting an unarmed man, we wonder? But how does the officer know the gun was a toy, or that the guy swinging the machete is not-really-dangerous-but-appears-to-be-so because of mental health issues, or that the thing the guy pulled from his waistband after being told to show hands five times is a shiny black cellphone rather than a shiny black Walther? Police don't have that luxury, and balancing the need (indeed, the public expectation) for law-enforcement action against the risk of error is the entire purpose of QI.

    Courts cannot employ hindsight like commentators do. (Although they often do, and perhaps that's why we see reversals). Instead, they ask whether, given the uncontroverted facts and the state of clearly established law, could the officer have believed her conduct was lawful at the moment she decided to act.?

  • David Nieporent||

    Courts cannot employ hindsight like commentators do.

    Every single tort case in every court in the country, from municipal court up to the U.S. Supreme Court, involves "hindsight." It involves a jury judging an action (or omission) after the fact to determine whether it was reasonable/justified at the time the action was undertaken/omitted. Why should police actions or omissions be any different?

    Instead, they ask whether, given the uncontroverted facts and the state of clearly established law, could the officer have believed her conduct was lawful at the moment she decided to act.?

    No. The test is objective -- what a reasonable officer would have thought -- not what the officer "could have believed."

    QI does not get a bad rap because people are looking at the case in hindsight. QI gets a bad rap because it shuts down the case before the victim has a chance to establish whether the officer was acting reasonably. Based on patently absurd reasoning about "clearly established law," like the majority in this case pretending that there needs to be a case in which a court ruled that an officer seeing a woman through a chain link fence holding a knife talking to someone can be shot.

    If I saw a woman standing in her yard holding a knife and talking to another woman, and I just ran up and shot her, no court would ever rule as a matter of law that I was acting reasonably.

  • bernard11||

    I agree with Brett that it is unrealistic to think the police are not going to know the decision rule, and be guided by that rather than the conduct rule.

    It may not be absolutely clear, but is it really that much less clear than the conduct rule? And even if it is less clear, surely the police will quickly learn that they have considerable leeway to violate the conduct rule.

    Aside from all else, a deliberately murky rule is hardly fair to police themselves. "A court let a cop in Indiana get away with this. Why do I get nailed?"

  • nonzenze||

    Indeed, the conduct rule is simpler to describe. But behavior drifts until it is corrected . . .

  • Eidde||

    Any 9th Amendment aficionados here? Any privileges and immunities aficionados?

    Do the rights retained by the people, and mentioned in the 9th Amendment, or the privileges and immunities mentioned in the 14th amendment, include rights to a judicial remedy for wrongs suffered, as in the constitutions of some states?

    If there's such a federal right, is it violated if you're denied justice for something that's admittedly illegal but the cop couldn't be expected to know was illegal?

  • Eidde||

    (I'm assuming that the people here are thought leaders who can suggest a correct course to the Supreme Court if the Supreme Court is wrong)

  • Martinned||

    Seems unlikely. The litigation we're talking about here is all on a statutory basis.

    (That's what qualified immunity is: the courts saying that Congress didn't really mean it when they said "Every person who (...) subjects (...) any (...) person (...) to the deprivation of any rights (...) shall be liable to the party injured (...)". )

  • FlameCCT||

    IOW QI is the extension of Congress not having to abide by the laws they create.

  • apedad||

    Perhaps the buffer aspect should be described more as a time aspect.

    The less time an officer has to make a decision (e.g. a split second decision--is the small black object in the suspect's hand a gun or a cell phone), the more QI.

    And the more time an officer has to make a decision (e.g. should we put this tracker on the suspect's vehicle), the less QI.

  • Martinned||

    That seems like an excellent way to put it to a jury. Not sure why such issues should occupy the Supreme Court, though.

  • nonzenze||

    It's already part of the standard. The reasonableness test is specifically as-observed, not given infinite reflection.

  • Rossami||

    That still doesn't solve the problem of the officer who willfully puts him/herself into the situation where a split-second decision is required.

  • FlameCCT||

    Isn't it the nature of law enforcement, similar to firefighters and others; to put oneself into situations where split-second decisions are required?

  • Dave W.||

    Lot of Lily gilding in this post. For the most part, the fourth amendment is much more important for poor people then rich people, at least in the sort of cases that lead to qualified immunity decisions. Think, for example, how the police treated John Ramsey. Qualified immunity is a subtle form of class Warfare. Anyway, good to see Professor Kerr finally blogging about qualified immunity. Too many Fourth Amendment Scholars have failed to pay sufficient attention to the issue for too long, in my humble opinion.

  • Martinned||

    Alternatively, maybe the Supreme Court just really likes the police.

    Well, yes. I think that's a given. That's how we ended up doing qualified immunity in the first place.

  • Peter Gerdes||

    I'm still a bit confused as to why the thickness of the acoustic separation is any different than the application of any other standard. I mean consider the application of intermediate scrutiny. There, just as here, there is an adjustable thickness which is supposed to apply uniformly to a range of different situations and misapplication of that standard to any particular case could be interpreted as changing the standard generally.

    Is the difference that somehow it's harder to articulate a principle specifying the size of the separation in the case of qualified immunity

  • nonzenze||

    I think the point is that in this case we explicitly want the police to believe that the separation is small but the courts to rule that it is not small, all at the same time.

  • Peter Gerdes||

    No, I understand that. However, Prof. Kerr is arguing that somehow correcting misapplications of the standard on individual fact patters is somehow more important here than it is to correct misapplications of other standards.

    In particular, there are a great many cases where a district or appellate court misapplies some other legal standard but the supreme court doesn't feel that it warrants a grant to correct that misapplication because it won't have some broader application outside of that one case. However, Prof. Kerr is arguing here that somehow denying QI is likely to have some kind of precedential style effect that we don't see in other mistaken applications of standards to a particular factual situation. I don't understand why that should be the case here.

  • Peter Gerdes||

    Well, actually to nitpick a bit I don't think the idea is that the police should believe the separation is small. Rather it's for them to be unsure about when they can count on that separation to protect them.

    For example, if a police officer is thinking about crossing the constitutional line in some particular way he can't think 'Ohh, no worries QI will protect me' because it's hard to figure out when and where QI might apply and for all they know yesterday the appellate court handed down a clear precedent covering just such a case.

  • Peter Gerdes||

    Also, wouldn't the fact that appellate courts get a particular application of QI wrong help serve the goal of incentivizing constitutional behavior while avoiding shirking.

    Presumably, for this argument to work at all, the model the justices have in mind is something like this. Officers will shirk if it seems the probability of being personally liable for stepping over the line is too high. However, we don't want officers being able to rely on getting away with unconstitutional conduct because in some circumstances that might encourage unconstitutional behavior so we will have a murky, hard to interpret gap where we don't hold them liable but they aren't able to rely on when making decisions.

    But on this model the supreme court should actually welcome a degree of variation from the lower courts as to when they hold QI applies. As long as ON AVERAGE the lower courts grant QI to every police action that is within 2.0 units of the constitutional line then officers are very rarely held personally liable for accidentally stepping across the line. However, if sometimes the appellate courts mess up and deny QI for an action that's only .5 units across and other times grant it for an action that's 2.5 units over that's good because it makes crossing the constitutional line at all somewhat dangerous.

  • Subconscious Security||

    This is an excellent point. Under this model it seems that the Supreme Court should welcome some variation for this reason. Instead, it seems as if the Supreme Court itself is attempting to create uncertainty the other way, by every so often diving in and summarily reversing a verdict against a police officer in questionable circumstances. In other words, it is as if the Supreme Court is trying to create and maintain uncertainty that makes it difficult for harmed civilians (and their contingent-fee counsel) to rely on the tort laws, rather than trying to create and maintain uncertainty that makes it difficult for the armed officers to rely on the "acoustic gap" to commit wrongs against the civilians. And so the harmed civilians (and counsel) who reasonably rely on the fact that there was wrongdoing to bring cases against the officers find that there is uncertainty whether their reasonable hard work and investment in the case will be for naught. Doesn't seem like the right result.

  • FlameCCT||

    Your example is basically saying that a 25% failure rate is acceptable. 0.5/2.0=.25

  • art guerrilla||

    1. had to surf nekkid (using edge with no adblock, noscript, ghostery, etc) in order to reply here, an all too common stupid web site trick... thing is WITHOUT adblock, etc, surfing the web becomes an unholy mess that is so-o-o-o-o unenjoyable, I probably wouldn't do 1/10th the online surfing I do if I were 'forced' to use a browser without all that stuff turned on FOR MY PROTECTION AGAINST YOUR anti-human policies...
    .
    2. to a semi-informed layperson, the actual 'issue' seems like arguing how many angels can dance on the head of a pin, rather than any actual moral decisions going on to promote the greatest good for the greatest number...
    .
    3. really, the last line of the op/ed was all that was necessary, no further complicated legalisms required: the supremes heart empire, water/wet, sky/blue, and 99%/fucked...
    .
    but -yeah- vote HARDER, mofos, THAT is surely going to fix Empire... *snort*

  • nastavm||

    Perhaps one more step that extends your reasoning might help explain the courts' interest in this topic, and I offer this theory.

    I suspect that the 'decision rule' is not quite a rule per se - rather it's a set of heuristics that must be discerned by the circuits (and everyone else) by extrapolating SCOTUS decisions. Extrapolation is, of course, necessary only when a situation arises that has not been explicitly addressed by a precedent already - which can and does happen.

    The problem, I think, is that there are now two competing authorities that can be used to fill-in the gap when lower courts must try to address a situation that is somewhat novel - (1) 'decision rule' heuristics extrapolated from SCOTUS precedents on the issue and (2) 'conduct rule' obtained from laws and the constitution, perhaps with some extrapolation and interpolation.

    Sometimes, lower courts choose to apply #2 over #1, and it would seem that SCOTUS would prefer that the lower courts did the reverse and always chose #1 instead. It would seem that it picks up cases for review where a lower court applied #2 instead of #1 and 'fixes' the 'erroronerous approach'.

  • Rev. Arthur L. Kirkland||

    Most of the justices -- and most judges -- are cop succors.

    Plenty of judges -- former prosecutors and Republicans, especially -- are no-good, lousy cop succors.

    If there is a better explanation for awkward judicial lurches to rescue law enforcement personnel from accountability for misconduct that is depressingly common and flagrant, I have not encountered it.

    From a libertarian perspective, the practical antidote includes fewer former prosecutors and fewer Republicans on the bench.

  • Ben_||

    Government operatives benefit from this double standard. For the rest of us "ignorance of the law is no excuse" and government operatives get to impose strict standards of conduct upon us. Except cronies get a pass. Great system, SCOTUS.

  • FlameCCT||

    Basic math word problem. As Peter Gerdes comments, similar to the way different review standards are used to change the bright line rule for conduct. Perhaps QI needs to be set at a specific review standard or become its own review standard. Anyway, I've tried to clean up the simple math problem for Prof Kerr's example below.

    If the bright line rule for conduct is:
    6.0 Arrest
    8.0 Wiretapping
    3.0 Terry
    4.0 Excessive use of force

    2.0 Qualified Immunity, which provides an additional buffer (aka reduction of the conduct line)

    The bright line rule for conduct under Qualified Immunity becomes:
    2.0 Arrest (4.0-2.0)
    6.0 Wiretapping (6.0-2.0)
    1.0 Terry (3.0-2.0)
    2.0 Excessive use of force (4.0-2.0)

  • BillB6||

    For a concise analysis of why the Supreme Court reviewed this case, see Andrew Branch:

    https://legalinsurrection.com/

  • jdgalt1||

    What a pile of ... sophistry.

    Having accused people go unprosecuted because they were abused by police does not do harm. They have already been punished. What does harm is letting bad police get away with their crimes.

    There needs to be no immunity of any kind for anybody. Only a law that applies to everybody will ever get or deserve respect. (Of course, this is not a sufficient condition, just a necessary one. Next we need to address overcriminalization.)

    The worst part of the current situation, of course, is that victims of unjustified use of force by police have no recourse except to shoot back. If the system doesn't approve of that outcome it should stop causing it.

  • Voize of Reazon||

    In my view the conduct vs. decision rule dichotomy is not as nonsensical as many here seem to think, it is just a manifestation of the risk-aversion inspired gap between rule and behavior that QI seeks to address. Elsewhere we have unqualified immunity, e.g. for judges and legislators, again because of the same concern: A prudent person in a role where they are continually required to make decisions that risk their well-being will keep some distance between themselves and the precipice, and do their jobs more timidly than society requires.

    That doesn't mean any form of immunity is the only way to deal with these problems. Doctors function without immunity by substituting malpractice insurance, maybe that would be another path to take, but immunity is where we are.

    I see QI as modeling that same prudence-informed gap between behavior and enforcement. If society wants officers to behave at 4.0 on the Kerr scale, and expects they will average about 2 points worth of prudence, then it will adopt a decision rule that holds them liable if they exceed 6. That rule will be publicized, and when officers individually or collectively avoid the edge the desired result is achieved. So, the "conduct rule" isn't a separately promulgated rule except maybe it is taught during training as a recommended way to avoid trouble.

  • Gunstar1||

    This doesnt explain why the court does not take as many cases going the other direction, where QI is granted but should not have been.

    It seems SCOTUS is not interested in making sure the buffer isnt artificially to large, only when it is too narrow. That will lead to a push closer to 0 on the scale rather than 2. As judges know if they err on the short side they will get reversed, on the large side, nothing will happen.

  • nonzenze||

    This assumes that judges care if they get reversed. I don't believe that's the case.

  • DjDiverDan||

    Qualified immunity - the only area of law in which I firmly believe that Justice Sotomayor got it absolutely right (in her dissent in Kisela v. Hughes) and the majority got it absolutely wrong, despite my firm belief that Sotomayor is otherwise the biggest idiot on the Court (based mostly, but not entirely, on her moronic dissent in Schuette v. Coalition to Defend Affirmative Action BAMN).

  • PDan||

    This would probably explain the AEDPA summary reversals too. But there are a lot of other legal issues that seem to share the same general features of the conduct versus decision rules. For example, criminal laws forbid conduct, but the proof-beyond-a-reasonable-doubt standard is a decision rule that says, even if a person has broken the law, we will not hold him liable if there is a reasonable doubt. And that standard, unlike qualified immunity, is important enough to be deemed a constitutional necessity. So why isn't the Supreme Court constantly deciding criminal sufficiency-of-the-evidence cases?

  • FlameCCT||

    Perhaps they dealt with enough cases in the past and/or that the line between the different standards is clear although they are constantly being pushed.

  • Voize of Reazon||

    Reasonable doubt is decided by the jury and reversing jury findings of fact requires a determination of clear error. More commonly standard of evidence appeals allege incorrect jury instruction, a legal error, and there were a number of high profile appeals of this in the 70s.

    That doesn't answer the more general question, why the Court hears so many QI denial appeals when there are other serious and unsettled areas of law. I can see a rationale that works like this:

    The Court's concern is always with the downstream effects of precedents, and not with correcting errors in individual cases, so the intensity of their concern depends on how widespread and destructive the justices anticipate those downstream effects will be.

    If you accept that QI exists to avoid having police frozen into inaction by a natural but excessive caution, then it follows that a few well-publicized cases of uncorrected QI denial could have a large enough chilling effect on police work to destabilize society.

    So, to pick an example, "Why doesn't the Court grant cert on more cases where QI was granted incorrectly?" it may be as simple as fearing lawlessness more than tyranny.

  • PDan||

    Sufficiency of the evidence is a legal question reviewed de novo, not for clear error.

  • Voize of Reazon||

    The court has to determine "whether, after viewing the evidence in the light most favorable to the prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt" (Jackson v. Virginia). De novo review in and of itself is not deferential, but this test is. We don't call it review for clear error but it amounts to the same thing.

  • PDan||

    Not much different from the QI test: whether, after viewing the evidence in the light most favorable to the plaintiff, any reasonable law enforcement officer would have understood his conduct to be unconstitutional. Or something along those lines.

  • Voize of Reazon||

    No, not at all. The sufficiency of evidence test is met if a rational (not "reasonable") juror could have found it sufficient. It requires only a rational basis. The QI test is met if a reasonable government official would (not "could") have known their actions were unconstitutional. There is a lot of room between could and would.

  • WJack||

    'Fearing lawlessness more than tyranny" . . . this simple answer is most likely the correct answer. Law enforcement is all that stands between order and disorder. Perhaps most Supreme Court Justices take this into consideration in QI cases and give the police the benefit of even the slightest doubt because otherwise, some police officers might not enforce the law for fear of punishment.

  • Michael Masinter||

    I think the Court's current majority would, if writing on a clean slate, decide Monroe v, Pape in accordance with Justice Frankfurter's dissent. But overruling Monroe v. Pape would be difficult to justify in light of conventional stare decisis principles in statutory construction cases, so instead it uses qualified immunity to relentlessly shrink the set of police misconduct cases in which section 1983 might afford individuals a damages remedy. Monroe v. Pape remains, but only as a mirage, or perhaps as the end of the rainbow, never quite within reach.

  • jsfreason||

    This discussion made me think of the long period of obscenity rulings. The Court had a principle that obscenity wasn't permitted by the First Amendment, and was then forced into long nights watching stag films to decide fact-bound individual cases, and not finding any particular film obscene. This continued until the world understood that the Court was no longer going to find anything not involving children obscene so they stopped bringing the cases. The Court seems to be saying hte same thing over and over here, but prosecutors don't seem to get the hint.

  • David Nieporent||

    What prosecutors? QI cases are virtually always civil suits by private citizens.

  • TwelveInchPianist||

    People have a difficult time accepting the truth that judges (and the justices) allow cops to do terrible things because they are bad people.

  • FlameCCT||

    It would be interesting to see how many QI cases and the circumstances surrounding them have been decided at the district court and appellate court levels. That might give a better indication of how the various courts are dealing with it.

    Personally I'm mystified as to how this particular case even made it to SCOTUS level as the use of force is authorized for law enforcement and civilians in defense of self and others. There is no way for LE or civilians to take into account the mental state of an armed individual. I remember two officers being gunned down while making a simple wellness check of a person with mental issues in Albuquerque.

  • BillBentnickel||

    I'm late to the thread, but very much enjoyed reading all of the thoughtful back-and-forth in the comments.

  • bernard11||

    Does it make sense to think about this idea as analogous to highway speed limits? There is a posted limit - the conduct rule - and a higher enforced limit - the decision rule.

    But most drivers ignore the conduct rule and try to obey the decision rule.

    Why is there a difference? Maybe partly because the speed measured by radar has some error in it, so you want to be sure the driver really was speeding. On this analogy QI just gives some leeway for imprecision. But that doesn't explain why egregious misconduct gets QI. After all, if you get stopped going 90 mph it's hard to claim the radar gun was off by enough that you weren't really over the limit.

  • Dave W.||

    Is probable cause a 2.0, 4.0, 6.0 or 8.0?

GET REASON MAGAZINE

Get Reason's print or digital edition before it’s posted online