As part of BIALL’s November Newsletter, Daniel Hoadley examined ‘Good’ law

Back in September, Jane Riley (of the Manchester Law Library), posted an important question on LIS-Law which I have paraphrased below:

“If there is no case status (no positive, considered or negative treatment, etc.) on either [name of online service] or [name of another online service] next to a particular Court of Appeal case, does that mean that the case has not been judicially considered? And, if so, can the case be classed as good law?”

At the heart of this apparently straightforward query is a broader, more fundamental question: can we rely on online legal research platforms to tell us whether cases we are thinking of deploying still represent “good law”? My view on this is simple: no, we cannot.

The era of the traffic light indicator

Most online platforms incorporate some visual mechanism to indicate what the future authoritative value of a case may be. The mechanism favoured by most online publishers is some sort of “traffic light” system.
Notwithstanding minor variations in presentation and terminology, a common formula is green for positive treatment, red for negative, blue for neutral and yellow for mixed judicial.  The traffic lights undoubtedly have their uses, but it is critical to understand they do not relieve the user of the responsibility to check that the case remains good law by reference to the purpose for which it is to be deployed.

What does “judicially considered” actually mean?

The first thing we need to be clear about is what we mean by “judicially considered”. Has a case really been “judicially considered” by virtue of fact that the judge refers to it once, along with an ocean of other authorities, in the course of a judgment running to a hundred paragraphs? The threshold must be higher than a mere mention. The better test is this: does the decision in the earlier case play a material role in the court’s reasoning in the later case? If so, then the earlier case has been “judicially considered” in the later one.
However, to carry out this operation properly, “someone” needs to read and analyse the judgment in the later case to determine whether the earlier case is being considered. And, preferably, that “someone” needs to possess the requisite legal knowledge and skill so that the rest of us can place some trust in their assessment.

Okay, so the case has been “judicially considered”.

What type of consideration did it receive?
The job doesn’t end there. Logically, once we’ve concluded that a case has been judicially considered, the next task is to categorise the class of treatment the later case is meting out against the earlier case.
This is where the simple traffic light system starts to wobble. The traffic lights can tell us the broad mode of treatment (good, not good, neutral and in-between), but they do not tell us the specific class of treatment. If the traffic light says the case has been positively treated, what type of positive treatment was it given? Was the case applied, followed or approved? If the case has mixed judicial treatment, was it distinguished or explained by the later case? This may strike some as pedantry, but when it comes down to it, these modalities may have a significant bearing on the future application of the case and the researcher able to identify and understand these distinctions is in a better position than the researcher who cannot.
Again, these categorisations do not magically materialise out of thin air – “someone” has to read the case and make an assessment.


I have given the algorithmic approach to judicial consideration pretty short shrift elsewhere and I’m going to do the same here. This is not the kind of research operation that is amenable to the pushing of a button and trusting whatever a computer churns out. The problem is that when a judge is considering a case, she will very rarely say something like, “I am considering/applying/distinguishing this case” in the judgment. More often than not, any conclusion on judicial consideration turns on being able to distil the ratio of the later case, comparing it against the ratio of the earlier case and then inferring the class of consideration from that inquiry.
No computer-driven service can perform this task better than a person.
Good law?

Good law for what?

An important limitation of the traffic light system, or indeed any system of case-based legal analysis, is that it will only tell you whether the status of a case, or of a particular proposition of law established by that case, has been affected by subsequent case law. That is not, however, the only factor potentially affecting the question whether a case is “good law”. A fairly stark example will suffice to point out this limitation.
In 2008, the House of Lords gave judgment in R v Davis [2008] UKHL 36; [2008] 1 AC 1128. R v Davis decided that where prosecution witnesses gave evidence from the witness box under conditions of anonymity, the conduct of the defence case would be so hampered as to render the trial unfair and a conviction unsafe.
Now, let’s suppose I’m appearing in the Crown Court with instructions to resist an application by the prosecution for their star witness to give evidence anonymously behind a screen in the witness box. It looks like R v Davis will deliver a deathblow to the prosecution’s plans, doesn’t it?

If you look up R v Davis on WestlawUK, Lexis Library, JustCite and ICLR Online (see, I can be objective), you will see that WestlawUK’s traffic light reports “positive or neutral treatment”; Lexis says “positive treatment indicated” and JustCite and ICLR Online rank Davis as having been “considered”. The traffic lights on four online services are telling me that I should be safe to say that Davis is good law for the proposition that prosecution witnesses cannot give evidence anonymously in a criminal trial, right?
Wrong. The traffic lights are not telling me one, absolutely whopping detail that blows a massive hole in Davis’s value as authority: the decision in Davis was overruled by Parliament (in the Criminal Evidence (Witness Anonymity) Act 2008, which was superseded by the Coroners and Justice Act 2009) only a month after the decision was handed down. If I had relied solely on the traffic light indicators, I would have been completely in the dark.

Moral of the story: the answer as to whether a case is good law for a proposition of law may not lie in the case law. The silver bullet or the fatal blow may actually be lurking in the statute book.
There are limits to what legal publishers are able to deliver
Comprehensively sweeping the case law landscape and analysing judgments requires massive quantities of time and qualified human resource. All of us publishers are trying to do the best job possible with the resources we have available. But there are limits to what we’re able to do and as usual things boil down to a trade-off between quality and quantity of output.

It is not the job of legal publishers to pummel you with an endless torrent of ubiquitous and potentially flawed information. Our job is to approach legal publishing as carefully and as thoughtfully as our resources permit, in order to guide and equip the researcher as accurately and reliably as possible.  From there, the responsibility shifts to the researcher to apply their knowledge and skill to reach conclusions on the law that they are sufficiently secure in to put there name on. 

Law is an art, not a science. There are no binary questions and there are no binary answers. Online research has opened up mega quantities of information, some of it good, some of it not so good. But, let’s not allow ourselves to be dazzled by the coloured lights and technical wizardry. At the end of the day, good old fashioned human research skills and savvy are what gets the job done best, both within and beyond the four corners of your computer monitor.

Daniel Hoadley, Barrister – The Incorporated Council of Law Reportingments retrievable using Search Center.