Deception Detection In Non Verbals And Linguistics

Susan Neill-Fraser Murder Conviction: Linguistic Analysis Of Interviews And Statements + DNA Evidence Part1

The murder trial that rocked and divided Hobart --  no body, no weapon, forensic evidence (which didn't link Susan Neill-Fraser to the murder) and circumstantial evidence combined with her multiple changing lies in police interviews sealed her fate and she was was convicted and jailed in 2010 for 23 years for murdering her partner Robert Chappell. Chappell disappeared from their yacht the Four Winds and has never been found. With all her appeals exhausted, Tasmanian law allows an application for leave of appeal by a single judge if new "compelling"evidence is available. This is to be heard in October 2017.

No Body
Many people can't get over the fact that no body was ever found, but as shown on wikipedia, no body convictions are unusual but not rare with countries such as Australia, Belgium, Brazil, Canada, France, Germany, India, Malaysia, New Zealand, Philippines, Portugal, Spain, U.K and United Stated having no body convictions.

David Butt is a lawyer in Canada that has successfully convicted a no body murder and gives 3 points to be successful. No body conviction
"The third key to success in a murder case without a body is to eliminate alternative explanations for the disappearance. People disappear without a trace for a very limited set of reasons: elopement, kidnapping, suicide and murder. If you convincingly eliminate elopement, kidnapping and suicide, the jury will themselves eliminate the malevolent pixie and the alien abduction theories, and you will be left with murder."

DNA Evidence
DNA evidence has been incredibly useful in securing convictions, but it has also been responsible for miscarriages of justice. The increased sensitivity of DNA analysis has meant increased likelihood of contamination and transfers (see below).

Before the linguistic analysis of what was said during the trial, let's look at the DNA evidence that was submitted in the trial, as well as an overview of DNA evidence in general.

The trial transcript runs just short of 1600 pages and can be found by a google search:

Unknown DNA found on yacht on day 3.

On the morning of 27th January 2009, the yacht Four Winds was reported to police that it was low in the water and sinking. When the police boarded they found blood on the staircase and a knife on the floor and Robert Chappell missing. The yacht was worked over by forensics, then the water was pumped out and after a brief stay at Constitution Dock, the yacht was  towed to a dry dock at Goodwood called Cleanlift for inspection and repairs by the insurance company.

On the third day after the disappearance of Robert Chappell, at 1.40pm on the 30th Jan, more DNA swabs were taken around the boat while it was at Cleanlift. A swab was taken on the walkway of the boat, near the gate, after a luminol reading showed positive.

It was not able to be determined what caused the luminol to show positive, but the DNA registered "strongly" but was unknown to the database.

An obvious problem was the contamination of the boat and the crime scene on the third day of the disappearance, with the police estimating at least twenty people that they knew of had been on the boat, including police, the insurance company, Robert Chappell's family and so on.

It could not be established that the gates were locked at Cleanlift during this period, but it was known that there were previous "break-ins", according to the trial. The boat was also still accessible by water at Cleanlift on the 30th January.

The unknown DNA from the walkway on the boat wasn't matched and that's where it stayed. A forensic report was prepared and then voila....a DNA match on the database. A young homeless girl, Meaghan Vass is arrested for stealing and her DNA is matched to the boat.

The swab from the yacht and the homeless girl Meaghan Vass were from the same lab, Forensic Scientific Services Tasmania. It never occurs to anyone in the trial that cross contamination could be a possibility even though this is exactly what happened with item 35 of evidence. (see below).

Unknown Yacht DNA Matches To Meaghan Elizabeth Vass

Prosecutor Tim Ellis:

MR ELLIS SC: - analysis didn’t match it with anyone, but then Ms
Vass fell foul of the law and when her DNA was taken it did match
that. And so we have this piece of DNA that seems to belong to her
on the deck of the Four Winds and no one quite knows how it has got
there. She has been uncooperative, and to be fair, I think she suffers
from some form of Aspergers or something like that, but she hasn’t
been cooperative with us or police in – truly cooperative in
furnishing a proof of her whereabouts and movements. So the point
about – oh my learned Mr Shapiro points out that – that she’s
mentioned at 1049, the DNA profiling of her.

She is adamant she has never been on the boat, never been to the dry dock at Goodwood or when the boat was at Constitution Dock. Police Intelligence finds nothing to link Meaghan Elizabeth Vass to the crime scene. Video footage from Hobart Ports show no one boarded the boat while it was at Constitution Dock.

Defence lawyer Mr. Gunson tried to place Vass on the boat (which would be advantageous to his client, Susan Neill-Fraser) by taking up a few pages of transcript, but in the end, she denied ever being on it and that's where it ended.

The forensic lab that analysed the DNA was next with Crown Prosecutor Mr. Ellis:

MR ELLIS SC: Thank you, your Honour. (Resuming): How – I’m
sorry, is it possible that a person who hasn’t been on that surface
from which the surface – from which the swab has been taken, is it
possible that she hasn’t been there, notwithstanding that a swab has
revealed her DNA?

Mr. Grosser: It is entirely possible. One of the things
about DNA – it’s fairly common in bodily fluids and those sorts of
things, blood saliva and once that’s outside of a person’s body, or off
a person’s body, there is a potential for that to be transferred in some
way, so if for example I was to bleed onto a tissue, somebody could
pick that tissue up and spot it against a wall and then there would be
a blood stain on a wall that I’d never seen that potentially carried my

Defense don't accept secondary DNA transfer explanation with Mr. Gunson calling it "near impossible" as he pushes the forensic witness Mr. Grosser over several transcript pages to admit that secondary transfer was highly unlikely, something Mr. Grosser does not concede.

Prosecution call the DNA a red herring saying the scenarios relating to how secondary DNA transfer could have happened are "limited only by your imagination."

Increased Sensitivity Lead to More Potential DNA Errors
Jonathon Koelher, University Of Texas -- DNA Is In Error 1% Of The Time

Jonathon Koelher from the University Of Texas--
Statisticians in Austin, Texas, gained access to the first accuracy tests carried out on DNA laboratories, which were conducted anonymously. Researchers had asked the labs to match a series of DNA samples. They knew which ones were from the same person, but found that in over 1 per cent of cases the labs falsely matched samples, or failed to notice a match. The statisticians then calculated that a substantial human error had occurred in 12 in every 1,000 tests.

Cross Contamination Of DNA Evidence
"Contamination can occur despite the procedures in place to deal with them."
- NSW Judge Andrew Haesler

The Potential for Error in Forensic DNA Testing (and How ThatComplicates the Use of DNA Databases for Criminal Identification)  by William C. Thompson Department of Criminology, Law & Society University of California, Irvine who wrote the above paper was asked to assist council for the Victorian State Coroner into the murder of toddler Jaidyn Leskie in Victoria in 1997.

Cross Contamination Case 1

He writes about the Jaidyn Leskie case:

"One of the best-known false cold hits occurred in a high-profile Australian case involving the murder of a toddler named Jaidyn Leskie. The toddler disappeared in 1997 under bizarre and mysterious circumstances while in the care of the boyfriend of the toddler’s mother.

The toddler’s body was found in a reservoir six months later, with a crushed skull, and the boyfriend was charged with murder. But the case was clouded by the discovery of DNA from an unknown woman in what appeared to be bloodstains on the toddler’s clothing. In late 1998, the boyfriend was acquitted.

In 2003, the unknown DNA was matched, via a database cold hit, to a young “mentally challenged” woman who lived hundreds of miles away and who, by all accounts, had never left her own village. 

Police could find no way to link the young woman to the toddler’s murder and at first dismissed the cold hit as an “adventitious” (coincidental) match. It was a seven-locus match and the estimated frequency of the matching profile was 1 in 227 million.

When the case became the subject of a Coronial investigation, I was asked to assist counsel for the Victorian State Coroner in reviewing the laboratory records. This review established that DNA from the young woman had been processed through the same laboratory at about the same time as the toddler’s clothing. 

The young woman had allegedly been the victim of a sexual assault involving a condom. Her DNA, which was extracted from the outside of the condom, had been in close proximity in the laboratory to extracts from the toddler’s clothing. 

Although laboratory personnel maintained that accidental transfer of samples between cases is impossible, I was able to document dozens of cases in which cross-contamination of samples had occurred under similar circumstances in other laboratories, and therefore suspected that accidental contamination explained the match with the young woman.

In order to test the alternative theory of a coincidental match, the Coroner had the matching samples tested at additional genetic loci. If the DNA on the toddler came from another person, and the seven-locus match to the young woman was coincidental, then one would expect testing at additional loci to exclude the young woman.

But the additional testing showed that the woman also matched at six additional loci. Furthermore, re-examination of the data produced in the first test revealed low-level matching alleles at two additional loci.

Altogether there were fifteen matching loci with an estimated frequency of less than 1 in 100 trillion, which made the theory of a coincidental match seem far less plausible than the alternative theory of cross contamination. The Victorian State Coroner issued a formal finding in 2006 that the evidence linking the young woman to the toddler was a false match caused by cross contamination in the laboratory. "

Cross Contamination Case 2

Farah Jama Cross Contamination And Wrongful Conviction

Farah Jama was wrongly convicted of rape and spent 15 months in jail before being exonerated, due to cross contamination of the sample with one from a rape victim. There was up to 40 hour time difference between Jama getting his DNA taken on an unrelated matter, and a rape victim being swabbed by the same lab which caused the contamination, Victoria Police Forensic Services.

"A judicial enquiry found that Jama’s DNA had been allowed to contaminate the crime scene sample because of faulty collection procedures."
-- From May 2010 Report of Justice Frank Vincent AM “Inquiry into the circumstances that led to the conviction of Mr Farah Abdulkadir Jamal.”

Cross Contamination Case 3

More DNA Contamination Errors, Again From Victoria Police Forensic Services Who Picked Up And Corrected The Error In Time.

William C. Thompson Department of Criminology --
 "Very recently, yet another false cold hit came to light in Victoria, Australia. On August 6, 2008, the government dropped charges against Russell Gesah, who was about to stand trial for the 1984 murder of a woman and her nine-year-old daughter. Gesah had been incriminated in the murders when a DNA sample allegedly from the murder scene was matched to his DNA profile through a cold hit. On the eve of trial, it was discovered that “an unrelated exhibit containing DNA from Mr. Gesah was tested on the same day and in the same place as material from the [murder] scene…” According to an officer who worked at the Victoria Police Forensic Services Centre, “crime scene samples, including bloodied clothing, were left on sinks and open shelves” in a manner that could have allowed items from different cases to be cross-contaminated. In light of the problem, a deputy police commissioner announced that the police force would re-examine DNA evidence in over 7000 previous cases looking for other such opportunities for cross-contamination."

Cross contamination by Paramedics
How an innocent man's DNA was found at a crime scene.
Here and here.

DNA Transfer During Autopsies.
Academic Report, Germany.
"Using DNA-free swabs, we took samples from instruments used during autopsy and autopsy tables. Surfaces and instruments were routinely cleaned before sampling. Swabs were subjected to different PCRs to quantify the total amount of DNA and to amplify individual specific STR-markers. In most samples, alleles that could be linked to bodies that had been autopsied before were found."

Secondary and subsequent DNA transfer during criminal investigation.
Academic Report
"We show that with use of the new highly sensitive technologies available in forensic DNA analysis there is an enhanced probability to obtain a DNA-profile which has not been directly deposited on the object but is an outcome of one or more transfer events. The nitrile-gloves used by investigators during exhibit examination can act as a vector for DNA transfer from one item to another."

Cross Contamination of DNA in Neill-Fraser Trial.
Mr.Ellis, Prosecution: You found the DNA profile of an FSST staff member on one of the
items. Does this happen occasionally?

Mr. Grosser, Forensic Scientific Service Tasmania: This does happen
occasionally. It was actually not necessarily on the item but on the
sample that was submitted for DNA so what happens when we’re
taking a sample from an item is we’ll look at something – in this
particular case I believe it was a swab that was taken at the scene,
and then in the laboratory we’ll do some handling of that swab and
sampling of that and placing that in a tube and what can potentially
happen is that somebody working in the laboratory may actually
inadvertently leave some DNA behind on that sample or have
contaminated the tubeware and plasticware that we’re using the
laboratory. We take a number of precautions to try and avoid this so
it is reasonably rare but it does happen from time to time.

DNA as only evidence of identity
When the Crimes (Forensic Procedures) Act 2000 (NSW) was introduced the Police Minister Paul Whelan was explicit: Hansard, NSW Legislative Assembly, 31 May 2000, p 6293.

“It is important to note that DNA will be only one tool in the police officer’s kit. They will still need to assemble a brief of evidence against the offender; DNA alone will not convict!”

NSW Judge Andrew Haesler wrote a paper Issues in Gathering, Interpreting and Delivering DNA Evidence

"In New South Wales the Court of Criminal Appeal has held that a DNA profile match could
not in the absence of other evidence prove beyond reasonable doubt that the accused was
responsible for leaving the crime scene stain: R v Green, unreported CCA NSW 26/3/1993;
R v Pantoja (1996) 88 A Crim R 554 & R v Milat (1996) 87 A Crim R 446 at 447"

NSW Judge Andrew Haesler (above) writes:

"To allow a jury (or a judge) to find guilt because of evidence of a DNA, profile match and
supporting statistics assumes four things, which are far from certain. It assumes:

1. That the statistical calculation given is real as opposed to one of a number of
possible statistical or mathematical construct that could be used to make an estimate
of the profile’s rarity.

2. That the allowances made for the impact of the distribution of and variations between
DNA profiles in families and general and specific populations are accurate.

3. That there has been no contamination in the collection or analysis of the sample.

4. That the statistical evidence interpreting the significance of the DNA match is
evidence of the probability that the appellant was the source of the incriminating DNA
rather than one of a number of circumstances that may be taken into account in
reaching that conclusion."

Judge Haesler:
"Courts cannot ignore the fact that we are dealing with such small samples that the possibility of secondary transfer and possible contamination and even corruption are ever present dangers. Contamination can occur despite the procedures in place to deal with them. Courts cannot ignore the limitations placed on any police prosecution or defence investigation, by cost constraints and other
resources issues."

Software, population stereotype corrections + interpretation errors in DNA:

The Fraser review made three key findings:

1. The software used by Victoria Police Forensic Services was not designed to deal with allele drop out and had led to systematic error in favour of the prosecution.

2. Interpretation errors in some profiles had wrongly led to them being identified as
mixtures or complex mixtures in too conservative a manner thus favouring the defence.

3. There had been inconsistency in interpretation of the DNA profiles by different case
managers in Victoria Police Forensic Services.

Secondary Touch DNA More Likely As DNA Tests Become More Sensitive.

Cynthia Cale (above) has done studies where a two minute handshake was conducted, then a knife picked up by one party, showed that in 85 percent of the samples the knife now had the DNA of both people on it. Touch DNA Transfer Article

A semen stain on a pair of pants and put in a washing machine will be able to be picked up on all the other washing afterwards. Academic Report

More Lab Errors Making News:

Austin crime lab bucked DNA standard for years, yet got passing grades.
Dozens of criminal cases could be in jeopardy after a litany of errors by a crime scene investigator at the Houston crime lab.
Detroit Police Lab Is Closed After Audit Finds Serious Errors in Many Cases.

UK 2016 Database Audit, error rates detected and corrected:

Statistics Of DNA

As relatives become more distant, the DNA that is shared becomes less. DNA statistics are worked on the basis of independence, as each point or allele is shared in the chain, the probabilities are multiplied, so errors are multiplied too. Independence means parents were random, not distant relations. Races have different probabilities at some alleles and so the exact population mix must be compensated for mathematically. The population in the U.K has a different average DNA mix to Australia or the U.S.

The genetics of relatives and percentage shared DNA is below:

The double helix DNA strand we are all familiar with is only measured at a few points to see if there is a match. An 8 or 9 loci (allele) point match is generally done, although the FBI is pushing to use a 13 loci match for it's database.

The probability of a match using 1 locus point is (1/10)^1 so a 3 point match is about 1/10^3=1000
This is assuming complete independence between loci points.

For example, in the Arizona DNA database, there are 65 000 samples, yet there are 144 people whose DNA matches at 9 loci (points). This casts doubt on the astronomical probabilities that are often quoted in DNA results.

Now the pairs that match in the Arizona database is similar to the Birthday Problem, the probability of any pair matching as opposed the the problem of matching a specific pair as in a DNA match. The birthday problem says that having 23 people in a room gives a greater than 50% chance that there will be a matching birthday date in that group. This is so because there are 253 pairs giving a greater than 50% chance of a match.

This does not describe a typical DNA match problem, but it can still be used as a check for the base rate probabilities. The probability of a 9 loci point matching pair in the Arizona database is calculated my mathematician Keith Devlin using the birthday formula approach at about 5 percent.

"A survey of DNA databases by the National Institute of Forensic Science examined 33,858
profiles, and found 206 matching pairs! This was significantly more than the statistical
model predicted. Matching pairs can be explained as twins, brothers or duplicates,
because offenders use aliases or because of genuine coincidence. Only by investigating
each match can the real reason be known. To date the investigation needed to explain the
matches has not been done."
-- Issues in Gathering, Interpreting and Delivering DNA Evidence, Judge Andrew Haesler

As DNA databases become larger, there is a problem called the Cold Hit Search.
Let's say they sell 14 million lotto tickets a week (the probability of winning lotto). This means each week (or every few weeks) there is a lotto winner. Now if someone wins the lotto at 14 million to one, what does this tell us about this person after the lotto win? Nothing! His story about his birthday numbers or lucky weekday or anything else mean nothing.

But what if now he phoned a reporter and said he would win lotto before the event? Now a journalists comes over and documents the event and voila...he wins the lotto. Now what can you say? Well, now you would think something is going on, a possible scam. Long odds before the event tell us nothing, long odds after the event tells us a lot.

A database match tells you there is a match, it doesn't determine guilt. Just like lotto, the identification needs to be done before the DNA match is done.

What DNA Is Not
JM Butler one of the most respected figures in the field of DNA analysis made this critical point:

“It is important to realise what a random match probability is not. It is not the chance
that someone else is guilty or that someone else left the biological material at the
crime scene. Likewise it is not the chance of the defendant being guilty or the chance
that someone else in reality would have that same genotype. Rather, a random
match probability is the estimated frequency at which a particular STR profile would
be expected to occur in the population. This random match probability may also be
thought of as the theoretical chance that if you sample one person at random from
the population they will have the particular profile in question.” 

As Daniel Kahneman, the only psychologist the win an Economics Nobel Prize said, human beings are very badly suited to deal with probabilities. We don't understand statistics and probabilities, even mathematics professors have a problem with the simple 1 in 3 Monte Hall problem.

Presenting DNA Probabilities In Court Or The Prosecutors Fallacy
The prosecutors fallacy is nearly always written up in newspapers when describing DNA probabilities.

wikipedia --: "The prosecutor's fallacy is a fallacy of statistical reasoning, typically used by the prosecution to argue for the guilt of a defendant during a criminal trial. ... At its heart, the fallacy involves assuming that the prior probability of a random match is equal to the probability that the defendant is guilty."

Professor Martin Neil puts The Prosecutors Fallacy like this:
  1. Incorrectly reporting the probabilistic impact: In reporting the impact of the DNA evidence it appears (based on the Telegraph report) that the prosecuting QC has yet again committed the prosecutor's fallacy. The statement that there is “a one billion-to-one chance that the DNA belongs to anyone else" is wrong (just as it was  here here and here). In fact, if the DNA profile was indeed such that it is found in one in a billion people, then it is likely to be shared with about six other (unknown and unrelated) people in the world. In the absence of any other evidence against the defendant there is actually therefore a 6/7 chance that it belongs to 'anyone' else.

Warning A Jury When Presenting DNA Evidence

"As Doyle CJ pointed out in Karger , the reason for warning a jury about DNA statistics is:
“The risk that the jury will reason that the evidence of the likelihood ratio or match
probability expresses the probability that the incriminating DNA was the DNA of the

 "The statistical evidence interpreting the significance of the DNA match is not
evidence of the probability that the appellant was the source of the incriminating
DNA. To so regard it would be to make an error."

-Issues in Gathering, Interpreting and Delivering DNA Evidence, NSW Judge Andrew Haesler

DNA Presented In Tasmanian Courts

Based on the court transcripts of the Susan Neill-Fraser case, is the Prosecutors Fallacy being invoked in Tasmanian Courts when presenting DNA statistics?

Crown Prosecutor Mr. Ellis seems aware of the Prosecutors Fallacy but doesn't really have a handle on it:

HIS HONOUR: All right, well what’s your submission, Mr Ellis?

MR ELLIS SC: The witness has said that he has a rudimentary
knowledge of the statistical interpretation gained by hearing people
with such expertise speak about it in the laboratory in which she
works. The cleaner might get that, with all respect to Ms McHoul,
the cleaner there might develop the same degree of expertise. And
to underline the danger of the question, as I understand it, my learned
friend is putting it in terms of what’s often referred to as ‘the
prosecutor’s fallacy’; that is odds of a hundred million to one,
whereas, the - well I won’t – I won’t say it in the presence of this
witness in case you allow the question, but he’s introducing the
prosecutor’s fallacy, so called, through a witness whose expertise
goes – is claimed to come from no more than hearing other people
with expertise talk about it.

HIS HONOUR: Mr Gunson?

MR GUNSON SC: I don’t think I need to make any further
submissions, your Honour, I submit the question is allowable on the basis that it’s being put with the expertise this witness has. She is, after all, a forensic scientist of quite some years experience.

HIS HONOUR: All right, well the position is that the – the witness
is a forensic scientist, she’s worked as one for nineteen years. She’s
got a Bachelor of Science Degree with Honours and a Master of
Science Degree in Forensic Science. She’s made it clear that the –
that the significance of DNA probability statistics is not her field but
that she has heard people that she works with and has worked with
who one would – who work in the area discussing such statistics.
s79 of the Evidence Act makes evidence of somebody’s opinion
admissible if that person has specialised knowledge based on the
person’s training, study or experience. The witness’ understanding
of the statistics would appear, from what she said, to have been based
more on experience than on training or study, but she’s a forensic
scientist who has specialised knowledge in relation to DNA and if she
has an opinion as to the significance of a particular statistic then in
my view she’s entitled to express that opinion, it falls within the
scope of what’s made admissible by s79. So I’ll permit the question,
we’ll have the jury back.

MR GUNSON SC (Resuming): Thank you, your Honour. What is
your understanding of what the expression ‘one in one hundred
million’ means in item 20, which relates to Meaghan Vass?

Ms. McHoul: My understanding, as not a DNA scientist, just as someone who has
worked in the forensic biology field, is that the figure means that the
chance of finding someone unrelated in the population that would
also match is around one in a hundred million.

Later on Carl Grosser from Forensic Scientific Service Tasmania is sworn in and he explains the DNA statistics like this:

Mr. Ellis: Thank you. And are you employed as a forensic scientist by
Forensic Science Service Tasmania?
Mr. Grosser: Yes, I am.

Mr. Ellis: Thank you. Now, the very first entry is your DNA profiling is it,
from a swab taken from an EverReady Dolphin torch which had been
given exhibit number 1.

Mr. Grosser: That’s right, yes.

Mr. Ellis: And in the last column, the match is “Robert Chappell, one in one
hundred million.” What does that mean?

Mr. Grosser: In this particular instance
in the two columns that are under the darkly hatched area I’ve got
two pieces of information there. The first piece indicates the DNA
profile type, so in this instance it was what we call a full DNA
profile, that means DNA profile that appears to come from a single
individual and from that DNA profile, one of the regions we look at
is a gender identifying region so we can tell whether a full DNA
profile is from a male or a female, and in this particular instance it
was from a male, and then in the adjacent column I’ve got a match to
Robert Chappell and it says “ One in a hundred million”. That means
that the chance of a second person that’s unrelated to Robert
Chappell, also matching this DNA profile is less than one in one
hundred million.

Mathematicians View Of Tasmanian DNA Statistics In Court

"Peter Donnelly, a statistics professor from London, argued that the DNA probabilities were presented in a way that was misunderstood by the jury; this is known as the prosecutor’s fallacy."

"He claimed that forensic evidence gives us the probability that the defendant's DNA profile matches the crime sample assuming the defendant is innocent, or P (Match | Innocent).
What the jury is looking for though is the probability that the defendant is innocent assuming the DNA profiles of the defendant and the crime sample match, or P (Innocent | Match). "
--Bayes Theorem in Court by Thompson, Danielson, Hartman, Bucheger, Koplitz

To sort this out, I contacted Coralie Colmez and Leila Schneps, French mathematicians and co-authors to Math on Trial: How Numbers Get Used and Abused in the Courtroom.

Coralie is a French born, Cambridge educated mathematician and spoke at TED with a talk called Maths on Trial. She also did an interview about forensic maths: Statistics can be useful in the courtroom—but only if they are applied correctly.

It's clear no interpretation warning is given to the jury when presenting DNA evidence in Tasmania, so I wanted to know what they thought of the bolded statement made by Mr. Grosser cut and paste below. Neither Leila nor Coralie were aware of the trial or the participants, and for all they knew, the Robert Chappell DNA on the torch was from the accused.

I just gave them the comment from Mr. Grosser from Forensic Scientific Service Tasmania --

"I've got a match to Robert Chappell and it says “ One in a hundred million”.
That means that the chance of a second person that’s unrelated to Robert
Chappell, also matching this DNA profile is less than one in one
hundred million."

Their response below--


"The prosecutor's fallacy is not happening in the cited statement (bold). However, it is written in a way that is somewhat misleading and could put the prosecutor's fallacy into the minds of listeners even without actually making any wrong statement.

The passage states: "the chance of a second person unrelated to Robert Chappell sharing the crime trace DNA is 1 in 100 million".  In itself this is not the prosecutor's fallacy, it is correct.  The chance of a random person having the crime scene DNA is just the RMP.  

However, the misleading suggestion, which is not explicitly stated, is the following: "If Robert Chappell is innocent, then the criminal is another person who shares the DNA, and the chance of there being such a person is 1 in 100 million, therefore Chappell has only 1 chance in 100 million of being innocent."

This would be the prosecutor's fallacy, however it is not actually said here, but it's clear that many people might understand this by reading the bolded statement.

In more detail: let's use the notation P(A|B) for "the probability of A given B"........"


They then go on to prove via Bayes Theorem, that the probability would be much lower accounting for the number of people living in a city or a large DNA database search. Bayes gives an exact and accurate method of combining evidence to account for prior probabilities.

So according to the two mathematicians, the phrasing from Forensic Scientific Service Tasmania regarding DNA probabilities is misleading to a jury and is likely to create a bias to the prosecutions advantage.

As well, they reject the 1 in 100 million odds (which are the raw match odds) because Bayes has not been used to to update or account for evidence ie the number of people in the pool who could be included in the  DNA pool.

The standard that is being pushed in Europe although not the USA, are likelihood ratios when using DNA Statistics (which use Bayes) and which have been shown to be slightly harder to explain but more accurate and less misleading and doesn't create a Jury bias.

Latest News
Three people charged with perverting the course of justice relating to Susan Neill-Fraser upcoming appeal.
"On Monday, police hit her with another charge of perverting justice. Police allege that in February Ms Keefe provided false evidence in an affidavit regarding Neill-Fraser and Meaghan Elizabeth Vass."
Hobart Mercury, Aug 24, 2017 

Incredible Account Of The Conspiracy here.

PART TWO To Follow In New Blog.

JonBenet Ransom Note Analysis Using Syntactic Ngrams -- Or Taking The Words Away And Looking At Structure.

New state of the art software is being released in various domains, much of which can help in stylometry analysis. I have decided to bite the bullet finally and move over from Matlab to R, the open source statistical software.

The best permutation and nonparametric combination test software is now on R -

This allows you to compare samples against base without worrying whether your data is complies with the normality curve, or if you have more variables than samples and so on. Devin Caughey has written some very nice papers on this, and now his software is available on R.

Now with the release of Stylo R package, I have well and truly moved over to R: stylistics/home

This is a superb stylometry package with some of the latest developments in stylo analysis such as Burrows Delta and Consensus Bootstrap Tree, rolling Delta etc. These guys know their stuff and have written a great program.

Two more bits of software to complete the analysis puzzle, the state  of the art Stanford Parser from the Stanford NLP Group -

And with the advent of Syntactic Ngrams by Google and others, some great ideas along these lines with with software to produce them, Dr. Gregori Sidorov has an interesting site along with some great papers he has written. He has done some interesting work on the syntactic ngrams and call them sngrams. His site and the software in Python --

Also worth mentioning is authorship software Toccata by Richard Forsyth, along with his other software. I bought Beagle from him in the eighties, and still have fond memories of it. All his new stuff is in Python:

That's a round up of the software, so lets put it together slowly.

The Problem:

A 374 word ransom note at the scene of a murder, or accidental homicide of JonBenet Ramsey. The FBI and police and lead investigator James Kolar agree the note was part of the "staging" of the crime scene.

A staged ransom note means it is trying to portray what it is not. The writer was aware that handwriting would be extensively analysed afterwards, this alone means that handwriting analysis (physically comparing writing) would be useless in a court of law because a lot of effort  would have been made to fake and randomise the appearance of the note, and it could never be "beyond reasonable doubt."

Linguistic Analysis:

Linguistic analysis is an option and has progressed in leaps and bounds over the last few years: (Koppel,Eder, Rybicki, Hoover et al).

It has been known for a long time that people tend to write with their own "style" and using function words, for example "at", "by", "be", "but" and "can" provide linguistic fingerprints because people are unaware of these tiny words and they are not context sensitive, making them a good marker in many cases. 

By themselves they are not enough however. And so the search is on for more markers and more software to separate the signal from the noise.

WritePrint which is embedded into Jstylo (earlier post) has about 800 different variables it creates, and used to be considered the gold standard.

Another clever method used with success in a stylometry competition was by the team of Koppel, Akiva and Dagan with their "Unstable" words as markers:

The JonBenet Ramsey Ransom Note:

Looking at the JonBenet ransom note, means that using content words would fail. In other words, pronouns probably need to be ignored, and content words cannot be used because all ransom notes bear similarities along these lines.

One ransom note would be linked to another if you used word frequencies of "you" and "money" and "die", for example. Since the JonBenet is staged or faked (she was dead when the note was written, the note was purported to be from a "faction"), it is likely that there would be red herrings in the writing in order to attribute it to a radical group.

Any spelling mistakes, hyphens and strange letter formations etc would be obvious and probably useless as markers because the writer knew the note would be analysed, and keeping in mind the dynamics of staging, you would expect conscious errors/red herrings etc.

What we need to do is look for unconscious style markers and text structure, things that are written as habit. It is likely that just as the handwriting experts noted that the last part of the note was the most fluid, it is also likely that the last part also has the most unconscious markers due to force of habit...concentrating on staging a note in the beginning, and it becoming more "free flowing" with habit taking over at the end.

It is also likely that if the crime was covered up by the parents after the son accidentally hit JonBenet on the head with a torch in a fit of rage for snatching some pineapple from him in a midnight snack as per the CBS show (which seems to line up the evidence as the most likely scenario), it would be natural to think that both parents are involved to some extent, one dictating some text or ideas, the other writing.

People write differently to how they talk, and use different parts of the brain to process written text and verbal, so one of the parents would be dominating in their unconscious writing style unless the letter was being quoted verbatim (unlikely.)

Parts-Of Speech Analysis:

The idea is to take away the words, leaving the lexical structure of the ransom note.

This is easily done with the Stanford Parser, and also the Stanford Tagger, both in Java and I have also used the MontyLingua Tagger written in Python.

What a Speech Tagger does is replace words with parts of speech lexical categories such as Verbs, Nouns, Pronouns, Determiners etc. The most used Tags are the Penn Tree Bank of tags, of which there are 36:

This means every word in language automatically gets tagged with one of the above parts of speech tags. There are 6 different Verbs, and depending on the context of the writing, it gets it's assigned Tag from this list.

As an example, lets look at a snippet of text from the ransom note using the word "hence", and one of Patsy's notes with the word "hence" and tag them:

1 /NN of/IN eternal/JJ life/NN and/CC hence/RB ,/, no/DT hope/NN 

2 /NN of/IN the/DT money/NN and/CC hence/RB a/DT earlier/JJR delivery/NN 

The top line tells us there is a Noun followed by a Preposition and then an Adjective in the Patsy note at the top, and the ransom note below is slightly different but the lexical structure is very similar. The actual words are followed by a slash and then a tag by the parser.

Looking at the ransom note now, and deleting all the words, only keeping the parts of speech tags, it looks like this:



 This is the ransom note with all the words and content deleted, leaving only the Penn Tree Bank Tags such as Nouns and Adjectives. So we have minimised the text to it's basic lexical structure of 36 tags.

We do this with all of Patsy notes, about 15 000 words, and John Ramsey's letter of 10 000 words. We also add in two genuine ransom notes, the short Robert Wiles notes and the very long 982 word ransom note from the Barbara Mackle kidnapping.

Running all the POS TAGS in R using the brilliant Stylo R Package and running the Consensus Bootstrap Tree, we get this output:

Using NO words, only parts of speech, the POS structure of one of Patsy's notes is similar to the ransom note, while the other ransom notes get binned together as being similar,and the two Christmas notes get put together too.

Using a clustering algorithm, where the closest most similar to clumped together, this dendogram is produced on the twenty most frequent POS tags:

This lumps Patsy with the ransom note, her other notes similar to John, and the real kidnapping notes from Wiles and Mackle are on the outskirts of Patsy and the JonBenet ransom note.

Now, asking the software to classify who wrote the note, or more accurately, who is the closest match and using one of the most best classifiers proven to have a good track record in authorship, the SVM classifier, Patsy is determined to be the author.

Using one of the most recent and powerful algorithms in  determining the distance ie the closeness of match is the Burrows Delta, which is included in the package, as well as modifications such as Eders Delta and Argamons Delta....the output is again Patsy as the author.

Is there a way to get more linguistic structure out of the writing ie more information than POS Tags can give us?

Yes there is. This brings us to:

Syntactic Ngrams

Part 1 - Parsing Text To Create A Dependency Tree:

Recall, POS Tags (above) give us lexical structure, a word is replaced with a verb or noun tag, but tells us nothing about the syntactic dependency tree structure; telling us what is the subject and object of the sentence is, which word is at the head (root) of the tree and so on.

We are now going to extract syntactic information. This is very different to POS Tags/ Parts Of Speech.

What we extract with syntactic parsing is the tree structure of a sentence -- which word is the object, which word is dependant on another, and to create a tree structure that is non linear. This means the words in a sentence are not listed by the parser in the order they are written, but in the order assessed to be syntactically correct according to a dependency tree.

The critical take away point from this is that syntactic structure is NON LINEAR, meaning the order of the sentence from the parser is different to how it was written. The state of the art Stanford Parser has an accuracy of about 97% and reveals reveals the syntactic structure of text without words, as a first step!

An example of  the parser output for the sentence:

The boy with the brown eyes ate the cake.

det(boy-2, The-1)
nsubj(ate-7, boy-2)
case(eyes-6, with-3)
det(eyes-6, the-4)
amod(eyes-6, brown-5)
nmod(boy-2, eyes-6)
root(ROOT-0, ate-7)
det(cake-9, the-8)
dobj(ate-7, cake-9)

Root is at the top of the tree, above that is a noun modifier, and brown at -5 (5th word) is dependent on eyes at -6. There are around 50 tags from the dependency parser, such as determiners, noun subjects etc.

Onwards now to:

Part 2- Ngrams

Ngrams have been used for a long time and are one of the most reliable indicators of authorship (Sidorov 2014). Ngrams can be characters or words. You can think of it as a sliding window:

Using the above sentence again which comes from Google powerpoint presentation about their ngrams:

The boy with the brown eyes ate the cake.

A bigram or 2 unit ngram is a 2 word sliding window:

The Boy, Boy With, With The, The Brown and so on.

A trigram is 3 words or character unit (word in our example) and goes like this:
The Boy With, Boy With The, The Brown Eyes, Brown Eyes Ate and so on.

Two to five ngram units are the most useful in authorship (Sidorov).

Part 3 - Syntactic Ngrams

The final piece to this puzzle is the syntactic ngram. Google has used them to index several million books and 320 billion ngrams, with it's ngram viewer:

This is a simplistic interface though, and can only be used for frequencies, however there is more sophisticated analysis possible by downloading the Google ngram data.

Notice a problem in the last trigram string above:

Brown Eyes Ate

This is obviously misleading and won't help with the text analysis of that sentence ie the subject is missing. You never get this output when you use syntactic ngrams, so they are far more powerful, contain more information and are more relevant to the text being analysed!

And once again, the beauty with syntactic ngrams is that they are non linear, they contain structure information in a different order according to the parser tree.

As mentioned, this example is from a Google presentation as they explain the purpose of their ngram viewer.

But there is more power in these little guys yet!
Thanks to Dr. Gregori Sidorov, we can produce mixed syntactic ngrams which he calls sngrams--you can mix the syntactic tags from the parser with POS tags (above) or words or lemma.

You now have mixed sngrams, or sngrams with relations, which he calls snrgrams.

He has a site and software in Python to create various sngrams in different sizes along with some interesting papers:

The take away point from this is that text goes into the Stanford Parser, that output from that goes into the sngram software, the output from that is sngrams or snrgrams (if you mixed them) of various sizes ie bigrams trigrams etc.

Long story short--these snrgrams have been shown the be the most powerful use of ngrams in various applications!

The JonBenet ransom note is coded as a 2 unit SNRGRAM (bigram) with Syntactic tags and POS tags.

WE are using the power of syntactic tags and syntactic POS tags containing more linguistic structure information than ever.

The output of the ransom note looks like this:

root[RB] nmod[IN] root[NNS] root[VBP] acl:relcl[NN] dobj[DT]

acl:relcl[WDT] root[PRP] dobj[JJ] root[DT] nmod[VBP] ccomp[IN]
ccomp[PRP] root[NNS] dobj[PRP$] cc[CC] root[VBZ] root[PRP] conj[DT]
dobj[RB] dobj[NN] nmod[IN] dobj[PRP$] nmod[DT] nmod[PRP$] root[NN]
root[PRP] conj[VBP] dobj[PRP$] advcl[IN] advcl[VB] advcl[PRP] conj[NNS]
nmod[DT] conj[PRP] root[JJ] conj[NN] nmod[TO] xcomp[TO] root[VBZ]
root[CC] root[PRP] root[VB] conj[MD] xcomp[CD] nmod[IN] dobj[$] root[CD]
nmod[PRP$] root[NN] root[PRP] root[MD] nmod[IN] nmod[$] conj[JJ]
conj[NNS] root[CD] nsubj[$] root[$] root[IN] root[CC] conj[DT] root[VB]
conj[$] amod[CD] root[MD] ccomp[IN] ccomp[PRP] root[VBP] dobj[JJ]
nmod[DT] dobj[DT] root[JJ] nmod[TO] ccomp[NN] dobj[NN] nmod[IN]
root[VBP] advcl[PRP] nmod[DT] advcl[NN] nmod[NN] root[NN] root[PRP]
nmod[JJ] advcl[WRB] root[MD] dobj[DT] dobj[NN] nmod[CC] nmod[IN]
nmod:tmod[RB] advcl[PRP] advcl[NN] root[NN] root[PRP] advcl[TO] root[VB]
dobj[CD] root[MD] nmod[CD] xcomp[VB] root[VBP] advcl[JJ] advcl[PRP]
advcl[IN] xcomp[TO] nsubj[DT] root[NN] root[VB] root[MD] xcomp[VB]
xcomp[PRP] conj[RB] advcl[VBG] root[JJ] nmod[PRP$] advcl[IN] dobj[JJR]
conj[DT] nmod[DT] root[PRP] root[MD] root[VBP] advcl[PRP] dobj[DT]
dep[PRP] xcomp[TO] dep[RB] nmod[IN] conj[NN] dep[NN] conj[JJR] dobj[CC]
xcomp[NN] dobj[NN] nmod[IN] nsubj[NNS] nmod[DT] nmod[PRP$] nmod[NN]
nsubj[DT] root[NN] nmod[JJ] root[MD] nmod[IN] root[NNS] dobj[PRP$]
root[NN] root[PRP] root[VB] nmod[JJ] root[RB] root[MD] xcomp[PRP]
root[NNS] dobj[PRP$] advcl[VB] advcl[PRP] root[RB] root[VBP] xcomp[RB]
acl[RP] xcomp[TO] nsubj[DT] root[PRP] advcl[IN] nsubj[VBG] nsubj[CD]
acl[NN] nmod[IN] nmod[VBN] nmod[FW] nmod[NNS] root[VBG] case[IN]
nmod[PRP$] nmod[TO] nmod[NNP] nmod[NN] root[NN] csubj[NN] nmod[JJ]
root[MD] acl[VBG] root[VBP] advcl[PRP] nmod[DT] advcl[VBG] dep[PRP]
nmod[TO] dep[NN] root[PRP] advcl[IN] nmod[JJ] advcl[PRP] root[VB]
advcl[NNS] root[PRP] advcl[IN] dobj[NN] advcl[VBN] acl[VBN] advcl[DT]
acl[IN] advcl[NN] advcl[VBZ] acl[CC] nsubj[DT] root[NN] root[PRP]
advcl[IN] nmod[IN] root[NNS] advcl[DT] advcl[IN] conj[PRP] root[VBZ]
root[CC] root[PRP] root[VB] nmod[JJ] root[MD] advcl[VBP] conj[VBN]
ccomp[IN] xcomp[PRP] conj[VB] ccomp[NNS] conj[JJ] nmod[IN] nmod[NNS]
ccomp[PRP] nmod[CC] nmod[NN] root[MD] xcomp[TO] root[CC] root[PRP]
ccomp[VBP] root[VB] nmod[NNP] root[VBN] xcomp[PRP] acl[NN] dobj[PRP$]
advcl[VB] advcl[PRP] xcomp[RB] dobj[DT] acl[IN] xcomp[TO] root[NN]
root[PRP] advcl[IN] dobj[NN] amod[CD] acl[VBP] dobj[VBG] root[NNS]
root[VBP] conj[PRP] dobj[DT] acl[IN] conj[NN] acl[NN] root[CC]
dobj[PRP$] dobj[VBG] amod[CD] dobj[NN] root[NNS] root[VBP] cc[RB]
nsubj[CC] root[JJ] root[IN] conj[PRP$] cc[IN] root[PRP] conj[DT]
nsubj[NN] root[RB] dobj[DT] xcomp[TO] root[VB] xcomp[NNP] root[RB]
dobj[NN] ccomp[IN] acl:relcl[VBP] root[VBP] root[RB] advmod[IN] root[JJ]
acl:relcl[JJ] ccomp[MD] ccomp[NN] root[PRP] amod[RB] root[VB] ccomp[VB]
root[DT] acl:relcl[RB] root[VB] xcomp[PRP] root[RB] root[NNP] nmod[IN]
dobj[NNP] dobj[DT] root[NN] dobj[JJ] advmod[RB] advmod[PRP] nmod[TO]
root[VBZ] root[PRP] root[IN] root[RB]

Again, there are no words here, just ngrams with syntactic structure that is NON LINEAR, not in the same order as written.

Doing this for all the text as before and using the Stylo R Package software gave the following results...

Using the single word analysis in Stylo  with various occurrences of the most frequent sngrams, was nearly the same as using only 2 characters from the sngrams, which was nearly the same as using 4 characters sngram combinations with frequencies up to 500 most used sngram character combinations--they all redflagged Patsy as the most likely author!

In other words, this was the most stable output of any analysis I have done over a whole range of settings, showing that the sngrams contain my relevant syntactic information, despite the lack of words!

As a final note, I should mention that I used the sngrams as input into Jstylo, the authorship attribution software from Drexel University, and just like the results above, increased the probability of Patsy being the author. Using the same Enron Corpus etc from my earlier post, the sngrams increased the likelihood of Patsy being the author.

Let me know if you have any questions. 

A project I have in the pipeline is use sngrams for lie detection in written statements.

Coming soon! 
© ElasticTruth

This site uses cookies from Google to deliver its services - Click here for information.

Professional Blog Designs by pipdig