Week 5 Flashcards
What is investigative psychology?
A field of applied psychology developed by David Canter to make police psychology and investigation more scientific and academically rigorous instead of subjective
Who is David Canter? (3)
Father of IP
Background as an environmental psychologist, looking at how space impacts our behavior
Got involved in policing when the police reached out to him for his environmental work during the Railway Rapist investigation
He created a geographic profile that assisted in finding the offender and started the field
What are the core features of IP? (4)
A scientific approach to assisting police
Utilizes research methods to handle police data
Attempts to understand both offenders and police
Develop practical strategies beyond inference and guess based prediction
What are some of the applications of IP? (6)
Linking serial crimes
Crime scene analysis (criminal profiling)
Geographic profiling
Truth verification
Hostage negotiation
Suicide note veracity
What is crime linkage analysis? (2)
Reliance on behavioral evidence since physical evidence is often not available
Techniques that rely on behaviors to link crimes is behavioral crime linkage analysis
Why is crime linkage important? (2)
Allows you to pool evidence and resources, creating a greater likelihood of solving crimes
Allows you to obtain more sever sentences for more offenses linked to one offender
What are the 2 different types of crime linkage scenarios? What are examples of each?
Comparative case analysis, where large samples of crimes are searched by a practitioner for potential linkages
A practitioner is given an index crime that they think isn’t a one off so they search a database for other behaviorally similar crimes
Crime linkage analysis, where limited crimes are brought to a practitioner who needs to determine whether the crimes are linked
A practitioner is presented with a pre-defined set of crimes and they must determine whether they are linked
What are the 2 assumptions underlying all linking tasks? Which assumption receives more focus?
Behavioral stability (to some degree, offenders will display similar behaviors across their crimes or there would be no reason to believe they could all be linked)
Behavioral distinctiveness (to some degree, offenders committing similar crimes will commit those crimes in a distinctive fashion so they cannot all be linked)
Detectives use stability more, which is why they are so bad because if there is no distinctiveness for a crime (B&E through a window, which is most), crimes will be linked without actually being done by the same offender
What is the main theory behind crime linkage? What does it examine? What is the problem with this?
Personality psychology
Examines behavioral consistency and how people consistently exhibit individual behavioral differences over situations due to internal predispositions and traits we possess
Traits are not the primary determinant of behavior but different situational factors can make us change how we act (how the victim acts)
What is CAPS and how does it relate to crime linkage? (5)
Cognitive-affective personality systems
Assumes that personality traits and situational factors influence our behavior, mediated by cognitive (thoughts) and affect (feelings)
Behavioral strategies are activated by internal triggers (traits, fantasy) and external triggers (victim, environment) across crimes
Strategies will be consistent because the same strategies are activated across similar situations (especially if the strategies work to commit the crime successfully)
Distinctiveness in strategies results from the fact that offenders understand/process information/situations differently based on learning theory
What are 4 linkage approaches?
Modus operandi based procedures
Behavioral signatures
Linkage databases (MO and signatures)
Linkage as diagnosis (MO based procedures)
What is modus operandi? (3) What is the challenge?
The method used by an offender to successfully commit a crime and get away with it
It is the oldest approach used to like serial crimes
Assumes offenders will exhibit the same behaviors each time they commit their crimes
An MO is not always stable, offenders change behaviors for a variety of reasons (situational factors, maturation)
What are behavioral signatures? (3) what are the challenges? (2)
Behaviors that are exhibited, not because they are necessary for crime commission, but because they satisfy some psychological/emotional need
Not necessary, more unique and for a reason other than committing the crime
It could even impede the ability to get away with the crime
Limited empirical research on behavioral signatures in terms of definition and recognition
Still susceptible to environmental factors
What are linkage databases? (5) what are the challenges? (3)
Databases like ViCLAS (RCMP) and ViCAP were developed to reduce linkage blindness
They capture information related to crimes committed across jurisdictions
Specially trained analyst attempt to identify linkages bracket (in Canada, this is done at the Canadian police college)
This allows analysts to compare many crimes using range of searchable linkage indicators, with them completely in control over linking
Allows for a more human subjective analysis, rather than a rigid AI based one that cannot pick up on human differences and behavior
Lack of research on data, reliability and accuracy, consistency and distinctiveness of MO and signatures is not defined, and the ability of analyst to identified potential linkages is inconsistent based on the analyst
What is linkage as a diagnosis task? (3) What are the outcomes? (4)
Treats linkage diagnostic tasks as we do every other diagnostic task so we can compare them to other studies on similar tasks, which use the same framework as there’s been a lot of research on those tasks (cancer, weather)
Views linkage as a two alternative yes- no diagnostic task to discriminate between linked and unlinked crimes using evidence
The goal of every diagnostic task is to maximize the correct responses and minimize the wrong responses
Hit (true positive where it was predicted to be linked, and it was the same offender)
Miss (false negative where it was predicted to be unlinked, but it was the same offender)
False alarm (false positive where it was predicted to be linked, but it was different offenders)
Correct rejection (true negative where it was predicted to be unlinked, and it was different offenders)
What is the best evidence for crime linkage?
Unambiguous (occurs only for linked crimes or unlinked crimes)
Accurately and reliably coded
What does an ideal evidence distribution look like? (3)
DNA for example is almost completely completely unambiguous
When comparing the distribution of correct rejections for DNA crime linkage and hits for DNA crime linkage, there will be virtually no overlap between the distributions with only a small overlap of false alarms
Basically, if you use the proper decision threshold where the distributions meet, you will almost always be right
What does a behavioral linkage evidence distribution look like? (3)
Behavioral similarity is less unambiguous, because behavior is harder to define and measure consistently
This means that there is more overlap between the distributions
It is more likely to have false alarms and misses even if using the decision threshold
How do you define useful evidence? (2)
Usefulness is defined by the degree of overlap between these distributions, meaning less overlap equals better evidence
The degree of overlap can be measured using receiver operating characteristic (ROC) analysis
What is ROC Analysis? (4)
Plot hit and false alarm rates as a function of similarity threshold used to link crimes
Basically determines what level of similarity needs to be present for crimes to be considered linked to maximize the accuracy of the decision
On a graph, linking accuracy is measured as the area under the ROC curve (AUC)
The higher the AUC and the closer to 1 (highest coefficient), the better the evidence is for linking
How good are we at linking crimes? (3)
It is hard to tell because studies have used different measures of linking accuracy
The question becomes do differences across studies represent differences and accuracy or differences in accuracy metrics?
Bennell et al. (2014) examined studies using the AUC and found that our abilities to link crime is moderate across all conditions, with some conditions leading to excellent abilities
What behaviors are best for linking crimes? (3)
The most useful linking behaviors appear to be those that are determined by the offender, rather than being situationally determined
Where and when offenders commit their crimes seem to be particularly useful indicators
Control behaviors also appear to have some value, at least in some crimes
Is it easier to link certain crimes? (2)
Linking accuracy scores are higher for interpersonal crimes than property crimes
There is a massive variation in AUCs for property crimes
Are people good at linking crimes? (4)
Three studies have examined this issue, and all suggest that linking crimes is difficult
Some evidence suggests that the problem doesn’t lie with identifying useful linking behaviors, but with processing that information to identify accurate linkages (a human cognitive limitation)
Another potential issue is cognitive conceit in professionals, who think they know better than the tools and psychologists (students tend to do better when trained because they do not have that built in conceitedness and confidence in their own abilities and methods)
Actuarial tools like ROC analysis can assist human judges
What are the 4 problems with linking research?
Over-reliance on solved and serial crimes in research when analysts are not looking to link solved crimes (those crimes may be solved because they were so easy to link)
Over-reliance on UK (and US) crimes
Over-reliance on property crimes
Ecological validity (the degree to which research overlaps with the ways thing are done in the real world is very low)
What is criminal profiling? (2) How is it different than RCMP profilers?
Predicting the background characteristics of offenders (personality, behavior, demographic) based on an assessment of crime scene behaviors
Potentially allows police to prioritize suspects by matching possible offender types to suspects
This is different than what RCMP profilers do, which is mainly threat assessment and investigative consultations (basically a free skilled investigator)
Who can become a profiler in different countries? (3)
In Canada, you must be a police officers with specific training and experience
In the UK, civilians and academics make up most profilers
In the US, basically anyone can be a profiler
What caveats are there to our discussions of profilers? (3)
Profilers do a lot more than just profiling
In most jurisdictions, they are not called profilers but Behavioral Investigative Analysts
We are only speaking of the profiling aspect of what they do
What are the 2 assumptions underlying the task of profiling? What is wrong with one of the assumptions?
Behavioral stability
Homology (if two crimes are committed in a similar way, offenders will have similar background characteristics)
There is no evidence of homology at all, some even saying the opposite, and this is usually where profiling tends to go wrong
What is the theory behind criminal profiling? (2)
Rooted in personality psychology
Latent traits are assumed to be the driving force behind behaviors exhibited in criminal and non-criminal contexts
What is the organized-disorganized model? (3)
Crime scenes will vary from organized (well planned) to disorganized (chaotic)
Backgrounds will vary from organized (high functioning with wife, kids, job, car, etc.) to disorganized (low functioning with criminal record, anti social, single, etc.)
Organized crimes mean organized offenders, disorganized crimes mean disorganized offenders
What are the two kinds of profiling approaches? Which is used more often?
Deductive and inductive approaches
Most profilers use both but inductive is at the core of many (like the organized-disorganized FBI model of profiling)
What is deductive profiling? (2)
Profiling an offender’s background characteristics based solely on evidence left at crime scenes
Ex: the victim is a man who had some military experience and his cheek is bruised, suggesting that the offender is a relatively strong man as he was able to overpower the victim with physical force
What is inductive profiling? (2)
Profiling an offender’s background characteristics based on what we know about other offenders who have committed similar (solved) crimes
Ex: the offender in this case exhibited behaviors A, B, C and D, and other offenders who did the same were male, married, with previous convictions
What is the IP approach to profiling? (3)
Determine if structure exists in the way offenders commit crime (observable behaviors)
Determine if structure exists in the background characteristics of offenders (observable characteristics)
Identify relationships between actions and characteristics
What is the arson example of IP? (5)
Used smallest space analysis
Coded behaviors of different arsonist cases
The more often 2 or more behaviors co-occur during the same event, the more clusters of different types of arsonists and their behaviors can be created
Can do the same with background characteristics of arsonists then look at correlations between behavior clusters and background clusters
Correlations tend to be moderate and weak
Are profiling models valid? (2)
Models have been used without evidence of their validity
Research is beginning to question their validity as mostly all are coming back with no or negative correlations
Are assumptions of profiling supported? (2)
Some evidence for stability assumption (but as seen in the linking section, how do we profile characteristics when offenders change their behavior cause they will)
No strong evidence for the homology assumption (different conditions can change behavior which means they won’t link to those characteristics anymore)
Is the content of profiles useful? (4)
Alison et al. (2003) examined the content of 21 profiles
Found only 28% of statements were profiling predictions
Of those 28, most were unsubstantiated, half were unverifiable, a quarter were ambiguous and a few had multiple outs (giving several options to be right)
Things have improved in the UK at least
Do end-users (police) see profiles as useful? (2)
Consumer satisfaction surveys generally show that police view profiling as a useful tool when cautious and appreciating limits
Snook et al. Found this in Canadian police investigators
What are the problems with research on profiling? (5)
Problematic data
Reluctance of profilers to participate
Reluctance to share profiles
Measuring accuracy vs. precision (saying the offender is between 20-40 years old vs. saying they are 32 is very different in terms of value)
Ecological validity (the degree to which research overlaps with the ways these things are done in the real world is very low)
