The bottom-up approach, also known as the data-driven approach, to offender profiling builds a picture of an unknown offender from the evidence they leave behind. Psychologists analyse patterns in behaviour, victim interaction, timing and location to generate evidence-based investigative hypotheses.

What is the Data-Driven Approach?

The data-driven approach begins with what investigators actually know about a crime.

They may examine:

• what the offender did;
• how the victim was approached;
• what the offender said;
• what evidence was removed or left behind;
• when offences occurred;
• where offences occurred;
• similarities between separate offences.

Investigators then search for meaningful patterns.

The basic process is:

Crime evidence → behavioural patterns → statistical analysis → offender hypotheses → investigative priorities

This makes the approach bottom-up because investigators work upwards from individual observations rather than beginning with a predetermined offender category.

This contrasts with the typology approach, which begins by deciding whether the crime resembles an existing type.

---

Investigative Psychology

Investigative psychology is particularly associated with British psychologist David Canter.

Canter wanted offender profiling to move away from intuition and towards psychology as an empirical science.

Rather than asking:

“What sort of person does this crime remind me of?”

investigative psychology asks:

“What conclusions can reasonably be supported by the behavioural evidence?”

Three important areas are:

• behavioural consistency;
• interpersonal coherence;
• the significance of time and place.

---

Behavioural Consistency

Behavioural consistency suggests that offenders may repeat aspects of their behaviour.

For example, a serial offender might repeatedly:

• approach victims in a particular way;
• use similar language;
• select similar victims;
• use similar methods of control;
• commit offences at similar times;
• take similar precautions;
• leave or remove particular objects.

These similarities may allow investigators to establish case linkage.

For example, police may initially believe that four offences have different perpetrators. However, analysis may reveal an unusual combination of behaviours across all four offences.

This does not prove that one person committed them. Nevertheless, it gives investigators a hypothesis that they can test against forensic evidence and other information.

---

Interpersonal Coherence

Canter proposed that the relationship an offender establishes with a victim can provide information about how that person normally interacts with others.

For example, imagine two offenders.

One offender:

• dominates the victim;
• issues commands;
• ignores distress;
• avoids conversation.

Another:

• communicates extensively;
• responds to requests;
• attempts to create cooperation;
• displays occasional concern.

Investigative psychologists may consider whether these behaviours reflect different interpersonal styles outside the crime.

This is known as interpersonal coherence.

However, this is not the same as saying personality remains perfectly consistent. Crime occurs in unusual circumstances, and behaviour can change according to the victim, situation, intoxication, stress or perceived risk.

---

Time and Place

The location and timing of offences can reveal information about an offender’s lifestyle.

For example, repeated offences:

• near railway stations;
• immediately after working hours;
• along a particular motorway;
• during weekday mornings;
• within walking distance of one neighbourhood

could reveal information about transport, employment or routine activity.

Therefore, investigators can combine psychological evidence with geographical evidence.

---

Forensic Awareness

Investigative psychologists may also consider how much effort an offender makes to avoid detection.

Examples include:

• using gloves;
• destroying evidence;
• changing methods following publicity;
• removing objects from a scene;
• avoiding CCTV;
• attempting to prevent identification.

High forensic awareness could indicate previous contact with the justice system or knowledge gained elsewhere.

However, investigators should remain cautious. Modern media means that someone can understand forensic techniques without any previous offending experience.

---

Canter and Heritage (1990)

Researcher and year

David Canter and Rupert Heritage, 1990.

Aim

To investigate whether patterns of behaviour could be identified across sexual offences and whether these patterns could contribute to scientific offender profiling.

Procedure / Method

Canter and Heritage analysed 66 sexual assaults committed by 27 offenders.

They identified 33 behavioural variables from victims’ statements.

These included features such as:

• the method of approach;
• language;
• restraints;
• use of weapons;
• victim interaction;
• sexual behaviour.

They then used Smallest Space Analysis to examine which behaviours tended to occur together.

Findings

Some behaviours occurred frequently across many offences, while different offenders displayed different combinations of less common behaviours.

The analysis therefore identified meaningful relationships between crime-scene behaviours.

Conclusion

Offence behaviour can be analysed systematically.

Therefore, investigators may be able to use statistical patterns to:

• differentiate offences;
• identify behavioural consistency;
• link crimes;
• generate hypotheses about offenders.

The study supported Canter’s argument that profiling could become more scientific rather than depending largely on intuition.

---

Smallest Space Analysis

Smallest Space Analysis, or SSA, is a type of multidimensional statistical analysis.

Imagine researchers identify 30 behaviours across hundreds of offences.

The software calculates how frequently different behaviours occur together.

If two behaviours frequently appear in the same crimes, they appear close together on the resulting statistical map.

If they rarely occur together, they appear further apart.

This can reveal clusters of behaviour that investigators might not notice simply by reading case files.

Importantly, SSA does not produce the name of an offender.

Instead, it helps investigators identify patterns.

---

Geographical Profiling

Geographical profiling examines where crimes happen.

Investigators may plot:

• where victims were approached;
• crime scenes;
• body disposal locations;
• abandoned property;
• escape routes;
• vehicle locations.

They then analyse relationships between these locations.

The underlying assumption is that offenders do not normally select locations completely randomly.

Instead, their choices may reflect:

• familiarity;
• opportunity;
• transport;
• perceived risk;
• daily routines.

---

Mental Maps

People develop mental maps of places they regularly encounter.

An offender’s mental map might contain:

• their home;
• workplace;
• previous addresses;
• relatives’ homes;
• shops;
• leisure locations;
• railway stations;
• frequently travelled roads.

Offenders may therefore choose locations from areas with which they already feel comfortable.

This creates spatial consistency.

---

The Least-Effort Principle

People generally prefer to achieve goals without unnecessary effort.

Applied to offending, this suggests that an offender may prefer suitable targets closer to familiar locations rather than travelling much further without a reason.

However, offenders do not necessarily choose the location closest to home because committing a crime immediately outside their house creates obvious risks.

---

Buffer Zones

An offender may deliberately avoid committing crimes extremely close to their home.

This produces a buffer zone.

Recognition becomes more likely close to home because:

• neighbours may recognise the offender;
• victims may know them;
• witnesses may identify their vehicle;
• CCTV may connect movements to their address.

Therefore, the highest concentration of offences may occur some distance away from the offender’s actual home.

---

Distance Decay

The principle of distance decay proposes that offending generally becomes less likely as distance from the offender’s base increases.

A simplified pattern could therefore be:

Home → buffer zone → high-probability crime area → decreasing probability with greater distance

The real pattern will be affected by environmental factors such as:

• roads;
• rivers;
• public transport;
• neighbourhood boundaries;
• workplaces.

---

Marauders and Commuters

Marauders

A marauder commits offences outward from a home or operational base.

Their base therefore lies within the general area formed by their crime locations.

For example:

Crime — Crime — HOME — Crime — Crime

If several offences form a cluster, investigators might prioritise suspects who live within that region.

---

Commuters

A commuter travels away from their normal home area before beginning to offend.

For example:

HOME → travel → Crime — Crime — Crime

In this case, searching the centre of the crime cluster may not reveal the offender’s home.

The marauder-commuter distinction therefore prevents geographical profiling from assuming that all offenders behave in exactly the same way.

---

Canter and Larkin (1993)

Researchers and year

David Canter and Paul Larkin, 1993.

Aim

To investigate geographical patterns among serial sexual offenders and test the distinction between marauders and commuters.

Procedure / Method

Canter and Larkin examined the spatial activity of 45 serial sexual offenders.

Each offender had committed at least two offences.

Researchers plotted crime locations in relation to each offender’s known home address and examined whether the offender’s base fell within their geographical offence region.

Findings

The majority showed a pattern consistent with marauding.

This meant that their known home location fell within or close to the overall region formed by their offences.

Conclusion

Serial offenders can show meaningful geographical consistency.

Therefore, geographical profiling may help investigators identify areas where an offender lives or regularly operates.

---

Case Study 1: John Duffy – The Railway Offender Investigation

One of the most famous applications of Canter’s developing approach concerned John Duffy.

During the 1980s, police investigated a series of serious sexual offences and murders connected with railway locations around London.

David Canter was invited to examine the evidence.

He analysed:

• crime locations;
• railway connections;
• patterns across offences;
• interactions with victims;
• likely familiarity with transport;
• behavioural characteristics.

Canter predicted that the offender probably lived in the Kilburn area, had knowledge of the railway system, was relatively young, had relationship difficulties and possessed several other behavioural characteristics.

Duffy did live in Kilburn and had worked as a British Rail carpenter. Contemporary accounts also describe how the geographical pattern helped direct attention towards the area of his earliest offences.

Why is the Duffy case important?

The investigation illustrates how:

behavioural evidence + geographical evidence + conventional police investigation

can work together.

However, profiling did not independently establish guilt.

Investigators still needed traditional evidence.

The case therefore provides a valuable reminder that offender profiling should prioritise enquiries rather than replace evidence.

---

Case Study 2: Adrian Babb

A second case associated with Canter involved a series of attacks against elderly women in Birmingham tower blocks between 1986 and 1988.

Investigators observed recurring features across the offences.

Canter considered information including:

• the locations;
• victim characteristics;
• the offender’s apparent physical fitness;
• his behaviour towards victims;
• his apparent lack of fear of recognition.

In one particularly interesting behavioural detail, the offender responded when a victim complained of physical discomfort.

Canter suggested that this behaviour could indicate someone who might normally appear helpful or considerate.

He also predicted characteristics including an interest in a solitary physical activity and regular contact with elderly people.

The offender, Adrian Babb, was eventually identified. He was a 20-year-old swimming-pool attendant and was linked to the investigation through fingerprint evidence before pleading guilty to seven rapes.

Why is this case important?

The case illustrates interpersonal coherence.

Apparently minor details of victim interaction potentially provided useful information about the offender’s everyday behaviour.

However, once again, the profile did not prove guilt.

Conventional forensic evidence was vital.

---

Reflection Question

Imagine a profile suggests that an offender is a male aged 20–30 who lives within three kilometres of the crimes and works irregular hours.

How should police use that information without unfairly treating every person who matches those characteristics as a suspect?

This highlights the most important principle of offender profiling: a profile is an investigative hypothesis, not evidence of guilt.

Evaluation

Strengths

One strength of the data-driven approach is its scientific basis. Investigative psychology attempts to replace intuition with measurable behavioural evidence and statistical analysis. For example, Canter and Heritage analysed 66 offences using 33 behavioural variables and Smallest Space Analysis. Therefore, relationships between behaviours can be identified systematically rather than because a profiler thinks two crimes simply “feel” similar. This increases objectivity and makes data-driven profiling more consistent with psychology as a science.

Another strength is research support for geographical profiling. Canter and Larkin examined 45 serial sexual offenders and found strong evidence for meaningful spatial patterns, particularly marauding behaviour. This supports the assumption that offenders often operate within geographical areas connected with their everyday lives. Therefore, geographical profiling can help investigators prioritise search locations rather than search an entire region.

The approach also has useful real-world applications. The John Duffy investigation demonstrates how geographical and behavioural information can contribute to police decision-making. Canter correctly identified several relevant characteristics and geographical features of Duffy’s behaviour. Therefore, profiling does not need to produce a person’s name to be useful; narrowing thousands of possibilities can still save investigative resources.

Investigative psychology can contribute to case linkage. Linking offences allows police forces to combine witness statements, geographical information and forensic evidence that might otherwise remain separated. Canter and Heritage showed that recurring combinations of behaviours can be statistically examined. As a result, data-driven profiling may reveal patterns that investigators would struggle to identify when looking at individual cases separately.

Finally, the approach avoids rigid offender categories. Human behaviour is complex, and offenders do not necessarily fit neatly into predetermined types. A data-driven profile can develop as new evidence appears. Therefore, the approach is more flexible and can accommodate offenders who show unusual combinations of behaviours.

---

Weaknesses

A major weakness is the problem of database bias. Investigative psychology often develops knowledge from solved offences because researchers need to know who committed the crime before connecting offender characteristics with behaviour. However, offenders in unsolved cases may behave differently. Therefore, databases may overrepresent offenders whose behaviour made them easier to identify, reducing the validity of predictions made about more difficult cases.

Sample bias also limits generalisation. Classic geographical research often focused on male serial sexual offenders. For example, Canter and Larkin studied 45 serial sexual offenders. Their geographical patterns may not apply equally to female offenders, younger offenders or people committing offences such as fraud or cybercrime. Therefore, researchers should avoid assuming one geographical model applies universally.

There may also be cultural bias. Geographical behaviour depends on transport, housing, employment and urban design. Patterns found among offenders in British cities may differ from those living in rural areas or countries where people routinely travel much greater distances. Consequently, geographical findings require local context rather than simple universal rules.

A further weakness is that offender behaviour may not remain consistent. Offenders can learn from experience, react to victims differently or deliberately change their methods following police publicity. An offender may also move home or change employment. Therefore, behavioural consistency should be treated as probabilistic rather than assumed to be perfect.

Geographical profiling may oversimplify complicated environments. A circle drawn around crimes cannot capture every factor affecting movement. Rivers, roads, railway lines, workplaces and social boundaries can all distort offending patterns. A mathematical centre of gravity should therefore never be interpreted as the offender’s exact home address.

There are also ethical concerns. A profile might describe characteristics shared by thousands of innocent people. Investigators who become overly confident in a profile could develop confirmation bias and concentrate on people who fit expectations while ignoring contradictory evidence. Therefore, profiling should inform police enquiries rather than determine guilt.

Temporal validity presents another problem. Much classic offender-profiling research predates widespread CCTV, smartphones, automatic number-plate recognition, digital mapping and modern forensic databases. These technologies may change how offenders choose locations and avoid detection. Consequently, older findings remain theoretically useful but should continue to be tested with contemporary crime data.

---

Quick Questions

What is meant by the data-driven approach to offender profiling?

A profiling approach that builds hypotheses from behavioural, geographical and other evidence from a crime rather than beginning with fixed offender categories.

What is interpersonal coherence?

The idea that an offender’s interactions with victims may reflect aspects of their everyday interpersonal behaviour.

What is Smallest Space Analysis used to identify?

Statistical relationships between behaviours and which behaviours tend to occur together.

What is the difference between a marauder and a commuter?

A marauder operates around a base within the crime region, whereas a commuter travels away from their usual area before offending.

What did Canter and Heritage investigate?

They analysed 66 sexual assaults committed by 27 offenders to identify patterns among 33 behavioural variables.

---

Key Terms

Offender profiling – Using information about an offence to make informed predictions about an unknown offender.

Data-driven approach – The current AQA term for bottom-up profiling. Investigators build a profile from the evidence rather than starting with an offender category.

Bottom-up approach – The traditional name for a profiling approach that works from individual crime evidence towards broader conclusions.

Investigative psychology – The scientific application of psychological theory and research to criminal investigation.

Interpersonal coherence – The idea that the way an offender interacts with a victim may reflect how they interact with people in everyday life.

Behavioural consistency – The idea that an offender may display similar patterns of behaviour across different offences.

Case linkage – Determining whether two or more offences are likely to have been committed by the same person.

Smallest Space Analysis (SSA) – A statistical technique that shows relationships between behaviours. Behaviours that commonly occur together appear closer together.

Geographical profiling – Analysing the locations of linked crimes to make predictions about an offender’s likely base or areas of activity.

Crime mapping – Plotting crime locations to identify spatial patterns.

Mental map – An individual’s internal representation of familiar locations, routes and environments.

Spatial consistency – The tendency for an offender to show predictable geographical patterns across offences.

Marauder – An offender who commits crimes around a home or operational base located within the general crime area.

Commuter – An offender who travels away from their home area before committing offences.

Circle hypothesis – The proposal that an offender’s home may lie within a circle based around their crime locations.

Distance decay – The tendency for offending to become less likely as distance from an offender’s base increases.

Buffer zone – An area immediately around an offender’s home where offending may be less likely because of the risk of recognition.

Forensic awareness – Behaviour suggesting that an offender understands investigative or forensic procedures and attempts to avoid leaving evidence.