The Data-Driven Approach | AQA A-Level Psychology Revision

Revision Notes
Aug 7
31 min read

Updated: Aug 14

For 7182 specification, first teach in September 2025

AQA A-Level Psychology | Free Revision Notes

Estimated study time: 65 to 80 minutes

The data-driven approach A-Level Psychology revision examines how investigators use crime-scene information, victim details, psychological theory and statistical analysis to construct a profile of an unknown offender. Rather than placing the offender into a pre-existing category, the profile is built from patterns found within the available evidence.

You will learn about investigative Psychology, behavioural consistency, interpersonal coherence and the use of statistical databases. You will also evaluate whether an approach based on data is necessarily objective, accurate and useful. This lesson follows profiling through organised and disorganised categories [Lesson 1: The typology approach to offender profiling] and prepares you for analysing the geographical distribution of offences [Lesson 3: Geographical profiling].

The current AQA specification requires the data-driven approach to offender profiling, including investigative Psychology. Geographical profiling is also required but is covered separately in the next lesson.

Learning Objectives 🎯

By the end of this revision page, you should be able to:

Define the data-driven approach and investigative Psychology.
Explain why the approach is described as bottom-up.
Explain how crime-scene and victim data are coded and analysed.
Explain interpersonal coherence and behavioural consistency.
Explain the role of statistical analysis, including smallest space analysis.
Apply investigative Psychology to unfamiliar offence information.
Evaluate the approach using evidence, methodological issues and practical considerations.
Compare data-driven profiling with the typology approach.

Revision Notes 📚

The Data-Driven Approach A-Level Psychology Revision Focus

The data-driven approach constructs an offender profile by analysing information from:

The current offence.
The victim.
Previous similar offences.
Databases of solved cases.
Established psychological research.
Statistical relationships between behaviours.

It is commonly described as a bottom-up approach because the profiler begins with the details of the individual case and gradually builds a profile.

The process moves from:

Specific offence data → identification of behavioural patterns → statistical and psychological interpretation → offender profile

This differs from the typology approach, which begins with the organised and disorganised categories and attempts to place the offender into one of them.

Why Is It Called Data-Driven?

The approach is data-driven because conclusions should be based on recorded evidence rather than primarily on intuition or professional judgement.

Relevant data may include:

The offender’s actions.
The sequence of events.
The offender’s interaction with the victim.
The amount of control shown.
Words spoken during the offence.
Property taken or left.
Attempts to conceal evidence.
The time of the offence.
The location.
Characteristics of the victim.
Similarities with other offences.

These details are converted into variables that can be compared across cases.

The purpose is not simply to collect large amounts of information. The data must be analysed to identify meaningful patterns.

Why Is It a Bottom-Up Approach?

The data-driven approach is commonly described as bottom-up because it begins with the current offence rather than a pre-existing offender type.

The profiler asks:

Which behaviours occurred?
Which behaviours tend to appear together?
Is the same pattern present in other offences?
What characteristics are associated with this pattern?
What can the pattern suggest about the unknown offender?

The profile is therefore built from the evidence upwards.

Top-down typology approach	Bottom-up data-driven approach
Begins with organised and disorganised types	Begins with the details of the individual offence
Matches the crime to an existing category	Builds the profile from behavioural patterns
Developed mainly from interviews with serious offenders	Uses databases, psychological theory and statistical analysis
May rely heavily on profiler judgement	Attempts to use systematic analysis
Produces categorical predictions	Produces probabilistic predictions
Traditionally focused on serious violent crime	Can potentially be applied to a broader range of offences

This comparison develops the organised and disorganised offender distinction [Lesson 1: The typology approach to offender profiling].

What Is Investigative Psychology?

Investigative Psychology is the application of psychological theory and research to criminal investigation.

It is associated particularly with David Canter and aims to make offender profiling more systematic and scientific.

Investigative Psychology attempts to answer questions such as:

Are several offences likely to have been committed by the same person?
What does the offender’s interaction with the victim suggest?
How consistent is the offender’s behaviour across offences?
What characteristics might the offender possess?
What knowledge or experience may the offender have?
What can the timing and location of offences suggest?
Which suspects should investigators prioritise?

Investigative Psychology does not rely on one simple offender classification. It examines relationships between multiple pieces of evidence.

Psychological Theory and Statistical Analysis

The data-driven approach combines two important elements:

Psychological theory
Statistical analysis

Psychological theory helps explain why a behavioural pattern might be meaningful.

Statistical analysis helps determine whether:

Behaviours genuinely occur together.
Similar offences form a cluster.
Particular patterns occur more often than chance would suggest.
A relationship found in one case is supported across many cases.

This combination is intended to reduce reliance on intuition.

AQA’s June 2022 mark scheme describes the approach as using crime-scene and victim data to construct statistical databases and analysing the information to identify relationships between offences.

The Data-Driven Profiling Process

A simplified data-driven investigation can be divided into seven stages.

Stage 1: Collect information

Investigators collect available information about:

The crime scene.
The victim.
Witness accounts.
The sequence of behaviour.
The time and location.
Evidence removed or left behind.
Other potentially related offences.

Stage 2: Identify behavioural variables

The offence is broken down into observable features.

For example:

Did the offender speak to the victim?
Was the victim controlled?
Was a weapon brought to the scene?
Was property taken?
Was evidence concealed?
Was the victim moved?
Did the offender apologise or reassure the victim?
Did the offender use unnecessary violence?

Stage 3: Code the variables

Each behaviour is placed into a defined category.

For example:

Behavioural variable	Coding
Offender spoke reassuringly to victim	Present or absent
Weapon brought to scene	Present or absent
Property taken	Present or absent
Victim moved	Present or absent
Evidence concealed	Present or absent

This allows information from different offences to be compared quantitatively.

The process resembles turning behavioural material into coded data [Lesson 8: Content analysis].

Stage 4: Compare cases

The coded offence is compared with:

Other crimes in the same investigation.
Solved offences held in databases.
Patterns identified in psychological research.

Investigators examine whether similar behaviours repeatedly appear together.

Stage 5: Use statistical analysis

Statistical techniques are used to identify:

Clusters of related behaviours.
Distinct behavioural themes.
Relationships between offence features.
Similarities between crime scenes.

Stage 6: Interpret the pattern psychologically

The statistical pattern is interpreted using concepts such as:

Interpersonal coherence.
Behavioural consistency.
Forensic awareness.
Significance of time and place.
The offender’s possible criminal experience.

Stage 7: Construct and test the profile

The resulting profile may suggest:

Social characteristics.
Criminal experience.
Interaction style.
Familiarity with the victim or location.
Likely employment or routines.
Possible links with other offences.

The profile should remain open to revision as new evidence becomes available.

Behavioural Consistency

The data-driven approach assumes some degree of behavioural consistency.

This means that an offender may behave in recognisable ways:

Across different offences.
Towards different victims.
During an offence and in ordinary relationships.
When responding to similar situations.

If an offender repeatedly:

Uses similar language.
Selects similar victims.
Controls victims in the same way.
Takes the same type of property.
Uses a similar sequence of actions.

investigators may infer that the offences are linked.

Behavioural consistency is not expected to be perfect. The approach looks for recurring patterns rather than identical crime scenes.

Behavioural Consistency and Case Linkage

Case linkage is the process of deciding whether several offences were probably committed by the same person.

Investigators may compare:

Victim characteristics.
Method of approach.
Type of threat.
Language used.
Treatment of the victim.
Evidence concealed.
Items taken.
Location and timing.

The greater the behavioural similarity, the stronger the possible link.

However, similarity does not prove that the same person committed the offences. Different offenders may use similar methods, particularly when those methods are common or learned from media coverage.

Modus Operandi

A modus operandi is the practical method used to commit an offence.

It may include:

How entry was gained.
How the victim was approached.
Which weapon was used.
How escape was achieved.
How evidence was concealed.

A modus operandi can change because the offender:

Learns from previous mistakes.
Encounters a different victim.
Gains access to new equipment.
Becomes more confident.
Responds to police publicity.
Commits the offence in a different setting.

Investigators should therefore avoid expecting complete consistency in every practical detail.

Interpersonal Coherence

Interpersonal coherence is the idea that the way an offender interacts with a victim may reflect the way they interact with other people in ordinary life.

For example, an offender may treat the victim as:

A person.
A vehicle for achieving a goal.
An object.
Someone requiring control.
Someone with whom a relationship is being created.

The offender’s behaviour may indicate:

Confidence or anxiety.
Social competence.
A desire for control.
Emotional distance.
Hostility.
Attempts to gain trust.
Awareness of the victim’s feelings.

The profiler uses the interaction to make cautious inferences about the offender’s wider interpersonal style.

Examples of Interpersonal Coherence

Consider two offenders.

Offender A

The offender:

Speaks calmly to the victim.
Uses persuasion.
Attempts to gain trust.
Gives detailed instructions.
Responds to the victim’s questions.

Investigators might infer that the offender is:

Relatively socially competent.
Confident in conversation.
Used to interacting with others.
Able to maintain control without immediate violence.

Offender B

The offender:

Avoids communication.
Treats the victim impersonally.
Uses immediate force.
Appears unable to respond to the victim’s reactions.

Investigators might infer:

Social isolation.
Anxiety during interpersonal contact.
Difficulty understanding others.
Reliance on force rather than communication.

These are hypotheses, not certainties.

Interpersonal Coherence Across Offences

If similar interaction patterns occur in several crimes, this may help:

Link the offences.
Distinguish one offender from another.
Predict how the offender may behave towards future victims.
Suggest an appropriate interview strategy.

For example, an offender who repeatedly attempts to create a personal relationship with victims may behave differently during questioning from an offender who consistently avoids interpersonal contact.

Limitations of Interpersonal Coherence

The offender’s behaviour towards a victim may not reflect ordinary life.

Behaviour during an offence may be influenced by:

Fear of discovery.
Emotional arousal.
Alcohol or drugs.
The victim’s resistance.
The physical setting.
The offender’s immediate goal.
The presence of another offender.
Time pressure.

A socially competent person may behave awkwardly during an offence. A socially isolated person may rehearse a confident approach.

Interpersonal coherence is therefore a probability rather than a guaranteed link.

Forensic Awareness

Forensic awareness refers to knowledge of police investigation and physical evidence.

An offender may demonstrate forensic awareness by:

Wearing gloves.
Removing a weapon.
Cleaning surfaces.
Avoiding cameras.
Concealing the victim.
Destroying evidence.
Changing clothing.
Selecting a place with limited surveillance.
Attempting to prevent identification.

This may suggest that the offender:

Has committed previous offences.
Has been investigated before.
Has knowledge of police procedures.
Has learned from media or other offenders.
Planned the offence carefully.

Forensic Awareness and Criminal Experience

Investigators may use forensic awareness to infer aspects of an offender’s criminal career.

For example:

An offender removes several different types of trace evidence and avoids known surveillance locations.

This could indicate:

Previous contact with the criminal justice system.
Knowledge gained through earlier offending.
Careful planning.
Access to specialist information.

However, forensic awareness does not prove previous offending. Information about evidence collection is widely available through news, films, documentaries and online sources.

Significance of Time

The time at which an offence occurs may provide clues about the offender’s routine.

Investigators may consider:

Day or night.
Weekday or weekend.
Regular intervals.
School or working hours.
Public transport availability.
The victim’s routine.
Periods when a location is empty.

For example:

Several offences occur on weekday afternoons but never during evenings or weekends.

This might suggest that the offender:

Is available during the day.
Works unusual hours.
Is unemployed.
Has access to the location at that time.
Knows when the area is quiet.

These are possible interpretations that must be checked against other evidence.

Significance of Place

The location of an offence may indicate:

Familiarity with the area.
Knowledge of access routes.
Access to transport.
Where the offender lives or works.
Places visited through leisure or family connections.
The amount of risk the offender is willing to take.

Detailed analysis of offence locations is the focus of using location patterns to infer an offender’s base [Lesson 3: Geographical profiling].

In this lesson, the important point is that time and place form part of the database used to understand offender behaviour.

Victim Information

The victim is an important source of data.

Investigators may examine:

Age.
Occupation.
Routine.
Lifestyle.
Relationship status.
Physical characteristics.
Known contact with the offender.
Level of vulnerability.
Why the victim was available at that time and place.

This process is sometimes called victimology.

Victim information may help investigators understand:

Whether the victim was selected deliberately.
Whether the offender had access to particular environments.
Whether several victims share important characteristics.
Whether the offender knew or observed the victim.
What risks the offender was prepared to take.

Care is needed to avoid implying that a victim caused or deserved the offence.

Statistical Databases

A crime database contains information from many offences.

It may include data about:

Crime-scene behaviour.
Victim characteristics.
Time and location.
Offender characteristics in solved cases.
Evidence recovered.
Convictions and criminal histories.
Links between offences.

A current offence can be compared with these recorded patterns.

The approach assumes that information from solved offences can help investigators make predictions about unsolved offences.

Why Databases Are Useful

A large database may reveal patterns that would be difficult to identify from one case.

For example:

Two behaviours may frequently occur together.
A particular victim type may be linked with a particular interaction style.
A method may be associated with experienced offenders.
A behavioural pattern may appear across several apparently separate crimes.

Statistical patterns allow investigators to move beyond personal impressions.

Primary and Secondary Data

Much of the information used in offender profiling is secondary data because it was originally collected for criminal investigation or administrative purposes rather than for the profiler’s current analysis.

Examples include:

Police reports.
Witness statements.
Crime-scene records.
Interview transcripts.
Criminal records.
Existing databases.

Secondary data can provide access to large numbers of real offences.

However, the profiler has no control over:

How the original information was collected.
Whether important details were recorded.
Whether definitions were consistent.
Whether mistakes were made.

This connects with evaluating primary and secondary psychological data [Lesson 30: Primary and secondary data].

Coding Behavioural Data

Crime-scene information is often qualitative and complex.

To analyse it statistically, investigators may convert it into coded quantitative data.

For example:

Original information	Possible code
The offender reassured the victim	Reassurance present
The offender avoided speaking	Communication absent
The offender brought restraints	Preparation present
The offender removed evidence	Forensic awareness present
The victim was selected from a workplace	Workplace access present

Coding allows comparisons, but it may reduce the detail and meaning contained in the original account.

Reliability of Coding

For coding to be reliable, researchers need:

Clear operational definitions.
Mutually exclusive categories.
Standardised coding instructions.
Trained coders.
Checks of inter-observer reliability.

Two coders should classify the same crime-scene behaviour in a similar way.

If one coder interprets an action as reassurance while another interprets it as control, the database may contain inconsistent information.

This connects with assessing and improving reliability [Lesson 24: Reliability].

Smallest Space Analysis

Smallest space analysis, often shortened to SSA, is a statistical technique used to examine relationships between behavioural variables.

The analysis identifies which behaviours tend to occur together across offences.

The results may be represented spatially:

Behaviours that frequently occur together are positioned closer together.
Behaviours with weaker relationships are positioned further apart.
Clusters may reveal distinct behavioural themes.

The “space” is a statistical representation. It is not a geographical map of where crimes occurred.

A Simple Example of Smallest Space Analysis

Imagine researchers code four behaviours across a large number of robberies:

Offender uses a disguise.
Offender removes surveillance equipment.
Offender brings tools.
Offender acts without preparation.

If the first three behaviours often appear together, SSA may place them close to one another. Acting without preparation may appear further away.

Researchers might interpret the cluster as a theme of:

Planning.
Forensic awareness.
Preparation.

A new offence containing the same cluster may be compared with known offenders who previously showed that pattern.

What Smallest Space Analysis Can Show

SSA may help identify:

Behavioural clusters.
Distinct themes within offences.
Similarities between apparently separate crimes.
Behaviours that are central or common.
Behaviours that distinguish one pattern from another.

It cannot directly reveal:

The offender’s identity.
The offender’s motive.
A certain personality type.
Whether one behaviour caused another.
Whether a statistical association is meaningful without interpretation.

Association Is Not Causation

Statistical analysis may reveal that two behaviours occur together.

For example:

Offenders who conceal evidence may also be more likely to bring equipment.

This does not show that concealing evidence causes preparation or that preparation causes concealment.

The relationship may reflect:

A third characteristic such as experience.
The type of offence.
The opportunity available.
The offender’s access to resources.

Profilers must distinguish statistical association from causal explanation.

This connects with interpreting relationships between co-variables [Lesson 7: Correlations].

From Statistical Pattern to Offender Profile

A statistical relationship becomes useful only when it is translated into a testable prediction.

For example:

The current offence includes evidence concealment, controlled interaction and equipment brought to the scene.
These behaviours form a pattern found in previous solved offences.
In the database, the pattern is associated with experienced offenders.
Investigators predict that the current offender may have a criminal history.
Police records are searched for suspects matching the evidence.
The prediction is accepted, revised or rejected as further information appears.

The conclusion must remain probabilistic.

An Example of Investigative Psychology

Consider this scenario:

Three robberies have occurred in different parts of the same town. In each case, the offender speaks calmly to staff, reassures customers that they will not be harmed and removes the same type of recording equipment before leaving.

A data-driven investigation could proceed as follows:

The repeated reassurance is coded as an interpersonal behaviour.
Calm speech and control suggest a consistent interaction style.
Removal of recording equipment indicates forensic awareness.
The repeated cluster across three robberies supports possible case linkage.
The pattern can be compared with solved offences in a database.
Investigators may predict an offender who is socially competent and experienced.
The profile can then be tested against suspects and physical evidence.

A weak application would state only:

The offender has interpersonal coherence.

A stronger answer explains which behaviours are consistent and what cautious inference may follow.

Applying Interpersonal Coherence

Consider another scenario:

During several offences, the offender asks victims personal questions, refers to them by name and becomes angry if they refuse to respond.

Investigators might infer:

The offender attempts to create a personal interaction.
The same interaction pattern across victims supports behavioural consistency.
The offender may place importance on controlling relationships.
Similar behaviour might appear in ordinary social relationships.
The pattern could help link the offences.

The conclusion should not be:

The offender definitely behaves this way with everyone.

Applying Forensic Awareness

An offender avoids security cameras, wears gloves and removes objects that may contain identifying evidence.

A profile might suggest:

Awareness of evidence collection.
Planning before the offence.
Possible previous contact with police.
Knowledge gained through earlier offending or media.

The evidence does not prove that the offender has a previous conviction.

Applying Time and Place

A series of offences occurs between 5 pm and 7 pm on weekdays near several railway stations.

Possible inferences include:

The offender may use public transport.
The times may fit travel after work.
The offender may know commuter routines.
The locations may be familiar through work or regular travel.

These hypotheses would need to be combined with geographical analysis and other evidence.

Evidence Supporting Investigative Psychology

Research has used statistical analysis to identify recurring behavioural patterns in offences.

AQA’s June 2022 mark scheme accepts:

Canter and Heritage’s analysis of distinct patterns of offender behaviour.
Lundrigan and Canter’s use of smallest space analysis.
The successful application of profiling in the John Duffy investigation.

These examples support the possibility that systematic analysis can identify useful relationships within offence data.

Canter and Heritage

Canter and Heritage analysed behavioural variables from sexual offences to examine whether distinct patterns of interaction could be identified.

The research is relevant because it:

Used actual offence information.
Converted behaviour into coded variables.
Applied statistical analysis.
Identified relationships between offence behaviours.
Supported the use of interpersonal patterns in profiling.

This provides evidence that behavioural data can be organised systematically rather than interpreted entirely through intuition.

However, patterns found in one offence type may not apply to other crimes.

Lundrigan and Canter

Lundrigan and Canter used smallest space analysis to investigate patterns associated with body disposal.

This research illustrates how statistical analysis can identify structured relationships within offending behaviour.

It also shows that:

Data from multiple offences can reveal patterns.
Behaviour and location may be analysed together.
Offending decisions are not necessarily random.

The study also connects with analysing patterns in offence locations [Lesson 3: Geographical profiling].

The John Duffy Investigation

David Canter contributed psychological profiling to the investigation that led to the identification of John Duffy.

The case is often used to illustrate the practical value of investigative Psychology because several features of the resulting profile were consistent with the offender.

A successful case can demonstrate that profiling may:

Generate useful hypotheses.
Help narrow an investigation.
Suggest characteristics not immediately obvious from physical evidence.
Support the connection of related offences.

However, one successful investigation cannot demonstrate that the approach is consistently accurate.

Strength: More Scientific Than the Typology Approach

A major strength is the emphasis on:

Empirical evidence.
Defined behavioural variables.
Databases.
Psychological theory.
Statistical analysis.
Testable predictions.

This may make the approach more objective and scientific than profiling based mainly on intuition or broad offender types.

The June 2022 mark scheme explicitly recognises the use of data, psychological theory and statistics as making the approach more objective than the typology approach.

Objectivity Is Not Complete

Although the approach uses numbers, human judgement is still involved.

Researchers must decide:

Which behaviours to record.
How to define each category.
Which cases to include.
How to interpret a cluster.
Which offender characteristic the pattern suggests.

If these decisions are subjective, statistical analysis does not automatically make the final profile objective.

A numerical output may create an appearance of certainty even when the original coding involved judgement.

Strength: Greater Reliability

The use of standardised coding and statistical procedures may increase reliability.

If investigators use:

The same operational definitions.
The same database.
The same statistical method.
The same decision rules.

they should be more likely to reach similar conclusions.

This could improve inter-rater reliability compared with a profiler relying mainly on intuition.

However, reliability depends on whether the original records and coding are consistent.

Strength: Applicable to a Wider Range of Crimes

The approach can potentially be applied to:

Violent offences.
Burglary.
Robbery.
Arson.
Fraud.
Theft.
Cybercrime.
Sexual offences.

It does not require an offence to fit an organised or disorganised category.

The June 2022 mark scheme recognises broader applicability as an advantage over the typology approach, which was developed mainly from serious violent offences.

This wider scope increases the practical value of the method.

Strength: Flexible Profile Construction

The data-driven approach does not force every offender into one of two types.

It can identify:

Several behavioural dimensions.
Mixed patterns.
Offence-specific themes.
Changes over time.
Similarities that do not fit a category.

This may provide a more individualised profile.

For example, an offence could show:

High forensic awareness.
Limited interpersonal confidence.
Considerable planning.
A local victim selection pattern.

The profiler does not have to label the offender entirely organised or disorganised.

Strength: Profiles Can Be Updated

A data-driven profile can be revised as new information becomes available.

New evidence may include:

Another linked offence.
A witness statement.
A change in method.
New location data.
Additional victim information.
Improved forensic evidence.

The profile can therefore develop with the investigation.

This is preferable to treating the first classification as fixed.

Strength: Practical Value

The approach may help investigators:

Link offences.
Prioritise suspects.
Identify likely offender characteristics.
Search records more efficiently.
Plan interviews.
Allocate investigative resources.
Predict possible future behaviour.

Its value is greatest when the profile provides specific, testable predictions.

For example:

The offender may work in or regularly visit the area at the times when the offences occur.

Investigators can check employment, transport and access records rather than merely describing the offender as “confident”.

Limitation: Database Accuracy

The quality of the profile depends on the quality of the data.

A database may contain:

Missing information.
Incorrect records.
Inconsistent terminology.
Details collected using different procedures.
Offences that were wrongly linked.
Behaviour interpreted differently by different officers.

AQA’s June 2022 mark scheme identifies accurate and detailed crime records as essential to the success of the data-driven approach.

The principle is:

Poor-quality input produces poor-quality output.

Limitation: Unsolved Offences Are Underrepresented

Databases of offender characteristics are usually built from solved cases.

This creates a possible sampling bias.

Offenders who are caught may differ from offenders who avoid detection.

Caught offenders may:

Make more mistakes.
Leave more evidence.
Be less experienced.
Use common methods.
Offend in areas with greater police attention.

A profile based on caught offenders may therefore be less accurate when applied to an offender who is especially skilled at avoiding detection.

Limitation: Recording Practices May Differ

Different police forces or investigators may record behaviour differently.

For example:

One report may describe an interaction as reassuring.
Another may describe similar behaviour as controlling.
A detail considered important in one case may not be recorded in another.
Older offences may contain less complete information.

This reduces reliability and may distort the statistical relationships.

Limitation: Subjective Coding

Converting complex behaviour into simple categories can involve judgement.

Consider the statement:

The offender told the victim that everything would be all right.

This could be coded as:

Reassurance.
Manipulation.
Control.
Attempted rapport.
Threat reduction.

Different coding decisions may produce different statistical patterns.

Clear operational definitions and inter-observer reliability checks are therefore essential.

Limitation: Loss of Qualitative Detail

Coding behaviour into present or absent categories may remove context.

For example, two offenders may both be coded as “speaks to victim”, even though:

One uses calm reassurance.
One asks personal questions.
One shouts threats.
One speaks only to give practical instructions.

Treating these interactions as identical may reduce validity.

Quantitative analysis provides comparability, but qualitative detail may be needed to understand meaning.

Limitation: Behaviour May Be Inconsistent

The approach assumes that offender behaviour is sufficiently consistent to support case linkage and profiling.

However, behaviour may change because:

The offender learns from previous crimes.
A victim resists.
An offence is interrupted.
Different equipment is available.
The offender is under the influence of alcohol or drugs.
The offender deliberately changes methods.
More than one offender is involved.

Inconsistent behaviour can lead investigators to:

Fail to link offences committed by the same person.
Link offences committed by different people.
Construct an inaccurate profile.

Limitation: Common Behaviours Have Limited Value

Some actions may occur in many offences.

For example:

Wearing gloves.
Threatening a victim.
Leaving quickly.
Selecting an isolated place.

If a behaviour is very common, it may provide little help in identifying a particular offender.

The most useful variables are behaviours that:

Occur consistently for one offender.
Are relatively uncommon among other offenders.
Can be recorded reliably.
Have a defensible psychological interpretation.

Limitation: Statistical Relationships Are Probabilistic

A database might show that a behavioural pattern is associated with a particular offender characteristic.

This means the characteristic is more likely, not certain.

For example:

A pattern may be associated with previous offending.

It would be incorrect to conclude:

Anyone showing this pattern must have a criminal record.

Statistical relationships describe tendencies across groups. They cannot reveal the complete identity of an individual offender.

Limitation: Profiles May Be Too Broad

A profile may contain characteristics that apply to many people.

For example:

Lives locally.
Is familiar with the area.
Has some knowledge of police procedures.
Is socially confident.
Has access to transport.

Such predictions may be accurate but provide limited help in narrowing suspects.

A useful profile must be both:

Accurate.
Specific enough to guide investigation.

Limitation: Successful Cases Do Not Prove General Effectiveness

The John Duffy investigation demonstrates that psychological profiling can contribute to a successful case.

However:

It is one investigation.
Many other pieces of evidence were involved.
Successful examples are more likely to be reported.
Unsuccessful profiles may receive less attention.
It can be difficult to separate the profiler’s contribution from ordinary police work.

A well-known success provides practical illustration but not reliable evidence of overall accuracy.

Limitation: Confirmation Bias and Tunnel Vision

Once a profile is produced, investigators may focus too strongly on suspects who match it.

They may:

Ignore suspects who do not fit.
Interpret ambiguous evidence as confirming the profile.
Overlook contradictions.
Ask leading questions.
Stop considering alternative explanations.

This is confirmation bias.

A profile should guide an investigation rather than determine guilt.

Physical, digital and witness evidence should remain more important than whether a person resembles a psychological profile.

Comparison with the Typology Approach

The data-driven approach may be stronger than the typology approach in several ways.

Comparison	Data-driven approach	Typology approach
Starting point	Individual offence data	Organised and disorganised categories
Type of reasoning	Bottom-up	Top-down
Evidence base	Databases and statistical patterns	Interviews and offender types
Classification	Behavioural dimensions and clusters	Two broad categories
Objectivity	Greater emphasis on systematic analysis	Greater reliance on profiler judgement
Flexibility	Can accommodate mixed patterns	May force an offender into one type
Range of offences	Potentially broad	Traditionally focused on serious violent offences
Main limitation	Depends on database and coding quality	Categories may be oversimplified

The June 2022 examiner report noted that students sometimes completely confused the top-down and bottom-up approaches.

Data-Driven Does Not Mean Entirely Automatic

A computer does not independently produce a complete offender profile.

Human investigators still need to:

Select relevant data.
Define variables.
Decide which analyses to use.
Interpret patterns.
Combine the findings with case evidence.
Test predictions.

The approach is systematic, but psychological expertise remains necessary.

Nomothetic Features

The data-driven approach is partly nomothetic because it uses data from many offences to identify general patterns.

It aims to discover relationships such as:

Offences containing pattern A are more likely to involve characteristic B.

This can support prediction and scientific testing.

However, the profile is then applied to a unique individual offence.

The approach therefore combines:

Nomothetic general patterns.
Idiographic attention to the current case.

This connects with general laws and individual case understanding [Lesson 6: Idiographic and nomothetic approaches].

Reductionism

The approach may be criticised as reductionist because complex offending behaviour is divided into coded variables.

This can reduce:

A relationship with a victim.
A sequence of emotional reactions.
A complicated decision.
A distinctive offence.

to categories such as present or absent.

Reduction has scientific advantages:

Behaviours can be compared.
Statistical patterns can be identified.
Procedures can be replicated.

However, it may remove context and meaning.

This connects with the strengths and limitations of reducing behaviour to measurable components [Lesson 5: Holism and reductionism].

Scientific Status

The data-driven approach has several scientific features:

Empirical data.
Operationalised variables.
Statistical testing.
Testable predictions.
Potential replication.
Use of psychological theory.

However, scientific quality depends on:

Reliable coding.
Representative databases.
Transparent methods.
Independent testing of predictions.
Willingness to revise an incorrect profile.

The use of statistics alone does not guarantee scientific validity.

Ethical and Social Implications

Offender profiles can have consequences for people who resemble the predicted characteristics.

Risks include:

Unfair suspicion.
Stereotyping.
Increased surveillance.
Confirmation bias.
Discrimination against particular groups.
Failure to investigate someone who does not match the profile.

Investigators must treat the profile as one source of information rather than proof.

Overall Evaluation

The data-driven approach attempts to make offender profiling more systematic by combining crime-scene information, victim data, psychological theory and statistical analysis. Investigative Psychology examines behavioural consistency, interpersonal coherence, forensic awareness and patterns across offences.

The approach is more flexible than dividing offenders into organised and disorganised types. It can be used with a wider range of offences, supports case linkage and produces predictions that can be tested.

However, its accuracy depends on the quality of crime databases, reliable coding and the assumption that offender behaviour is sufficiently consistent. Statistical associations are probabilistic and may produce profiles that are broad or inaccurate.

The most balanced conclusion is that data-driven profiling can provide useful investigative hypotheses, particularly when based on high-quality data and transparent analysis. It should be used alongside physical evidence and continually revised rather than treated as a certain identification of the offender.

Key Words 🔑

Key word	Student-friendly definition	How it may be used in an exam
Data-driven approach	Profiling based on systematic analysis of offence, victim and database information	Introduce the required offender-profiling approach
Bottom-up approach	Building a profile from the details of an individual offence	Explain how the approach differs from top-down profiling
Investigative Psychology	The application of psychological theory and research to criminal investigation	Explain the psychological basis of data-driven profiling
Offender profiling	Using offence information to infer characteristics of an unknown offender	Introduce the purpose of the approach
Crime database	A collection of coded information from previous offences and offenders	Explain the source of comparison data
Behavioural variable	A defined feature of an offender’s actions that can be recorded and analysed	Explain how crime-scene behaviour becomes data
Coding	Placing qualitative behaviour into defined categories	Explain how statistical analysis becomes possible
Behavioural consistency	The assumption that an offender shows recurring patterns across situations or offences	Explain case linkage and prediction
Case linkage	Deciding whether separate offences were probably committed by the same person	Apply behavioural consistency to an investigation
Modus operandi	The practical method used to commit an offence	Explain similarities and changes in offending methods
Interpersonal coherence	The idea that interaction with victims may reflect wider interpersonal behaviour	Infer possible social characteristics
Forensic awareness	Knowledge of evidence collection and police investigation	Infer possible experience or preparation
Victimology	The systematic examination of victim characteristics and circumstances	Explain why victim data are included in a profile
Smallest space analysis	A statistical technique showing which behavioural variables tend to occur together	Explain how behavioural clusters are identified
Behavioural cluster	A group of offence behaviours that frequently appear together	Explain how a pattern may be compared with solved cases
Secondary data	Information originally collected for another purpose	Evaluate the use of police records and databases
Inter-rater reliability	The extent to which different coders classify behaviour consistently	Evaluate the reliability of coding
Operational definition	A clear and measurable definition of a variable	Explain how coding can be standardised
Statistical association	A relationship between variables that does not necessarily show causation	Evaluate conclusions drawn from databases
Selective sample	A sample that does not represent the full target population	Evaluate databases containing mainly solved crimes
Confirmation bias	Focusing on information that supports an existing profile	Explain the risk of investigative tunnel vision
Probabilistic prediction	A conclusion stating that a characteristic is more likely rather than certain	Use cautious language when constructing profiles
Nomothetic approach	An approach that identifies general patterns across many cases	Explain the use of crime databases
Idiographic approach	An approach focused on the distinctive features of one case	Explain how the current offence is examined individually
Reductionism	Breaking complex behaviour into smaller measurable components	Evaluate the coding of crime-scene behaviour

Hints from the Examiner Reports 💡

State That the Approach Is Data-Driven

Examiner hint: Do not overlook the defining feature of the approach.

The June 2022 examiner report noted that some detailed answers neglected the fundamental point that the bottom-up approach is data-driven.

A strong opening sentence is:

The data-driven approach is a bottom-up method that constructs a profile using crime-scene and victim information, statistical databases and psychological theory.

Do Not Produce a List of Terms

Examiner hint: Naming concepts does not demonstrate understanding.

Some answers were restricted to the lower level because they listed terms such as interpersonal coherence and forensic awareness without explaining them.

For each term:

Define it.
Explain how it is identified from offence data.
State what investigators may infer.
Use cautious language.

For example:

Interpersonal coherence suggests that the way an offender interacts with a victim may reflect their wider interpersonal style. Reassuring and persuading a victim might therefore suggest relative social confidence.

Keep Bottom-Up and Top-Down Profiling Separate

Examiner hint: Students sometimes completely confused the two approaches.

Remember:

Top-down: Begins with organised and disorganised types.
Bottom-up: Begins with the evidence and builds the profile using data.

Do not describe the FBI interviews or organised offender characteristics as part of investigative Psychology.

Explain How the Data Are Used

A complete answer should not stop at:

Information is collected from the crime scene.

Explain the next stages:

The information is coded into behavioural variables, compared with offence databases and analysed statistically to identify relationships and behavioural clusters.

Use Smallest Space Analysis Accurately

Smallest space analysis is not a map showing where offences occurred.

It is a statistical representation in which:

Closely related behaviours appear near one another.
Less closely related behaviours appear further apart.
Clusters may reveal behavioural themes.

Physical location analysis belongs mainly to Lesson 3: Geographical profiling.

Explain Interpersonal Coherence

Do not write:

The offender is coherent.

Write:

Interpersonal coherence is the assumption that the way an offender treats a victim may reflect their usual style of interacting with other people.

Then apply it to the scenario.

Use Evidence to Construct an Argument

AQA’s June 2022 mark scheme accepts evidence, scientific objectivity, broader applicability and database quality as relevant evaluation.

A developed evidence paragraph might state:

Statistical research identifying recurring interaction patterns supports investigative Psychology because it suggests that offence behaviour is not random. However, the patterns depend on how the original behaviours were coded, so subjective classification may reduce validity.

Explain Why Statistics Are a Strength

Do not write only:

It is scientific because it uses statistics.

Develop the point:

Statistical analysis makes the relationships between behaviours explicit and allows the method to be replicated. This reduces reliance on the intuition of one profiler. However, statistical output remains dependent on the quality and coding of the original data.

Evaluate Database Quality

A developed database criticism should identify:

The problem with the data.
How it changes the statistical analysis.
Why this may produce an inaccurate profile.

For example:

Databases contain mainly solved offences, so offenders who are caught may be overrepresented. Patterns identified from these cases may not generalise to more experienced offenders who avoid detection.

Distinguish Usefulness from Accuracy

A profile may be useful without being completely accurate.

It may help investigators:

Generate questions.
Link offences.
Prioritise evidence.
Narrow an initial search.

Do not claim that a profile identifies the offender with certainty.

Compare Directly with the Typology Approach

A strong comparison is:

Unlike the typology approach, investigative Psychology does not force an offender into one of two categories. It identifies behavioural dimensions using data, which may allow mixed or unusual patterns to be represented more accurately.

Use Cautious Language

Appropriate language includes:

“May suggest.”
“Is associated with.”
“Could indicate.”
“Makes the characteristic more likely.”
“Is consistent with.”

Avoid:

“Proves.”
“Definitely means.”
“Identifies the offender.”
“Shows exactly what the offender is like.”

Remember:

Smaller statistical distance means a stronger relationship between behaviours.

Evidence → behavioural interpretation → cautious offender inference

Mistake 13

Mistake: Assuming that behavioural consistency is perfect.

Why this is incorrect:

Offenders may change their methods or respond differently to victims and situations.

How to improve:

Look for recurring patterns while acknowledging situational variation.

Mistake 14

Mistake: Letting the profile replace physical evidence.

Why this is scientifically weak:

Profiles are probabilistic hypotheses and may create confirmation bias.

How to improve:

State that profiles should guide investigation alongside forensic, digital and witness evidence.

Exam-Style Questions ✍️

Question 1

What is meant by a data-driven approach to offender profiling? [2 marks]

Question 2

Explain why the data-driven approach is described as bottom-up. [2 marks]

Question 3

Outline interpersonal coherence as used in investigative Psychology. [4 marks]

Question 4

Explain how smallest space analysis may be used in offender profiling. [4 marks]

Question 5

Three robberies have occurred in the same region. In each offence, the offender speaks calmly to staff, reassures customers and removes the recording equipment before leaving.

Using investigative Psychology, explain how investigators might use this information. [4 marks]

Question 6

A police database contains the following information from solved offences:

Behavioural pattern	Offenders with previous convictions	Offenders without previous convictions
Evidence removed and gloves worn	38	12
No attempt to conceal evidence	14	36

Describe two patterns in the table and explain one inference that investigators might make about an unsolved offence in which evidence was removed and gloves were worn. [6 marks]

Question 7

Explain one reason why information contained in a crime database might produce an inaccurate offender profile. [3 marks]

Question 8

Explain one difference between the data-driven and typology approaches to offender profiling. [4 marks]

Question 9

Briefly evaluate the data-driven approach to offender profiling. [4 marks]

Question 10

Discuss the data-driven approach to offender profiling, including investigative Psychology. [16 marks]

Answers and Mark Scheme

Question 1

Award one mark for each of the following:

The profile is constructed from crime-scene, victim or previous offence data.
Statistical analysis and psychological theory are used to infer characteristics of the unknown offender.

Maximum: 2 marks

Question 2

Award marks as follows:

One mark for stating that the approach begins with the information from the individual offence.
One mark for explaining that behavioural patterns are analysed to build the profile rather than applying a pre-existing offender category.

Maximum: 2 marks

Question 3

Award one mark for each relevant point, up to four marks:

Interpersonal coherence concerns the way the offender interacts with the victim.
The interaction may be consistent across offences.
It may reflect how the offender interacts with people in ordinary life.
Investigators may examine language, reassurance, control or response to the victim.
The interaction can be used to infer possible social competence or relationship style.
These conclusions are probabilistic rather than certain.

A full-mark answer should explain the connection between offence behaviour and wider interpersonal characteristics.

Question 4

Award one mark for each relevant point, up to four marks:

Crime-scene behaviours are coded as variables.
The variables are collected across a number of offences.
Statistical analysis identifies which behaviours frequently occur together.
Closely associated behaviours are represented nearer to one another.
Clusters may reveal distinct behavioural themes.
A current offence can be compared with these patterns.
The pattern may help link offences or infer characteristics of the offender.

Do not credit a purely geographical description.

Question 5

Award up to four marks for effective application:

The repeated behaviours can be coded and compared across the three robberies.
Similar calm speech and reassurance suggest behavioural consistency.
The repeated interaction style may support case linkage.
Interpersonal coherence may suggest a socially confident offender.
Removing recording equipment suggests forensic awareness.
Forensic awareness may indicate planning, experience or knowledge of police procedures.
The pattern can be compared with solved offences in a database.
The profile should be treated as a hypothesis and checked against other evidence.

Question 6

Award up to two marks for describing patterns:

Evidence removal and glove use are more common among offenders with previous convictions.
Thirty-eight offenders with previous convictions showed this pattern, compared with 12 without previous convictions.
Failure to conceal evidence is more common among offenders without previous convictions.
Thirty-six offenders without previous convictions made no attempt to conceal evidence, compared with 14 with previous convictions.

Award up to four marks for inference and explanation:

The current offender may have previous criminal experience.
Removing evidence and wearing gloves suggest forensic awareness.
In the database, this pattern is associated with previous convictions.
Investigators might prioritise suspects with relevant criminal histories.
The inference is probabilistic because 12 offenders without previous convictions also showed the pattern.
The relationship does not prove that the current offender has a conviction.
The database may be incomplete or unrepresentative.

Question 7

Award one mark for identifying a valid problem and up to two marks for explanation.

Possible answers include:

Incomplete records

Police reports may omit relevant behaviours. Missing data could change the statistical pattern and result in an inaccurate profile.

Inconsistent coding

Different investigators may classify the same behaviour differently. Unreliable coding would reduce the accuracy of relationships identified within the database.

Solved-case bias

The database may contain mainly offenders who were caught. These offenders may differ from those who avoid detection, so the profile may not generalise to the unknown offender.

Incorrect information

Witness statements or original crime-scene records may contain errors. Statistical analysis cannot correct inaccurate input data.

Maximum: 3 marks

Question 8

Possible answer:

The data-driven approach is bottom-up because it begins with the behavioural information from an individual offence and uses databases and statistical analysis to construct a profile. The typology approach is top-down because it begins with the organised and disorganised categories and matches the offence to one of them. The data-driven approach therefore builds a profile from patterns, whereas the typology approach applies a pre-existing type.

For four marks, both approaches should be accurate and the contrast should be explicit.

Question 9

Award up to four marks for clear and effective evaluation.

Possible content:

Greater objectivity through data, psychological theory and statistical analysis.
Reduced reliance on profiler intuition.
Ability to profile a broader range of offences.
Support from research using behavioural variables and SSA.
Practical successes such as the John Duffy investigation.
Dependence on accurate and detailed databases.
Problems with subjective coding.
Solved-case sampling bias.
Behaviour may be inconsistent.
Statistical relationships are probabilistic.
Risk of broad profiles or confirmation bias.

Example full-mark response:

The approach is more systematic than the typology approach because behavioural variables are analysed statistically rather than matched to broad categories. This may increase objectivity and allow the method to be replicated. However, the analysis depends on accurate crime databases. If records are incomplete or coding is inconsistent, the statistical pattern and resulting profile may be invalid.

Question 10

A high-level response should include accurate knowledge and developed evaluation.

Indicative AO1 content

Definition and purpose of offender profiling.
Data-driven or bottom-up approach.
Investigative Psychology.
Collection of crime-scene and victim data.
Coding of behavioural variables.
Use of solved-case databases.
Statistical analysis.
Smallest space analysis.
Behavioural clusters.
Behavioural consistency.
Case linkage.
Interpersonal coherence.
Forensic awareness.
Criminal experience.
Significance of time and place.
Construction of probabilistic offender characteristics.

Indicative AO3 content

Supporting evidence from statistical analyses of offence behaviour.
Practical application in the John Duffy investigation.
More objective and scientific than intuition-based profiling.
Standardisation and possible replicability.
Wider range of crimes than the typology approach.
Flexibility and avoidance of rigid offender types.
Dependence on accurate and detailed databases.
Inconsistent recording practices.
Subjective coding and inter-rater reliability.
Loss of qualitative context.
Solved-case sampling bias.
Behavioural inconsistency.
Common behaviours having limited discriminating value.
Statistical associations not showing causation.
Probabilistic and potentially broad predictions.
Confirmation bias and tunnel vision.
Comparison with the typology approach.
Nomothetic and reductionist features.
Need to use profiles alongside objective physical evidence.

Thirteen to sixteen marks: Knowledge is accurate and generally detailed. Discussion is thorough, focused and effective. Evidence is used to construct arguments, methodological issues are explained and comparison with the typology approach is relevant.

Nine to twelve marks: Knowledge is mostly accurate and discussion is generally effective. Investigative Psychology and data use are explained, although some evaluation points may lack detail or precision.

Five to eight marks: Some relevant knowledge is present, but the answer is mainly descriptive. Specialist terms may be listed without explanation, or evaluation may be generic.

One to four marks: Knowledge is very limited or confused. The answer may describe organised and disorganised offenders instead of the data-driven approach, or provide little effective evaluation.