Easier Said Than Done Episode 8 - Understanding Grounds for Stop & Search: New Insights from MOPAC’s Landmark Analysis
- Helen Khezrzadeh
- 14 hours ago
- 3 min read
Introduction
The latest session in SEBP’s Easier Said Than Done series brought together practitioners, researchers, and oversight professionals to explore one of the most contested areas of policing: the grounds recorded for stop and search. This session featured Dr. Paul Dawson, Head of Evidence & Insight at the Mayor’s Office for Policing and Crime (MOPAC), joined by Dr. Anthony Dugay, a data science specialist from the same team.
The focus was a major new piece of research analysing all 152,000 stop and search records in London during 2023, including nearly 17 million words of free‑text grounds written by officers. Paul described this dataset as “fantastic” in scale and offers an unprecedented opportunity to understand how grounds are constructed, how they vary, and what this means for fairness, legitimacy, and operational practice.
Why This Research Matters
Stop and search remains one of the most scrutinised police powers, particularly due to persistent disproportionality affecting Black Londoners. Previous studies suggested that Black individuals were more likely to receive vague or subjective grounds, but these were based on small samples. MOPAC’s new analysis scales this up dramatically, using a combination of:
Human-coded samples (600 grounds texts)
A bespoke coding framework developed with community scrutiny experts
Natural language processing models trained to code all 152,000 records
This hybrid approach allowed the team to analyse the entire dataset with consistency and rigour.
What the Research Found
1. Black Londoners receive more vague and subjective grounds
The analysis confirmed significant differences in the types of grounds recorded. Black Londoners were more likely to have grounds referencing:
“concealing clothing”
“area known for weapons or robbery”
“attempts to avoid police contact”
“smell of drugs”
White Londoners, by contrast, were more likely to have evidential grounds recorded, such as:
“subject seen stealing”
“subject admitted possession”
Clearer articulation of what was being searched for
This aligns with earlier small‑scale studies but now rests on a dataset of unprecedented size.
2. A new ‘quality of grounds’ metric
MOPAC developed a statistical measure of “quality” based on whether particular grounds were associated with positive outcomes (e.g., finding the item searched for). Paul clarified: “When I say quality, I’m talking about the statistical likelihood that it will find something or not.”
The key findings were:
Grounds like “attempting to avoid contact” or “concealing clothing” were not associated with positive outcomes.
These factors were instead linked to no further action.
Black Londoners had significantly lower quality grounds in several search types, including weapons, going equipped, and stolen property.
3. Quality varies by time of day, age, and borough
Quality was highest during low‑volume early morning hours.
For white Londoners, quality remained stable throughout the day.
For Black Londoners, quality dropped sharply during peak operational hours.
Young people (14–19) received the lowest quality grounds.
Inner London boroughs showed lower average quality than outer boroughs.
4. A strong link between quality and public perceptions
Boroughs with higher‑quality grounds also had more positive public attitudes toward the fairness of stop and search. As Paul described: “The public are able to pick up when stop and search is being done well.”
Introducing the ‘Groundskeeper’ Tool
One of the most practical outputs of the project is a new tool — the Groundskeeper — which allows users to paste in a grounds text and receive:
A quality score
Automated feedback
Identification of key factors present in the text
Paul described it as “a really interesting tool that we can share with supervisors, with community groups…”
SEBP attendees were enthusiastic about its potential for scrutiny panels, training, and reflective practice.
What's Next?
MOPAC plans to:
Repeat the analysis annually (2023, 2024, 2025…)
Track changes over time, including the impact of the Met’s Stop & Search Charter
Explore new variables such as officer length of service and repeat stops
Expand the methodology to other forces if interest grows
The Met has already committed to using the findings to strengthen supervision and local learning.
As Paul summarised: “This is a good example of really actionable analytics.”
Access the slides that Paul shared on the link below:
