Impact Portal · 2024–25

Longitudinal Survey — Early Findings Report

Phase 1 findings · Spring 2024 – Autumn 2025 · 1,277 unique participants across four survey waves

📊 At a Glance

This report presents findings from 1,277 unique participants surveyed across four waves (Spring 2024 – Autumn 2025). The analysis examines whether Bloomsbury Football’s programmes are achieving their intended outcomes: improving physical health, building confidence and wellbeing, and fostering social connections.

Our Key Metrics (Oct 2025)

86%

of young people are active every day

90%

of parents say their young person shows resilience — keeping going when things get hard

96%

of parents say their young person feels more confident having played football at Bloomsbury

95%

of young people have made friends from different national, ethnic or socioeconomic backgrounds

Methods

Origins

This survey began in early 2024 because we wanted more evidence behind what our coaches were already seeing on the pitch. We needed a way to test our theory of change rigorously and track those shifts over time, rather than relying only on observation.

We co-created the survey with the University of Westminster, adapting validated items from Sport England’s Active Lives Survey (for physical activity and wellbeing) and the CASEL framework (for social-emotional outcomes such as resilience, self-awareness, and relationship skills). Each wave runs across our youth, adolescent, and parent populations so that we can triangulate self-report with parent-report and track the same young people as they progress through the programme.

The Data

We analysed survey responses collected across four waves over 2 years. Rather than treating each response independently, we tracked individual participants over time to understand how outcomes change with continued participation.

Sample composition by population Responses · unique participants · tracked longitudinally 0 200 400 600 800 745 532 162 Parents 479 350 94 Participants 13–18 482 395 74 Participants U13 Total responses Unique participants Tracked over time
Three Populations, Three Perspectives
Population What They Report Sample Size
Parents/Guardians Young person’s physical health, confidence, social development, school focus 532 unique
Participants 13–18 Self-reported happiness, life satisfaction, loneliness, confidence 349 unique
Participants Under 13 Simplified survey: enjoyment, friendships, basic social-emotional skills 395 unique

Tracking participants over time (330 individuals across 2+ waves) allows us to examine within-person change, a more robust approach than comparing different groups of people.

Ethics, consent and safeguarding

Surveys are administered as part of normal programme participation, with separate forms for parents/guardians and for young people aged 13–18 and under 13. Participation is voluntary. Survey responses are linked across waves using anonymised participant identifiers before analysis; named records are not used in any analytical output. This work is governed by Bloomsbury Football’s safeguarding policy and data-protection practices for working with young people and families.

Analytical Approach

Our analysis strategy was designed to answer three questions: Are participants changing over time? Who benefits most? And does Bloomsbury reach everyone equitably?

Statistical Methods

1. Physical Activity Transitions We compared participants’ activity levels before joining Bloomsbury to their current levels, categorising each person as having improved, declined, or stayed the same.

2. Longitudinal Change Models Linear Mixed Effects (LME) models tracked how outcomes changed across survey waves, accounting for the fact that individuals were measured repeatedly.

3. Differential Analysis Early analysis revealed high ceiling effects; many participants already scored 8-10 out of 10 at baseline, leaving little room for measurable improvement. To address this, we separately analysed participants who started with lower scores (≤7) or scored under the mean, where meaningful change was possible.

4. Equity Analysis We tested whether outcomes differed by gender, socioeconomic status, and ethnicity using interaction terms in our models.

Important Caveats

No control group: We draw useful comparisons using national and local averages, but without a defined comparison group of young people not participating in Bloomsbury, we can’t yet isolate the programme’s specific contribution to the changes we see. Some of the improvement likely reflects natural development or the influence of other activities. We plan to address this with more rigorous MEL designs over the coming years.

Attribution questions: Some survey items ask about changes “since joining Bloomsbury,” which may lead to overestimation of programme effects due to recall bias and social desirability.

Who Are Our Participants?

The figures below describe the demographics of the people who completed our surveys during this reporting period. They are not a precise breakdown of Bloomsbury Football’s full programme reach. Survey participation is voluntary, so the sample is broadly representative of programme participation but should be read as the survey sample, not as the charity’s overall participant composition.

Gender breakdown by population Each bar shows percentage of participants by gender Parents 67.7% 31.7% Participants 13–18 55.3% 44.3% Participants U13 75.5% 23.7% Boys Girls Other

Note: The figures describe the demographics of survey respondents, not a precise breakdown of Bloomsbury Football’s full programme reach.

Ethnic minority share by population Around half of respondents identify as ethnic minority (parents and young people 13–18) Parents 47.0% 53.0% Participants 13–18 46.6% 53.4% Ethnic minority background Other

Note: The figures describe the demographics of survey respondents, not a precise breakdown of Bloomsbury Football’s full programme reach.

Families receiving means-tested benefits Parents only · n = 456 valid responses Parents 50.7% 49.3% Receives means-tested benefits Does not receive benefits

Note: The figures describe the demographics of survey respondents, not a precise breakdown of Bloomsbury Football’s full programme reach.

How long participants have been with Bloomsbury Most participants have been engaged for a year or more Parents 19.2% 12.2% 13.2% 55.4% Participants 13–18 9.4% 10.9% 18.3% 61.4% Participants U13 22.8% 11.6% 19.5% 46.1% 0–3 months 3–6 months 6 months – 1 year 1 year+

Note: The figures describe the demographics of survey respondents, not a precise breakdown of Bloomsbury Football’s full programme reach.

Change Over Time

This section presents the core findings, organised by outcome domain in line with Bloomsbury’s Theory of Change. For each domain, we highlight statistically significant results and contextualise them against the broader evidence base.

🎯 Bloomsbury’s Aim

“Young people increase physical activity levels; young people’s physical health is improved.”

Physical Health Improvement (Parents)

Finding: Large Effect Size (d = 0.94)

For young people who started with lower physical health ratings, parents reported a 1.9-point improvement on the 0-10 scale. This is a large effect size by conventional standards, and considerably larger than the effects typically found in school-based physical activity interventions, which meta-analyses place at d = 0.35-0.50.

Young people who started lower in physical health: parent ratings Parent-rated physical health (0–10) before and after, for young people with baseline ≤ 7 0 2 4 6 8 10 5.4 Before 7.3 After Cohen’s d = 0.94 · n = 60

Physical Activity Levels (Participants 13–18)

Finding: 7-to-1 Improvement Ratio

28.3% of young people 13–18 moved up an activity category between baseline and follow-up — from Less Active to Fairly Active or Active — while only 4.0% moved down. This 7:1 ratio of improvers to decliners is statistically significant (p < 0.001).

The UK Chief Medical Officers (2019) recommend that young people achieve an average of 60 minutes of moderate-to-vigorous physical activity daily. Sport England’s Active Lives Children and Young People Survey (2024) reports that only 48% of young people in England meet this guideline; in comparison, 86% of young people in our programmes report being physically active. Bloomsbury’s role in keeping young people active and shifting young people into higher activity categories speaks directly to this national challenge.

Activity category change (Participants 13–18) For every adolescent who moved down, seven moved up · n = 350 paired observations 0% 10% 20% 30% Moved down 4.0% Moved up 28.3% 67.7% of young people 13–18 maintained their activity category between waves.
🎯 Bloomsbury’s Aim

“Young people’s mental health is improved; young people have increased confidence and resilience.”

Confidence (Parents)

Finding: Large Effect Size (d = 0.86)

Young people who initially scored lower on confidence showed a 2.4-point improvement on the 0-10 scale. Building confidence in sport is foundational; research shows that sport-based confidence often transfers to other areas of life, including school and social situations.

This finding is particularly meaningful in light of Women in Sport’s Reframing Sport for Teenage Girls research (2022), which identifies lack of confidence as a primary barrier to girls’ continued sport participation.

Young people who started lower in confidence: parent ratings Parent-rated confidence (0–10) before and after, for young people with baseline ≤ 7 0 2 4 6 8 10 5.9 Before 8.3 After Cohen’s d = 0.86 · n = 29

Feelings of Loneliness (Participants 13-18)

Finding: Medium Effect Size (d = 0.52)

Among young people 13–18 who initially reported higher loneliness, scores improved by 1.6 points on the 0-10 scale. This medium effect size is meaningful; loneliness during adolescence is associated with depression, anxiety, and reduced academic performance.

The Longitudinal Study of Australian Children (2024) found that consistent sport participation is associated with 31% lower risk of loneliness.

Adolescents who started lonelier: 'not often lonely' score Self-reported 'not often lonely' (0–10) before and after, for adolescents below the median at baseline 0 2 4 6 8 10 6.3 Before 7.9 After Cohen’s d = 0.52 · n = 42
🎯 Bloomsbury’s Aim

“Young people increase positive relationships across backgrounds; young people improve social and emotional literacy.”

Cross-Background Friendships (Parents)

Finding: Medium-Large Effect Size (d = 0.72)

For young people whose parents initially rated cross-background friendships lower, scores improved by 1.9 points. In an increasingly diverse society, the ability to form friendships across ethnic, socioeconomic, and cultural lines is a valuable life skill.

The UEFA Foundation’s Football for Inclusion programme evaluations (2023) and GIZ’s Sport for Development portfolio reviews (2022) both document sport’s role in promoting social cohesion and cross-cultural connections, while emphasising that intentional programme design matters. Bloomsbury’s diverse participant base creates natural opportunities for these connections.

Young people who started lower in cross-background friendships Parent-rated score (0–10) before and after, for young people with baseline ≤ 7 0 2 4 6 8 10 4.9 Before 6.8 After Cohen’s d = 0.72 · n = 61

School Concentration (Parents)

Finding: Medium Effect Size (d = 0.70)

Parents reported a 1.8-point improvement in their young person’s school concentration for young people who started lower. The link between physical activity and academic focus is well-documented: Public Health England’s evidence briefing (2015) confirms associations between physical activity and educational attainment.

Young people who started lower in school concentration Parent-rated score (0–10) before and after, for young people below the median at baseline 0 2 4 6 8 10 4.3 Before 6.1 After Cohen’s d = 0.70 · n = 77

Peer Relationships (Participants Under 13)

Finding: Large Effect Size (d = 1.16)

Among young people under 13 who initially scored lower on peer relationships, we observed a 2.9-point improvement, the largest effect in our analysis. Team sport naturally develops social skills: cooperation, communication, and conflict resolution. For young people who enter with weaker peer relationship skills, Bloomsbury provides a structured environment to practice and develop these competencies.

Young people under 13 who started lower in peer relationships Self-reported 'gets along with friends' (0–10), for under-13s below the median at baseline 0 2 4 6 8 10 7.8 Before 10.0 After Cohen’s d = 1.16 · n = 20

The chart below summarises all significant findings from the differential analysis, showing the magnitude of improvement for participants who started with lower scores.

Significant improvements across all populations Effect sizes (Cohen’s d) for participants with room to improve at baseline 0.0 0.2 0.5 0.8 1.0 1.2 Small Medium Large Physical health 0.94 Confidence 0.86 Cross-background friendships 0.72 School concentration 0.70 Reduced loneliness 0.52 Peer relationships 1.16 Parents Participants 13–18 Participants U13
Interpreting Effect Sizes

Effect sizes tell us the magnitude of change, not just whether change occurred. In education and health research, Cohen’s d of 0.2 is considered small, 0.5 medium, and 0.8 large.

For comparison, a meta-analysis of school-based physical activity interventions (Sports Medicine Open, 2020) found effect sizes of d = 0.35 for anxiety reduction, d = 0.75 for resilience, and d = 0.88 for wellbeing. Bloomsbury’s effect sizes, ranging from 0.52 to 1.16, are at or above these benchmarks, suggesting the programme is achieving meaningful impact.

Longer Participation, Better Outcomes

Does staying with Bloomsbury longer lead to better outcomes?

Longer participation, better outcomes Parent-reported scores (0–10) by how long the young person has been with Bloomsbury 6 7 8 9 10 0–3 mo 3–6 mo 6–12 mo 1 year+ Physical health Enjoyment of sports Confidence Cross-background friendships
Interpretation Note

This cross-sectional pattern cannot prove causation. Young people who remain engaged may differ from those who leave. However, the consistent upward trend across multiple outcomes suggests that sustained participation may yield cumulative benefits.

Equitable Impact

A key question for any youth programme is whether it benefits all participants equally, or whether traditionally vulnerable groups are left behind. We tested this for gender, socioeconomic status, and ethnicity.

Gender: Girls Gain More Confidence

Confidence change among young people who started below median Parent-reported confidence on a 0–10 scale, before and after · interaction p = 0.02 4 5 6 7 8 9 10 BASELINE FOLLOW-UP 5.5 8.5 +30% Girls 6.0 7.8 +18% Boys Subgroup with below-median baseline confidence · n = 53 girls, 105 boys
Finding: Girls Show Greater Confidence Gains

Among young people who started with below-median confidence, girls’ parent-reported confidence rose from 5.5 to 8.5 (+30% of the 0–10 scale), while boys’ scores rose from 6.0 to 7.8 (+18%). The interaction is statistically significant (p = 0.02), meaning the difference in the size of those gains between girls and boys is unlikely to be due to chance.

This finding directly counters national trends. Women in Sport’s Reframing Sport for Teenage Girls research (2022) found that 43% of previously ‘sporty’ girls disengage from sport after primary school, with lack of confidence and fear of judgement as primary barriers. UNESCO’s Fit for Life analysis (2024) reports that 49% of adolescent girls drop out of sport — six times the rate observed in boys.

Bloomsbury appears to be creating an environment where girls can build confidence in a traditionally male-dominated sport, a meaningful achievement.

Socioeconomic Status & Ethnicity: Mostly Equitable

Finding: No Differences by Family Income

We found no significant differences in outcomes between young people from families receiving benefits and those who do not. Lower-income families benefit equally from Bloomsbury programmes.

Outcome Receives Benefits Does Not Receive Significant?
Physical Health 8.3 8.4 No
Confidence 8.5 8.4 No
Cross-Background Friendships 8.1 8.2 No
School Concentration 7.4 7.5 No

This is particularly noteworthy given that Sport England’s Active Lives Children and Young People Survey consistently shows young people from affluent families are significantly more likely to be active. Bloomsbury is successfully reaching lower-income families and delivering equivalent outcomes.

Results at a Glance

A summary of the outcomes examined in this analysis, grouped by strength of evidence. Effect sizes are Cohen’s d; ratios are improvers-to-decliners between baseline and follow-up.

Strong & Significant

Outcomes that showed clear, statistically significant improvement with medium-to-large effect sizes.

  • Physical activity (13–18): 7:1 improvement ratio
  • Physical health (Parents): d = 0.94
  • Confidence (Parents): d = 0.86
  • Cross-background friendships (Parents): d = 0.72
  • School concentration (Parents): d = 0.70
  • Peer relationships (U13): d = 1.16
  • Girls’ confidence: significantly greater gains than boys
  • Socioeconomic equity: equivalent outcomes across income groups

Moderate or Emerging

Outcomes that showed positive trends with smaller effects, or signals approaching significance.

  • Reduced loneliness (13–18): d = 0.52 (medium effect)
  • Physical activity (U13): positive but smaller ratio (1.4:1)
  • Trust in peers (13–18): positive trend, small effect
  • Resilience: modest improvement, not statistically significant
  • Communication skills: positive trend, not statistically significant

What we did not detect: overall happiness/life satisfaction, help-seeking behaviour, decision-making skills, and parent-reported communication showed no statistically significant change in this wave. Several of these are subject to ceiling effects (baseline scores already above 8.5/10), which limits our ability to detect movement. They are discussed further in the technical-details section, and remain priorities for the next survey wave.

Technical Details

Three blocks — click any to expand the underlying methodology and limitations.

Data Cleaning
  • Combined four survey waves into longitudinal datasets
  • Used fuzzy name matching (Jaro-Winkler similarity, threshold 0.80) to link participants across waves
  • Removed duplicate responses, keeping the most complete
  • Created standardised variables with consistent coding across waves
  • Final datasets: Parents (745 records, 532 unique, 162 longitudinal); Participants 13-18 (479 records, 349 unique, 94 longitudinal); Participants U13 (482 records, 395 unique, 74 longitudinal)
Statistical Methods

Physical Activity Transitions: Compared “before joining” vs “current” activity levels. Used McNemar-style tests for matched pairs to assess whether the proportion improving exceeded the proportion declining.

Longitudinal Change (LME): Linear Mixed Effects models with random intercepts for participants: Outcome ~ Wave + (1 | Participant). This approach accounts for repeated measures on the same individuals.

Differential Analysis: Filtered to participants with baseline scores ≤7 (indicating room for improvement), then ran LME models on this subset. This addresses ceiling effects that can mask true change.

Equity Analysis: LME models with interaction terms: Outcome ~ Wave * Subgroup + (1 | Participant). Significant interactions indicate different rates of change between groups.

Effect Sizes: Cohen’s d calculated as (Mean_after − Mean_before) / SD_pooled. Interpretation: 0.2 = small, 0.5 = medium, 0.8 = large.

Limitations
  1. No control group — Cannot definitively attribute changes to Bloomsbury vs. natural development
  2. Self-report bias — Particularly for retrospective “before joining” questions
  3. Ceiling effects — Many participants scored high at baseline. This was especially true for young people under 13, which is a well documented effect on survey research with kids. This limited our ability to detect improvement, but was addressed using differential analysis.
  4. Attrition — Participants retained across waves may be more engaged. There was a significant number of participants who did not have subsequent waves.
  5. Social desirability — “Attribution” questions may overestimate programme effects.

Further Evidence

Each topic below pairs the wider research literature with what we saw in our own data. Click any title to expand.

Physical Activity & Health

The UK Chief Medical Officers (2019) recommend 60 minutes of daily activity for young people. Sport England’s Active Lives data (2023/24) report that only 48% of young people meet this guideline, with boys (51%) more likely than girls (45%). The WHO Global Action Plan on Physical Activity 2018–2030 targets a 15% relative reduction in physical inactivity by 2030.

How Bloomsbury Compares

Bloomsbury’s 7:1 improvement ratio among young people 13–18 demonstrates the programme is successfully shifting young people toward more active lifestyles, directly contributing to national activity goals.

Academic Performance & Concentration

Public Health England (2015) confirms strong associations between physical activity and educational outcomes, citing evidence of improvements in classroom behaviour, cognition, and attainment among more active young people.

How Bloomsbury Compares

Parents report medium effect sizes (d = 0.70) for school concentration, consistent with the evidence linking physical activity to academic focus.

Mental Health & Loneliness

Meta-analyses of school-based interventions find effect sizes of d = 0.35–0.88 for anxiety, resilience, and wellbeing. The Longitudinal Study of Australian Children found a 31% lower loneliness risk among consistent sport participants.

How Bloomsbury Compares

Bloomsbury’s effect size in loneliness (d = 0.52) meets meta-analytic benchmarks, suggesting the programme achieves meaningful wellbeing improvements.

Gender in Sport

Women in Sport (2022) reports 43% of ‘sporty’ girls disengage after primary school. Key barriers: lack of confidence (61%), fear of judgement (68%). UNESCO (2024) notes 49% of girls drop out during adolescence — six times the rate observed in boys.

How Bloomsbury Compares

Bloomsbury’s finding that girls show greater confidence gains than boys directly counters national trends. The programme creates an environment where girls can thrive.

Disadvantaged Communities

The Chiles-Webster-Batson Commission (2022) documented how community sport organisations mitigate health inequalities. Disadvantaged young people face greater barriers but also potentially greater benefits from sport participation.

How Bloomsbury Compares

Bloomsbury reaches lower-income families (35% receive benefits) and delivers equivalent outcomes, demonstrating successful delivery of equitable impact.

← Back to portal