Start with the question
Before any method is chosen, the research question, the population, the outcome and the decision the study is meant to inform are made explicit. Most statistical problems are really question-definition problems.
About JH Stat
An independent statistical consultancy for health research, built on academic practice, methodological rigour and plain explanation.
The statistician
About Jamal Hossain
Jamal Hossain is a UK-based applied health statistician, researcher and university academic with more than ten years of experience in statistics, research methods and quantitative analysis.
He holds a PhD in Social Statistics and Demography and works as a Senior Lecturer in Applied Health Statistics. He is a Fellow of the Higher Education Academy, a recognition of sustained, evidence-informed practice in higher education teaching.
His work sits at the meeting point of statistical methodology and applied health research. He has contributed to multidisciplinary health research and to externally funded research projects, and has extensive experience in teaching, doctoral supervision and statistical consultancy. Across those roles the constant has been translating between statistical method and substantive research question — helping clinicians, epidemiologists, public health researchers and social scientists get to analyses that genuinely answer what they set out to ask.
Alongside consultancy he continues to teach and supervise, which keeps the methods in active use and the explanations tested against real audiences.
Founder and Principal Statistician, JH Stat. Senior Lecturer in Applied Health Statistics.
Academic profiles
Background
Academic and professional background
Qualifications and recognition
- PhD in Social Statistics and Demography
- Advanced quantitative training in the statistical modelling of complex, hierarchical and longitudinal population data.
- Senior Lecturer in Applied Health Statistics
- Current academic post covering teaching, supervision and research in applied health statistics.
- Fellow of the Higher Education Academy
- Professional recognition for sustained, evidence-informed practice in higher education teaching and learning support.
Professional experience
- More than ten years working in statistics, research methods and quantitative analysis
- Experience working on multidisciplinary health research alongside clinicians, epidemiologists and social scientists
- Experience contributing to externally funded research projects, including design and analysis responsibilities
- Teaching statistics and research methods to postgraduate and professional audiences
- Supervision of postgraduate and doctoral researchers
- Statistical consultancy for academic and applied research teams
- Peer review and statistical review of research outputs
Approach
Approach to statistical consultancy
Every project is different, but the working method is consistent.
Match method to data
The analysis follows from the design and the structure of the data — clustering, repeated measures, censoring, missingness — rather than from habit or convenience.
Plan before you analyse
Wherever possible, a written statistical analysis plan is agreed in advance. This protects the study against selective reporting and makes peer review considerably easier.
Test the assumptions
Model assumptions are checked and reported, and sensitivity analyses are used to show how much conclusions depend on particular choices.
Explain in plain English
Findings are written so that a clinician, funder or policy colleague can use them, with the technical detail available but not obstructive.
Leave capability behind
Code, documentation and explanation are handed over so your team can re-run, adapt and defend the analysis after the engagement ends.
Standards
Commitment to rigorous, transparent and reproducible research
Reproducibility is treated as a practical working standard rather than an aspiration. An analysis that cannot be re-run is difficult to defend, hard to extend and expensive to correct.
In practice this means analyses are scripted rather than performed through point-and-click menus wherever the software allows; data preparation is documented as carefully as the modelling; and outputs are generated directly from code so that tables and figures cannot drift out of step with the underlying results.
Reporting follows the relevant guidelines for the study type — for example CONSORT, STROBE, PRISMA or TRIPOD — and analysis plans are shared with collaborators before results are seen wherever the design allows.
What this looks like in delivery
- Version-controlled, commented analysis scripts
- A documented data-preparation trail from raw file to analysis dataset
- Pre-specified analysis plans where the design permits
- Explicit handling and reporting of missing data
- Assumption checks and sensitivity analyses reported, not just performed
- Reporting aligned to the relevant EQUATOR guideline
- Outputs generated reproducibly from code using Quarto or R Markdown
- A handover session so your team understands what was done and why
Tools
Software expertise
Work can be delivered in whichever environment your team already uses, so that the analysis remains maintainable after handover.
R
Primary environment for modelling, simulation and reproducible reporting, including tidyverse workflows, mixed-effects and survival packages.
Stata
Widely used in health and epidemiological research; well suited to survey data, panel models and clearly documented do-file workflows.
SPSS
Common in applied clinical and health services teams. Analyses can be delivered with syntax files so the work remains auditable and repeatable.
AMOS
Structural equation modelling, confirmatory factor analysis and path modelling for measurement and latent variable work.
Git, Quarto and R Markdown are used alongside the packages above to support version control and reproducible reporting. They are working tools rather than statistical analysis software in their own right.
Collaboration
Research collaboration
Beyond time-limited consultancy, JH Stat takes on longer-term collaborative roles within research teams. This can mean acting as the named statistician on a grant application, joining a trial or study steering group, or working as a co-investigator with responsibility for design and analysis.
Collaborative arrangements suit projects where the statistical questions will evolve — multi-stage studies, programme grants, methodological development work, or research where measurement and analysis are themselves part of the contribution.
Authorship is discussed openly at the outset. Where a contribution meets recognised authorship criteria, such as those set out by the ICMJE, authorship is expected; where it does not, acknowledgement is appropriate. Either arrangement is agreed in writing before work begins.
Forms collaboration can take
- Named statistician on a grant application
- Co-investigator with design and analysis responsibility
- Member of a trial or study steering committee
- Independent statistical adviser to a research programme
- Analytical partner on a methodological paper
- Retained advisory time for a research group or department
Principles
Ethical working principles
These principles are not negotiable and apply to every engagement.
Research integrity comes first
Analyses are conducted to answer the research question, not to reach a preferred result. JH Stat will not selectively report, re-specify models until a threshold is crossed, or present exploratory findings as confirmatory.
Academic honesty
JH Stat supports researchers with methodology, analysis and learning. It does not complete assessed academic work on a student's behalf or in any way that would breach an institution's academic integrity rules. Support for doctoral researchers is offered on the same basis as any legitimate methodological training or statistical advice — and students are encouraged to confirm the arrangement with their supervisor.
Confidentiality
Study details, data and unpublished findings are treated as confidential. Non-disclosure agreements, institutional data-sharing agreements and information governance requirements are welcomed and complied with.
Data protection
Wherever possible, work is carried out on de-identified data or within your own secure environment. Data handling follows UK GDPR and the Data Protection Act 2018, and the arrangement is agreed in writing before any data are transferred.
Honest scoping
If a question falls outside JH Stat's expertise, or if the data cannot support the intended analysis, that is said at the outset. Where useful, referral to a more appropriate specialist is offered.
Transparent authorship and credit
Contribution, authorship and acknowledgement are agreed in advance against recognised criteria, so that credit is neither assumed nor disputed later.
Work together
If you have a study in development, an analysis that has stalled, or a team that would benefit from training, an initial scoping conversation can be arranged. The scope, timescale and fees are agreed in writing before any chargeable work begins.