top of page
Research.jpg

Our Research

At BAACRD we are an established foundation created to address interconnected social, economic, educational, and well-being challenges through early, targeted, and empowering support interventions.

Because our interventions and recommendations are outcome-based, the integrity of our research data is not merely a technical requirement it is a foundational promise to our partners. This document outlines the formal procedure for gathering, verifying, and analyzing research data to ensure the accuracy, integrity, and reliability of all BAACRD analytics reports.


Purpose & Scope

This layout statement defines the mandatory protocols for:

  • Collecting primary and secondary research data

  • Verifying the authenticity and reliability of all data sources

  • Validating data accuracy, completeness, and consistency

  • Processing and cleaning raw data for analysis

  • Producing auditable analytics reports

  • Governing data with accountability and transparency


These procedures apply to all BAACRD interventions, including Early Intervention Risk Management, Stress Testing, Infrastructure Stabilization, and Technology & Workflow Engineering.


Data Gathering Process

BAACRD employs a hybrid approach to data collection, combining internal primary data with authoritative secondary sources to ensure comprehensive, multi-dimensional insights.


A. Primary Data Collection

Method | Description | Description

Internal CRM/Database

  • Description: Data pulled directly from internal systems to analyze  organizational performance, client history, and intervention outcomes.

  • Tools & Standards: In-house weekly extraction audits.

Direct Surveys & Forms

  • Description: Custom surveys deployed to stakeholders, clients, and  community members. Built-in validation rules prevent incomplete or illogical entries.

  • Tools & Standards: Jackrabit forms; validation rules (range, required fields).

Structured Interviews

  • Description: One-on-one or group interviews conducted using standardized protocols. Sessions are recorded, transcribed verbatim, and checked for transcription accuracy.

  • Tools & Standards: Recording software, transcription services; independent review of 10% of transcripts.

Observational Data

  • Description: Direct observation of organizational processes, workflows, or community interactions. Observers use standardized checklists.

  • Tools & Standards: Observation templates; inter-observer reliability checks.


B. Secondary Data Collection

Source Type | Description | Examples

Industry Reports

  • Description: Data from reputable industry associations and market research firms.

  • Examples: Statista, IBISWorld, McKinsey & Company, Deloitte Insights.

Academic Journals

  • Description: Peer-reviewed   studies from recognized academic sources.

  • Examples: JSTOR, Google Scholar, PubMed, Scopus-indexed journals.

Government & Public Data

  • Description: Official databases, regulatory bodies, and public sector statistics.

  • Examples: ONS (UK), World Bank, IMF, Eurostat, U.S. Census Bureau, DfE.

Regulatory Filings

  • Description: Publicly available legal and compliance documents.

  • Examples: SEC filings, Companies House, Charity Commission records.


Protocol: No secondary data is accepted without full citation and source documentation. Anonymous or unverifiable online sources (e.g., uncredited blogs, forums) are explicitly excluded.


Verification & Validation Procedures

Data verification is mandatory before any analysis begins. It confirms that gathered data meets BAACRD's quality standards: accuracy, completeness, consistency, and reliability.


A. Source Verification – The "C-R-A-P" Method + Triangulation
Every data source is subjected to the C-R-A-P test

Criterion | Question to Answer / Passing Standard

Currency

  • Is  the data up-to-date? For time-sensitive analysis, data must be current within  defined thresholds (e.g., economic data: ≤12 months; historical trends: ≤5 years).

Relevance

  • Does the data directly answer the research question? Data must map explicitly to the intervention's scope.

Authority

  • Who is the author or publisher? Are they qualified? Sources must have verifiable institutional affiliation, peer review, or recognized expertise.

Purpose

  • Why was this data produced? Is it unbiased? Commercially funded or advocacy-driven data is flagged and used only with transparent disclosure.


In addition, BAACRD employs Triangulation:


Triangulation: Cross-referencing every key data point against at least three independent sources before acceptance. If three sources do not align, the data is flagged for further investigation or discarded.


B. Data Validation Techniques

Technique | Description | Example

Range Checks

  • Description: Ensuring numerical values fall within logical boundaries.

  • Example: Percentage values must be between 0–100. Age values cannot be negative.

Consistency Checks

  • Description: Ensuring alignment across related data sets.

  • Example: Regional  sales totals must match the sum of district-level data.

Format Checks

  • Description: Enforcing standardized data formats.

  • Example: Dates in YYYY-MM-DD format. Currency in standardized units (e.g., USD, GBP).

Uniqueness Checks

  • Description: Identifying and removing duplicate records.

  • Example: Duplicate client entries merged or deleted.

Logic Checks

  • Description: Verifying that relationships between fields are plausible.

  • Example: "Date of intervention" cannot precede "date of initial contact."


C. Exception Handling

Issue | Action

Missing values (<5% of dataset)

  • Imputed using statistical methods (mean, median, or regression) with clear documentation.

Missing values (>5% of dataset)

  • Flagged; client notified; alternative data sources sought.

Outliers

  • Investigated for validity. If genuine, retained with annotation. If error, corrected or removed.

Source contradiction

  • Escalated to Data Steward; resolution requires majority consensus among three independent verifiers.


Data Processing & Cleaning

Before analysis, raw data undergoes a structured cleaning process to enhance reliability and usability.


A. Data Cleaning Steps

Step | Action

1. Irrelevant Data Removal

  • Fields or records not aligned with research objectives are excluded.

2. Structural Error Correction

  • Typos, inconsistent naming conventions, and misaligned formats are standardized.

3. Missing Value Management

  • As per exception handling above (Section III.C).

4. Duplicate Elimination

  • Automated and manual de-duplication processes applied.


B. Data Transformation

Technique | Purpose

Normalization

  • Scaling data to a common range for comparative analysis.

Aggregation

  • Summarizing granular data (e.g., daily transactions to monthly totals).

Conversion

  • Transforming data types (e.g., text to categorical codes, currency conversions).

Enrichment

  • Appending additional verified data from secondary sources.


All transformations are logged in a data lineage document, enabling full reverse traceability.


Data Analytics & Reporting
A. Analytics Method Selection

Research Goal | Typical Methods

Descriptive (what happened?)

  • Mean, median, frequency istributions, standard deviation.

Diagnostic (why did it happen?)

  • Correlation  analysis, regression, cohort analysis

Predictive (what could happen?)

  • Time series forecasting, machine learning models, scenario simulation.

Prescriptive (what should we do?)

  • Optimization models, decision trees, risk scoring.


B. Visualization Standards

  • Tools:Tableau, Microsoft Power BI, or Python (Matplotlib/Seaborn)

  • Principles:Clarity, accuracy, accessibility (colorblind-friendly palettes), and clear labeling

  • Required Elements: Title, source citation, date of data extraction, confidence intervals where applicable


C. Review Process

Stage: Draft Report

  • Responsible Party: Lead Analyst

  • Action: Initial analysis and visualization.

Stage: Peer Review

  • Responsible Party: Second Analyst

  • Action: Independent verification of methodology, calculations, and source citations.

Stage: Data Steward Review

  • Responsible Party: Designated Data Steward

  • Action: Confirms compliance with governance framework and validation protocols

Stage: Final Approval

  • Responsible Party: BAACRD Research Director

  • Action: Sign-off before client delivery.


D. Final Report Structure

All BAACRD analytics reports contain the following sections:

  1. Introduction:Research objectives and scope

  2. Methodology:Data sources, collection dates, sample size, verification methods

  3. Results:Findings with visualizations and statistical summaries

  4. Limitations:Acknowledged data gaps, assumptions, and confidence bounds

  5. Conclusion: Key insights and actionable recommendations

  6. Appendix: Full source citations, data lineage, validation logs


Transparency Requirement: Any data point that could not be verified through triangulation is explicitly marked as "Unverified – For Illustrative Purposes Only."


Data Governance Framework

BAACRD's data practices are governed by a formal framework that defines roles, responsibilities, and standards for data management.


A. Accountability Structure

Role | Responsibility

Chief Data Officer (CDO)

  • Overall ownership of data governance strategy and compliance.

Data Stewards

  • Day-to-day data quality, validation, and documentation. Assigned per intervention.

Analysts

  • Proper application of collection, cleaning, and analytical methods.

External Auditors

  • Annual independent audit of data processes and sample of reports.


B. Security & Privacy

Element | Standard

Storage

  • Encrypted cloud storage (AWS, Azure) with geographic redundancy.

Access Control

  • Role-based  access; minimum necessary privilege; all access logged.

Compliance

  • GDPR, UK Data Protection Act 2018, and sector-specific privacy standards.

Retention

  • Raw data retained for 7 years; anonymized after 2 years for non-essential identifiers.


C. Documentation & Auditability

All data, sources, and verification steps are meticulously documented to enable external auditing and validation.


Documentation Requirements:

  • Source citation with URL/DOI and access date

  • Verification checklist (C-R-A-P + triangulation) signed off by Data Steward

  • Data cleaning and transformation logs

  • Review sign-offs at each stage


Summary of BAACRD Data Integrity Commitments

Principle | BAACRD Commitment

Accuracy

  • Every data point is verified through range, consistency, format, and logic checks

Reliability

  • Triangulation across three independent sources before acceptance.

Transparency

  • Full methodology and source disclosure in every report.

Auditability

  • Complete   data lineage from collection to final insight.

Accountability

  • Designated Data Stewards and external audits.

Ethical Use

  • No unverified or biased data used without clear disclosure.


In an era where decisions are increasingly data-driven but information is increasingly fragmented, BAACRD stands on a non-negotiable principle: insight without integrity is not insight it is misinformation. Our systematic, multi-layered approach to data gathering from primary source collection to rigorous C-R-A-P verification and mandatory triangulation ensures that every statistical finding, every analytical report, and every strategic recommendation we deliver rests upon a foundation of verifiable truth. We do not assume accuracy; we prove it. We do not trust sources; we test them. Every data point is challenged, cross-referenced, and validated before it ever informs a client recommendation. This is not an extra step in our process it is the foundation of our process. When BAACRD delivers an analytics report, our partners receive not mere numbers, but certified intelligence auditable, defensible, and actionable. That is our methodology. That is our standard. That is our promise to every organization and individual who trusts us with their future.

bottom of page