Skip to main content
Evidence For Social Impact Lab

RCT & Longitudinal Operations

Protecting experimental protocols from assignment to follow-up.

The problem we solve

Experimental integrity is decided in operations. Assignment errors, contamination between arms and unmanaged attrition damage a trial long before analysis begins.

Scope — what is included

  • Implementation of the randomisation procedure approved by the investigator
  • Intervention delivery tracking against assignment
  • Contamination monitoring between study arms
  • Longitudinal respondent management across follow-up rounds

Methodology

  • Blocked randomisation implementation
  • Permanent respondent IDs
  • Assignment reconciliation
  • Compliance monitoring
  • Attrition tracking

Deliverables

  • Assignment records with reproducible seed and code
  • Implementation monitoring reports
  • Protocol-deviation logs
  • Panel-linked data across rounds

What we need from you

Research decisions stay with your team. To deliver this work we need you to:

  • Define and approve the experimental design, unit of randomisation and analysis plan
  • Approve the randomisation code before it is executed
  • Decide how protocol deviations are treated in analysis

Illustrative workflow

A synthetic example, included to show how the work runs. It is not a past engagement and does not describe a real client.

A two-arm trial randomises at the school level and requires three follow-up rounds over eighteen months.

  1. 1Execute the investigator-approved randomisation script and archive the seed
  2. 2Issue permanent IDs and lock the assignment file
  3. 3Monitor delivery against assignment and log every deviation
  4. 4Reconcile each follow-up round to the original panel and report attrition

Other services