Technical evidence should be read as a collection of studies with identifiable questions, versions, and populations. A historical result retains documentary value, but its scope does not automatically transfer to a current version or every context of use.
Research has a history
The AMITAI Honestidad foundations manual brings together background from different stages of the instrument. This history helps explain which questions were explored and how development was documented.
Accurate reading preserves each study’s identity: version, date, population, method, and purpose. Combining results from different stages as though they describe a single assessment removes essential information.
Version, scale, and population
An instrument’s components or administration may change. Evidence from an earlier version provides background, but its continuing relevance needs examination when the instrument or use changes.
The level of analysis also matters. Findings about one scale do not automatically describe an entire battery. A sample from a particular population does not by itself represent every country, sector, or role.
Understanding sample sizes
Identify whether a number counts people, administrations, or records. One person may participate more than once, and a subset may come from a larger database; adding both figures would duplicate information.
- What was the unit of analysis?
- How were records selected?
- Do samples overlap?
- Which data were excluded, and why?
- To which population do conclusions apply?
Match the method to the claim
Internal consistency, stability over time, scale structure, and relationships with an external criterion address different questions. A performance claim needs a method capable of examining that performance.
Group comparisons and statistical associations need context. Correlation does not establish causation, and a score distribution is not equivalent to a study of fairness or absence of bias.
Build an evidence record
- Research question and proposed use.
- Instrument, version, and scale.
- Period, population, sample, and selection.
- Procedure and analysis.
- Results and limitations.
- Reference to the complete report.
This record supports comparison without erasing differences. It also reveals what information is missing when a new use raises a new question.
From background to current use
Before adopting a conclusion, request documentation relevant to the version and context you will use. Check whether the study supports the precise interpretation you need.
Research history provides memory and context. Sound technical communication explains findings, limits, and questions requiring additional evidence, without turning background research into promises of outcomes.
Frequently asked questions
Is an older study still useful?
Yes, as background with a defined scope. Relevance to a different use or version needs examination.
Can sample sizes from different reports be added?
Only when composition is known and overlap is ruled out. People, administrations, and records are different units.
Sources for this reading
AMITAI Honestidad foundations manual, pages 32, 38–44. An educational guide adapted from its concepts; it reproduces no individual reports and does not replace documentation for the version used.
- AERA, APA y NCME · Estándares para Pruebas Educativas y Psicológicas ↗
- Reidl-Martínez · Confiabilidad en la medición ↗
An AMITAI Help Center publication.
