Reading focus
Syllabus: What Is Intelligence?
Audit the task choices and reliability before interpreting the factor results. Give particular attention to the authors' caution about the musical measures.
Why you’re reading it. Bring a broad theoretical claim into contact with evidence while evaluating whether the test represents that claim fairly.
A question to consider. If the programme's proposed social and spatial domains correlate with its general score, what pattern would challenge independence, and what measurement objections would need to be addressed first?
Why this reading
A broad theory needs an interpretable empirical test. Visser and colleagues construct tasks intended to represent eight proposed intelligence domains and examine how performance covaries. The study exposes both the force of shared cognitive variance and the difficulty of operationalizing an opponent's theory fairly.
Before you read
The sample comprises 200 adults, mostly university students. The design includes two tasks per proposed domain and an external cognitive test. Factor loadings describe associations with modeled dimensions, not the importance of an ability in life. Poor reliability can weaken observed correlations, making apparent independence ambiguous.
Read for
- Inspect how each proposed domain is translated into tasks before deciding whether the analysis tests the intended theory.
- Distinguish evidence of a shared general factor among cognitive tasks from residual group structure and from results involving less cognitively demanding tasks.
- Attend to the low reliability of the musical measures: the authors explicitly limit conclusions about that domain. Consider sample selection and restricted variation as further limits.
Consider afterwards
- Two tasks intended to measure the same intelligence correlate weakly with each other and with other tasks. What evidence would distinguish a genuinely separate capacity from unreliable or poorly chosen measures?
Editorial context
The findings challenge strong independence claims for several measured domains; they do not settle the value of every human capacity or every educational use of varied activities. In a published response, Gardner disputes whether the measures adequately represent his theory; the authors' reply defends the test. This is a separate exchange from Gardner and Moran's response to Waterhouse. The core question is whether disagreements can be translated into a better test, rather than resolved by relabeling unwanted findings.