Reading focus
Syllabus: How Do We Build and Test Psychological Theories?
Ask what the data can distinguish before interpreting the name of a fitted model.
Why you’re reading it. Expose a different gap: excellent fit can leave the generating process unresolved.
A question to consider. What intervention or repeated observation would make two otherwise similar accounts predict different outcomes?
Why this reading
Fried distinguishes fitting a statistical model from supporting an explanation of what generated the data. Factor and network examples make the consequences of confusing those achievements concrete.
Before you read
A factor model summarizes shared variation through unobserved variables; a network model represents relations among variables. Different explanations can sometimes produce statistical patterns that the available observations cannot distinguish.
Read for
- Compare the generating process with the fitted model in the simulation.
- Identify where a claim about a construct or intervention needs assumptions beyond model fit.
Consider afterwards
- Two models fit the same symptom correlations. What additional evidence might distinguish their predictions about changing one symptom?
Editorial context
The target article critiques inferential practices, not every use of factor or network analysis. Its demonstrations do not establish that all models are equivalent under every design. Formalization can clarify a theory without validating its assumptions. The surrounding commentaries offer objections and extensions; they are separate works.