Rotating eigenvalues obliquely clarifies clustered survey items orthogonal rotation blurs

Evidence

In the results for the Societal scale, the first component had an Eigenvalue of 7.06 (37% of the total variance), which dropped to 2.06 (10% of the total variance) in the second component (Table 4). Parallel analysis findings supported retaining the first two components. However, an orthogonal rotation of the components did not result in clarity regarding item groupings by component. Therefore, an oblique rotation of the two components was assessed. Oblique rotation resulted in four items loading strongly on the second component, an increase of three items from the orthogonal rotation. These four items dealt with public regard of Muslims (media depictions of Muslims) and structural anti-Muslim discrimination (government surveillance). — Naheed Ahmed, Measurement of Perceived Interpersonal…, 69

Categorization

• XREF: Connects to the ‘structure of the psyche mapped statistically’ idea — eigenvalue clustering as a way of finding latent dimensions like the doer/done-to axis.
• QUESTION: Do media depiction and government surveillance genuinely form one distinct dimension of perceived discrimination, or is this a sample artifact?