# Relational Integration and Attentional Control Are Crucial to Fluid Intelligence Together but Not Alone—An Experimental Investigation of Individual Difference in Relational Monitoring Processes

## TL;DR

This study explores how working memory and fluid intelligence are linked through processes like relational integration and attentional control.

## Contribution

The study introduces a new experimental approach to better understand how relational integration and attentional control jointly relate to fluid intelligence.

## Key findings

- Neither relational integration nor attentional control alone significantly relates to fluid intelligence.
- Combining these processes provides the strongest predictive power for fluid intelligence.
- How performance is measured significantly impacts the conclusions about cognitive processes.

## Abstract

Working memory (WM) and fluid intelligence (Gf) are highly correlated, which provides the basis for the claim that they share common cognitive processes. Attentional Control Theory and the Relational Integration Hypothesis are two process theories linking WM and Gf. Additionally, both have empirical evidence to support them; the strength of this evidence can be limited by the experimental manipulations used and the operationalisation of performance metrics. To investigate the cognitive processes related to Gf, levels of relational integration and attentional control in the relation monitoring task (RMT) were manipulated. Study 1 (N = 39) focused on calibrating RMT response time windows for different levels of relational integration to strengthen validity claims by reducing possible ceiling effects in RMT performance observed in prior research. Study 2 (N = 146) examined how Gf was related to manipulations of relational integration and attentional control. The research extends previous studies by (a) using experimental manipulations that align more closely to underlying process accounts, and (b) contrasting simple-composite scores, a common operationalisation of performance, with a variance decomposition approach that statistically isolates the hypothetical processes aligned with the experimental manipulations. Results suggest that the way performance is operationalised matters, and that neither relational integration nor attentional control processes alone relate to Gf; instead, predictive utility is greatest when they are operationalised together.