Cognitive Skills and the Measurement of Social Networks

How to cite

Herath, Nilmini; Piraino, Patrizio; Caldeira, Raquel. 2026. "Cognitive Skills and the Measurement of Social Networks." AEA RCT Registry. AEARCTR-0017988. https://doi.org/10.1257/rct.17988-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Nilmini Herath* , Patrizio Piraino , Raquel Caldeira

* Primary investigator

Registry ID
AEARCTR-0017988
DOI
https://doi.org/10.1257/rct.17988-1.0
First registered
Last updated
Status
On going
Trial dates
2025-11-01 – 2026-04-30
Countries
South Africa ()
JEL codes
C81, C93, D81
Registry record
http://www.socialscienceregistry.org/trials/17988

Abstract

Measuring social networks accurately presents significant methodological challenges, as survey-based network elicitation requires respondents to recall and report complex relational information — a cognitively demanding task. Despite growing recognition that poverty and scarcity consume mental "bandwidth", the implications for network measurement remain under-explored. This study addresses this gap through a survey experiment with approximately 900 entry-level workers in South Africa, in which respondents are randomly assigned to one of three elicitation methods for their labor market networks. The first arm uses standard auditory elicitation, in which questions are read aloud by an enumerator and respondents recall contacts sequentially from memory. The second arm uses a decomposed auditory format, which scaffolds the recall task by prompting respondents to think through their networks systematically by setting or relationship type before providing counts. The third arm uses a visual elicitation method — implemented via Network Canvas software — which allows respondents to see their network as they build it, with contacts represented as visual nodes, reducing demands on working memory and making the elicitation process more intuitive.

A key feature of the study is the systematic measurement of respondents' cognitive skills prior to the network elicitation task. We measure cognitive skills in three ways: through a Digit Span task, a self-reported measure and an enumerator-reported measure of participant comprehension. This allows us to test not only whether more structured and/or visually supported formats produce larger and more internally consistent network reports on average, but also whether gains are concentrated among respondents with lower cognitive capacity, for whom standard auditory elicitation may be most burdensome. We evaluate elicitation formats on reported network size, network composition, and the predictive validity of reported networks against an external criterion: whether respondents obtained their current job through a network contact. These findings will generate evidence on whether commonly used auditory survey methods systematically misestimate network size and composition for precisely those populations most targeted by network-based development interventions.

Intervention

This is a test of different methods of ARD elicitation for labor market networks.

T1: Standard auditory elicitation. Questions are read aloud by an enumerator to define the network and respondents recall contacts from memory.

T2: Decomposed auditory elicitation. This scaffolds the recall task by prompting respondents to think through their networks systematically by setting or relationship type before providing counts.

T3: A decomposed visual elicitation method implemented via Network Canvas software. This allows respondents to see their network as they build it, with contacts represented as visual nodes, reducing demands on working memory and making the elicitation process more intuitive.

Experimental design

We conduct a three-arm randomized controlled trial with approximately 900 entry-level workers in South Africa. Respondents are randomly assigned at the individual level to one of three network elicitation methods. Arm 1 (standard auditory) uses traditional survey questions read aloud by an enumerator, requiring respondents to recall contacts sequentially from memory. Arm 2 (decomposed auditory) scaffolds the recall task by prompting respondents to think through their networks systematically by setting or relationship type before providing counts, delivered verbally. Arm 3 (decomposed visual) uses Network Canvas software, which allows respondents to see their network as they build it with contacts represented as visual nodes, reducing working memory demands while also decomposing the recall task.

The primary comparison of interest is between Arm 1 and Arm 3, which represent the most distinct elicitation approaches. To maximise statistical power for this comparison, these arms are each allocated approximately 375 respondents. The remainder of the sample is assigned to Arm 2, which serves a diagnostic role: by isolating the effect of decomposition from the effect of visual presentation, it helps identify the mechanism through which any differences between Arms 1 and 3 operate — whether differences in network reporting are driven by the structured decomposition of the recall task, the reduction in working memory demands afforded by visual elicitation, or some combination of the two.

All respondents complete a battery of cognitive skills assessments. The primary measure, performed prior to the network elicitation, is the Digit Span subtest from the Wechsler Adult Intelligence Scale, in which participants repeat sequences of numbers in forward and backward order at increasing levels of difficulty, capturing numerical working memory. This is complemented by two subjective measures: a self-reported measure in which respondents rate the ease, comprehensibility, and effort required by the survey task on 5-point Likert scales; and an enumerator-evaluated measure in which field officers independently rate each participant's comprehension and communication. These measures allow us to examine whether treatment effects on network reporting are heterogeneous by cognitive capacity, and whether objective and subjective indicators of cognitive load predict measurement quality differently.

Design details

Randomization method
Randomization is performed in Qualtrics prior to the survey. Respondents are allocated to one of three groups.
Randomization unit
Individual
Clusters
900
Observations
900
Arms
Note that the main comparison is between T1 and T3. A smaller number is allocated to T2, an intermediate approach, which is included to understand the mechanism through which any differences occur. T1 375 participants T2 150 participants T3 375 participants
Primary outcomes
We explore three measures of network 'quality', in terms of labour market informativeness. Total network size Number (or proportion) of network members in a formal job Number (or proportion) of network members with tertiary education
Secondary outcomes
For the validation exercise, we look at the correlation between primary outcomes and whether the participant obtained their current job through their network - a sign of network quality.

Cited by

No citation data available for this registration yet.

Keywords