Sensitivity analysis for contamination in egocentric-network randomized trials with interference

Bar Weinstein, Daniel Nevo · 2026

Read Sensitivity analysis for contamination in egocentric-network randomized trials with interference by Bar Weinstein, Daniel Nevo (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-05553.txt.

Introduction

Trials that randomize treatment across people in a social network face a subtle flaw: networks sampled around each person look separate, but hidden links between them can carry effects across. Bar Weinstein and Daniel Nevo study this contamination in egocentric-network randomized trials, where each sampled person recruits some neighbors and treatment is randomized across the sampled people. Under a design-based framework, they show standard estimators go biased whenever such hidden links exist.

The authors derive bias-corrected estimators for direct and indirect effects. They also build a sensitivity analysis that uses parameters for the chance or count of missing edges, run through both grid checks and probabilistic bias analysis. Applied to a well-known HIV prevention trial study, the method suggests ignoring contamination can understate indirect effects and overstate direct ones. The tools give trial designers a way to test how fragile their conclusions are.

Catalog record

CREATORBar Weinstein, Daniel Nevo
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE121 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-05553

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Sensitivity analysis for contamination in egocentric-network randomized trials with interference Bar Weinstein, Daniel Nevo Abstract Egocentric-Network Randomized Trials (ENRTs) are increasingly used to estimate causal effects under interference when measuring complete sociocentric network data is infeasible. ENRTs rely on egocentric network sampling, where a set of egos is first sampled, and each ego recruits a subset of its neighbors as alters. Treatments are then randomized across egos. While the observed ego-networks are disjoint by design, the underlying population network may contain edges connecting them, leading to contamination. Under a design-based framework, we show that the Horvitz-Thompson estimators of direct and indirect effects are biased whenever contamination is present. To address this, we derive bias-corrected estimators and propose a novel sensitivity analysis framework based on sensitivity parameters representing the probability or expected number of missing edges. This framework is implemented via both grid sensitivity analysis and probabilistic bias analysis, providing researchers with a flexible tool to assess the robustness of the causal estimators to contamination. We apply our methodology to the HIV Prevention Trials Network 037 study, finding that ignoring contamination may lead to underestimation of indirect effects and overestimation of direct effects. Keywords: Causal inference; Network experiments; Network sampling; Probabilistic bias analysis; Spillovers. 1 Introduction In causal inference and design of experiments, it is commonly assumed that there is no interference between units, that is, the treatment assigned to one unit does not affect the outcomes of other units ( 11 ) . However, researchers are often interested in settings where the treatment of interest can spill over from treated units to untreated ones, leading to interference between units. For example, in infectious disease epidemiology, vaccinating some individuals may protect unvaccinated individuals ( 22 ) . To estimate causal effects under interference, researchers often rely on detailed information about the network structure that encodes the relationships between units. Such data is typically assumed to be on a complete sociocentric network , where all units in the network and their connections are observed ( 26 ; 2 ; 17 ; 47 ; 39 ) . This finite population can be composed of a collection of disjoint clusters or a single connected network. However, collecting complete sociocentric network data is often logistically challenging and expensive, limiting the application of network-based causal inference methods or resorting to sub-networks sampled from the population network. A common sampling design is egocentric network sampling ( 36 ; 40 ) , where a set of egos (also known as indexes) are sampled from the population, and each ego recruits a subset of its neighbors in the population network as alters (non-indexes). The observed network is then assumed to be composed of disjoint ego-networks comprising each ego and its recruited alters. This design is attractive due to its logistical feasibility and accessibility. Such egocentric sampling can be combined with randomized experiments, where treatments are randomly assigned to the recruited egos, resulting in an egocentric-network randomized trial (ENRT) ( 7 ; 8 ; 10 ; 16 ) . ENRTs are increasingly used in various studies, such as HIV prevention ( 28 ; 29 ; 5 ; 38 ) , substance addiction ( 45 ) , and mental health ( 3 ; 13 ) . However, while most methods assume that the ego-networks are disjoint, alters may be connected to multiple egos in the population network, and egos may be connected to other egos, resulting in contamination between ego-networks. Thus, the egocentric sampling process may lead to missing edges in the observed network, resulting in misspecification of the network interference structure ( 50 ) . Previous analyses of ENRTs typically assumed that the ego-networks are non-overlapping in the population network ( 7 ; 8 ; 10 ; 16 ) . That is, alters are connected only to one ego, and egos are not connected to other egos. Such an assumption may not hold in practice and be difficult to verify with the observed data alone, absent additional validation studies or reference data. In this paper, we relax the non-overlapping networks assumption and allow for the possibility of overlapping ego-networks. Under a design-based inference paradigm, where the only source of randomness is the treatment assignment, we show that the Horvitz-Thompson estimators of the indirect and direct effects based on the observed ENRT data are biased whenever there is contamination between ego-networks. We develop a sensitivity analysis framework to assess the impact of possible contamination between ego-networks on the causal estimates, and we couple it with bias-corrected estimators of the indirect and direct effects in the presence of contamination between ego-networks. The bias-corrected estimators depend on sensitivity parameters that encode the expected number or the probability of missing alter–ego and ego–ego edges in the population network. These parameters can be specified using covariates and prior knowledge about suspected contamination. For the direct effect on egos, we require an additional sensitivity parameter that relates to the magnitude of the treatment and exposure interaction effect on the egos. We provide practical guidelines on how to specify these parameters in practice, under various scenarios. We show how our results can be implemented within a grid sensitivity analysis ( 41 ; 20 ) or a probabilistic bias analysis ( 21 ; 27 ; 18 ) to assess the robustness of the causal estimates to possible contamination. The proposed framework is illustrated through simulations and a real-world application. The HIV Prevention Trials Network (HPTN) 037 study ( 28 ) serves as a primary motivating example for our work. HPTN 037 was an ENRT that aimed to evaluate the efficacy of a peer education intervention to reduce HIV risk behaviors among people who inject drugs. In the trial, treated egos received training on the prevention of risky behaviors and were encouraged to share this information with their alters. While previous analyses of this study ( 28 ; 7 ; 10 ) assumed that ego-networks are disjoint, contamination remains a significant concern. Specifically, alters might inject drugs or have sexual relations with egos from other ego-networks, or egos might be connected to each other. In this paper, we relax the assumption of disjoint ego-networks and re-analyze the HPTN 037 study, assessing the robustness of the causal estimates to such contamination. Our analysis suggests that ignoring contamination may lead to underestimation of the indirect effect and overestimation of the direct effect. Our work relates to several strands of literature. First, it contributes to the growing literature on causal inference under interference where the true network structure is latent and only proxies are observed by the researchers ( 48 ; 31 ; 23 ; 6 ; 51 ) . Second, it adds to the literature on interference with sampled network data which has primarily focused on random node sampling or snowball sampling designs under linear-in-means parametric models ( 9 ; 44 ; 52 ; 35 ) . Third, it contributes to the literature on sensitivity analysis under interference, which has focused on unmeasured confounding ( 49 ) , or on settings where interference depends on two networks, but only one of which is observed ( 15 ) . Finally, it relates to the literature on cross-cluster contamination in cluster-randomized trials, which often recommend changing the clusters boundaries to reduce contamination (

Read the complete text