The effect of number of children on loneliness: a nationally-representative instrumental-variable study in India

Abstract Background This study aimed to investigate the causal effect of the number of children on loneliness among older parents in India. Methods We used nationally representative data from Wave 1 of the Longitudinal Ageing Study in India (LASI). Our analytical sample comprised individuals aged 50 years and older with at least one child. We employed two-stage least squares estimation using sex of firstborn child as an instrumental variable to estimate causal effects. We stratified analyses by sex and residential setting, and conducted exploratory analyses of downstream social and family-arrangement indicators, including social engagement and proximity to children. Results In the full sample, each additional child was associated with higher loneliness scores in later life (β = 0.076, 95% CI 0.019–0.132; P = 0.009). However, significant sex differences were observed. Among fathers, the effects of each additional child significantly reduced loneliness (β = -0.095, 95% CI: [-0.187, -0.003], P = 0.043) but increased loneliness among mothers (β = 0.218, 95% CI: [0.137, 0.298], P < 0.001). Residence-stratified analyses suggested that these sex-differentiated patterns were most evident in rural areas. Exploratory analyses further suggested that children’s proximity moderated these associations: co-residence and nearby residence attenuated adverse effects for mothers by 16.87% and 22.21%, respectively, while strengthening protective effects for fathers by 4.52% and 14.00%. Conclusions Additional children were associated with lower loneliness causally among fathers but higher loneliness among mothers, particularly in rural India. These findings challenge the assumption that larger families uniformly protect older adults from loneliness and highlight the need for gender-sensitive and community-based approaches to later-life social support.