Assets, Sector or Ethnicity? Structural Divisions in Political Attitudes and Online Participation among University Students in Lahore

Abstract

Analyses of political behaviour in Pakistan conventionally treat economic position as the organising social cleavage, and research on digital politics has inherited that assumption largely intact. Among university students who are already online, however, it remains an open question whether household economic standing differentiates the political experience they have there. This article places three candidate divisions in competition — household asset ownership, ethno-linguistic group, and the public or private character of the university attended — against four outcomes: perceived echo-chamber enclosure, institutional trust, conspiracy mentality, and online political action. The data come from a cross-sectional survey of 779 students at twelve universities in Lahore, conditioned by design on students who are already online. Because institutional sector is a property of only twelve institutions, two of which supply 59 per cent of respondents, inference is cluster-robust with Satterthwaite degrees of freedom, and every institutional claim is stress-tested by dropping the two dominant universities and by weighting universities equally. Measured household asset ownership shows no statistically detectable association with any outcome among students who disclosed it, and the null holds under count, principal-components, categorical and item-by-item specifications. One institutional association retains its sign and nominal significance across every composition check: public-sector students report less online political action. Differences by ethno-linguistic group, and a large association between governing-party identification and institutional trust, are reported as exploratory; none of the thirty-two primary associations survives Benjamini–Hochberg correction across that family. Within this sample, institutional and partisan differences are more visible than differences associated with measured household assets, and a design with twelve institutional clusters can pose the question of institutional environment but cannot settle it.