The Same Information, Two Gendered Responses

The Delhi experiment's gender contrast is not a story about men and women having opposite levels of motivation. It is a story about the same information meeting different starting conditions. This chapter traces how belief revision and access to search channels produced lower search for some men and a shift toward digital search for women. The evidence supports a useful distinction: information can change the expected return from searching, or it can make a previously difficult route more usable. Those mechanisms can point in different directions, and some links between them remain suggestive rather than directly measured.

The difference begins with the kind of problem the information is addressing. Men in the sample entered with more upward-biased salary expectations and greater reliance on family, friends, and neighbors for leads. Women had lower employment exposure, were less likely to apply even when they had searched, and depended relatively more on formal or digital channels. A common intervention could therefore correct a forecast for one group while opening access for another.

Beliefs moved before choices fully separated

Six to twelve months after the intervention, people assigned to Personalized messages were less likely to report that they were working in or toward their ideal job. The effect was 6.2 percentage points at the first endline and 5.0 points at the second for the full Personalized group. Among men, the corresponding reductions were 8.4 and 5.6 percentage points in the main estimates. Personalized treatment also lowered the perceived earnings-gain category for an ideal job: the estimate was -0.398 for the full sample and -0.458 for men.

These measures suggest that the intervention changed expectations about attainability and pay, not merely the amount of search. The strongest evidence appears among men whose baseline wage expectations were upward-biased. For that subgroup, the Personalized estimate for perceived earnings gain was -0.426. In other words, people who began furthest above the benchmark showed the clearest downward revision when exposed to tailored information.

The outcome is not a direct measure of a revised rupee forecast. Respondents selected an ordered category describing the earnings gain they expected from an ideal or aspiring job. Nor does a lower report of being on track mean that a person no longer values the goal. It may mean that the person now sees the goal as more distant, has altered its timing, or has become less confident that the current route will reach it. The survey establishes a change in belief, while the reason and consequence of that change must be inferred from related behavior.

Women's expectations also moved downward in some measures, but the effects were generally imprecise. That difference is consistent with the baseline pattern: women's expected earnings were still above the benchmark on average, but less so than men's. Information had less inflated optimism to correct. Their more distinctive constraint may have been finding opportunities through a usable channel.

There are hints that the belief response affected plans beyond search. Men in the Personalized group were more likely to remain enrolled in education at the later endline, particularly if they had been enrolled at baseline.