Norms Behind Closed Doors:
Misperceptions and Maternal Employment in Couples

Main results.

• Beliefs: the binding misperception (men about other men's support) falls −11 pp**; spousal belief about partner +6 pp** (spillover, both genders).
• Intra-household decision: treated men are 23 % more likely (+9.1 pp) to nominate their wife for an online career-development course instead of taking it themselves.
• Labor outcomes: treated women search for work / change jobs more (+9.6 pp***, RW p = 0.013); men's work–family balance preference +11 pp*. Aspirations null.
• Boundary: the chain runs end-to-end only among labor-attached women — where structural barriers are already low.

The LAC puzzle. Female LFP stalled around 2002 even as fertility fell, education rose, and private support for women's work approached parity (Gasparini & Marchionni 2015; Marchionni, Edo & Berniell 2024). The deceleration is a normative puzzle, not an income or institutional one.

The question. If norms — not preferences, not institutions — are the binding constraint, then correcting beliefs about the norm should move behaviour.

Why complementary. Without ①, the community correction would not trigger discussion within the household nor reach beliefs about the partner; without ②, updated beliefs would not translate into intra-household action. We cannot fully isolate ①'s channel — but the couples design lets us measure how within-household beliefs and decisions move together after treatment.

Pluralistic ignorance runs on both sides. Both husbands and wives underestimate community support for maternal employment — sharply for men's support (≈ 30 pp), more mildly for women's (≈ 10 pp). The intervention corrects misperceptions about both by giving each spouse the true Bogotá distribution.

Step 2 — IPWRA within each sample S

Step-1 weights w_i^S enter as pweights. The estimator combines:

• Treatment model e(X_i^D, strata) = Pr(D=1│X, strata) → propensity score
• Outcome models m_d(X_i^D, strata) = E[Y│D=d, X, strata], d ∈ {0,1}

Reweights the control group toward the covariate distribution of the treated; adjusts flexibly for remaining differences via the outcome regression.

Doubly-robust property. The ATT is consistently estimated if either the treatment model e(·) or the outcome models m_d(·) are correctly specified — we do not need both. Reported with cluster-robust analytic variances at the household level.

Multiple-testing correction · Romano-Wolf step-down (Appendix Table A·rwolf)

• Step-down adjustment within each outcome family — F1 community beliefs · F2 spousal beliefs · F3 course outcomes · F4 labor outcomes
• Controls family-wise error rate while preserving power vs. Bonferroni
• Computed on unweighted OLS → conservative relative to our preferred IPWRA estimates (treats RW as a stress test of the headline pattern)

Headline robustness. Women's job mobility survives Romano-Wolf double-starred (RW p = 0.013). Men's course nomination and work–family balance are suggestive under RW but the Fisher p-values remain below 0.10.

Policy implication. Information is a low-cost lever where structural conditions already permit. Where they don't, norms and structure are complements, not substitutes — childcare, demand, and certification are the binding tools.

⑤ Spillover to partner-specific beliefs (RQ2)

Treated respondents update beliefs about their own partner, whose views they were never told. Inventing partner-specific beliefs to please the experimenter is implausible — these are private, idiosyncratic targets. The spillover (+6 pp for both) is hard to reconcile with SDB.

One exception we flag honestly: men's work–family balance preference (+11 pp ★) is a stated preference; the timing placebo helps but a demand effect cannot be fully ruled out. We interpret it conservatively (paper §6.3).

⚠ Caveat — endogeneity: we did not randomize which spouse engaged with the WhatsApp module. Direct / indirect / joint exposure configurations are endogenous. We therefore interpret these results as descriptive, not causal.

Across all 8 gender-norm items, the same pattern holds: progressive private attitudes coexist with sizable misperceptions about others' views, particularly men's support. The placebo shows near-universal agreement and no misperceptions — confirming misperceptions are norm-specific, not general pessimism.

What we do: Romano-Wolf adjusts each p-value by simulating the joint distribution of all test statistics under the null (1,000 bootstrap replications, clustered by household). It produces a family-wise error rate–controlled p-value for every outcome, accounting for the dependence structure between them.

Plain English: "Even if I throw lots of outcomes at this experiment, here is the p-value adjusted for the fact that I am fishing in many ponds. Results that survive RW are not lucky strikes."

Notes: Romano-Wolf computed on OLS (unweighted) — conservative vs. IPWRA. 1,000 replications, seed(12345), clustered by household. Women's job mobility survives stepdown correction (RW p=0.013). Course nomination for men is marginal (RW p=0.077).

What we do: Trim observations from the lower-attrition arm to make response rates equal across treatment and control. Then compute the worst-case and best-case ATT — the interval brackets all possible values consistent with monotonicity (treatment doesn't change who attrits).

Plain English: "Imagine the absolute worst possible scenario about who dropped out. Even then, my treatment effect lies somewhere in this range. If both ends of the range exclude zero, my result holds even under unobservable bias."

Interpretation: Lee bounds apply when treatment monotonically increases probability of being in sample. For the course (Sample 2), this is satisfied by design (engagers). Both bounds strictly positive for men's course nomination and women's job search — key results hold under worst-case attrition scenarios consistent with monotonicity.

What we do: re-estimate the IPWRA ATT under 4 alternative weight constructions to check that headline results don't depend on the most extreme weights:

• (i) Baseline: probit propensity score (preferred)
• (ii) Winsorised: cap weights at 95th percentile
• (iii) Trimmed: drop observations with PS < 0.10 (≈ 1% of obs)
• (iv) Logit PS: alternative functional form for the selection model

Plain English: "If a handful of unusual observations were driving my result, the estimate would change a lot when I cap or drop them. It doesn't change → my result is robust, not artifact of extreme weights."

After IPWRA weighting, maximum absolute standardized mean differences (SMDs) are below 0.10 in all samples and genders. Some covariates show marginal imbalance in Sample 2∩3 (joint F tests reject), but effect sizes are small, and post-weighting balance is tight. The key variable — second-order belief about men's community support — does not differ significantly across treatment and control arms in any sample.

Propensity scores range from 0.06 to 0.76; overlap is adequate in all samples. Effective sample sizes remain large after weighting; mass outside common support is small. P-score densities from 0.06–0.76 → no extreme regions of non-overlap that would invalidate IPWRA.

IPWRA is the preferred specification chosen a priori to address selection into midline take-up. OLS and weighted-OLS produce nearly identical point estimates across all headline results — the choice of estimator does not drive the findings.

Job mobility effects are concentrated among employed (+14.8 pp) and unemployed (+8.9 pp) women. Inactive women show near-zero effects. The course nomination effect is also largest when the wife is employed (+13.4 pp, p<0.10). Together, these results suggest information works at the margin where action is already feasible.

Engager characterization (selective take-up): only 36% engage with the WhatsApp module. We compare engagers vs. non-engagers on (a) demographics and (b) the targeted second-order belief.

Plain English: "Engagers are more inactive (selection on demographics — fix with IPW step 2). But they hold the same prior on community support as non-engagers — so the disclosed norm is accurate for them, the people who actually received it."

Reference-group accuracy: max deviation of any subgroup's mean SOB from city-wide average is 3.3 pp. The misperception we are correcting is 28 pp → reference-group mismatch is < 12% of the corrected signal. Disclosed Bogotá-average norm is a valid proxy for every demographic subgroup.

Reference group accuracy: Max deviation of any subgroup's mean SOB from city-wide average = 3.3 pp (high-SES). The corrected misperception is ~28 pp → reference group mismatch is <12% of the corrected signal. Disclosed norm is valid for all demographic subgroups in the sample.

The mediation exercise supplements rather than replaces the reduced-form evidence. We treat it as a consistency check on the sign pattern and direction of the channels.

Replication data and code available at doi.org/10.7910/DVN/QYWHLA.