{"id":"5640c099-1190-4ed0-8cb3-739ba79d1f41","arxiv_id":"2606.02503","paper_version":3,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Biased pay beliefs distort amenity-pay tradeoffs in stated preferences, with full disclosure restoring alignment to full-info benchmarks while short-term info does not.","lead":"The paper finds that workers underestimate starting salaries by 18% and overestimate how much more high-amenity jobs pay. Full pay disclosure corrects their choices to match full-information benchmarks, while short-term information only shifts beliefs without changing perceived tradeoffs.","discovery_kind":"new_application","skeptic_critique":{"model":"grok-4.3","headline":"Hypothetical choice experiments may not recover real-world amenity-pay tradeoffs even with incentives","rationale":"The reader's weakest_assumption correctly isolates the load-bearing assumption. With the full manuscript now available, the experimental details (randomization, incentive structure, belief elicitation) can be inspected, but they do not address the external-validity gap between stated and revealed preferences. This concern is internal to the argument rather than a disagreement with external consensus; it directly conditions how much weight the 4% disclosure effect and the recovered WTP alignment can receive. The verdict therefore moves from UNVERDICTED (information-limited) to CONDITIONAL pending external validation.","tokens_in":1678,"tokens_out":401,"duration_ms":15887,"concrete_test":"Field-validate a random subset of respondents by offering them real, incentivized job opportunities whose pay and amenity attributes are drawn from the same design space used in the survey; compare acceptance probabilities and implied marginal rates of substitution against the survey WTP estimates. A statistically significant divergence would indicate that the disclosure effects observed in the experiment do not generalize.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim—that full pay disclosure raises chosen pay by ~4% and aligns WTP estimates with full-information benchmarks—rests entirely on within-survey variation in stated choices across information conditions. The design elicits beliefs and presents hypothetical job profiles whose pay is either disclosed or not; all outcome measures (chosen pay, implied amenity WTP, belief updating) are derived from these responses. For the 4% effect and the alignment result to speak to actual labor-market behavior, the hypothetical choices must map to real tradeoffs. Nothing in the reported design (incentivized survey, multi-stage structure) rules out systematic hypothetical bias arising from low real stakes, social-desirability effects on amenity valuation, or the absence of search frictions and outside options that exist in actual job markets.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.3","summary":"The paper reports results from a multi-stage incentivized survey experiment that elicits beliefs about starting salaries in real jobs and presents hypothetical job profiles with randomized pay disclosure. It finds that respondents under-predict salaries by 18% and over-predict the amenity-pay gradient; a short-term information intervention raises mean pay beliefs by 4% and reduces dispersion by 15% but leaves perceived tradeoffs unchanged; full pay disclosure in the choice stage raises chosen pay by ~4% and produces WTP estimates aligned with full-information hypothetical benchmarks.","tokens_in":1807,"tokens_out":403,"duration_ms":14159,"significance":"If the mapping from stated choices to real labor-market behavior holds, the results provide direct evidence that information about pay can correct biased beliefs and restore amenity-pay tradeoffs, with implications for pay-transparency policies. The design's combination of belief elicitation, randomization of disclosure, and within-subject choice variation is a methodological strength that cleanly separates belief updating from choice effects.","major_comments":[{"comment":"Abstract and experimental-design section: the central claim that full disclosure 'recovers willingness-to-pay estimates closely aligned with full-information hypothetical-choice benchmarks' and raises chosen pay by 4% is load-bearing for the paper's interpretation, yet the design contains no validation (e.g., comparison to actual job offers, revealed-preference data, or external benchmarks) that the hypothetical choices are free of hypothetical bias. Without such evidence the 4% effect and alignment result cannot be taken to speak to real amenity-pay tradeoffs.","section":"Abstract / experimental design"}],"minor_comments":[{"comment":"The abstract states that the short-term intervention 'does not alter the perceived amenity-pay slope,' but the manuscript should report the precise statistical test and standard error for this null result to allow readers to assess power.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":{"model":"grok-4.3","summary":"We thank the referee for the constructive feedback. Below we respond point-by-point to the major comment on the interpretation of hypothetical choices.","responses":[{"response":"We agree that the experiment relies on stated choices and contains no external validation against revealed-preference data, actual job offers, or other real-world benchmarks. The manuscript's claims are scoped to the stated-preference setting: the abstract explicitly compares results to 'full-information hypothetical-choice benchmarks,' and the design uses randomization of disclosure to identify how information affects choices within this controlled environment. The 4% shift in chosen pay and the alignment result therefore speak directly to the correction of biased beliefs in hypothetical tradeoffs. While we acknowledge that hypothetical bias remains a concern for extrapolating to real labor-market behavior, the paper does not claim to provide direct evidence on real amenity-pay tradeoffs; implications for policy are presented as suggestive only. We do not view the absence of external validation as undermining the internal validity of the information-provision effects documented.","revision_made":"no","referee_comment":"[Abstract / experimental design] Abstract and experimental-design section: the central claim that full disclosure 'recovers willingness-to-pay estimates closely aligned with full-information hypothetical-choice benchmarks' and raises chosen pay by 4% is load-bearing for the paper's interpretation, yet the design contains no validation (e.g., comparison to actual job offers, revealed-preference data, or external benchmarks) that the hypothetical choices are free of hypothetical bias. Without such evidence the 4% effect and alignment result cannot be taken to speak to real amenity-pay tradeoffs."}],"tokens_in":1314,"tokens_out":345,"duration_ms":22495,"standing_objections":[]},"desk_editor":{"model":"grok-4.3","letter":"The main takeaway is that workers underpredict starting salaries by 18 percent and overstate how much higher-amenity jobs pay, and that giving full pay information for the exact jobs in the choice task corrects the implied willingness-to-pay numbers while a one-time information nudge does not. The design that produces these numbers is straightforward: elicit beliefs about real jobs, run hypothetical choice experiments, and randomly vary whether pay is shown or not. That separation between temporary and full disclosure is the clearest new piece.\n\nThe paper does a good job keeping the experiment incentivized and multi-stage, which reduces some obvious demand effects. The quantitative patterns line up internally: short disclosure moves average beliefs and cuts dispersion but leaves the amenity-pay slope untouched, while full disclosure shifts both beliefs and the choices themselves by about 4 percent in pay. Those contrasts are useful for thinking about pay transparency rules.\n\nThe soft spot is the usual one for this style of work. All the outcome measures come from stated choices over hypothetical profiles. Nothing in the design directly tests whether the 4 percent shift or the recovered willingness-to-pay numbers would appear in actual applications or job offers where search costs, outside options, and real stakes are present. The abstract does not report any validation against administrative data or field behavior, so the policy implications rest on the assumption that the survey maps closely enough.\n\nThis is the kind of paper that belongs in a labor or behavioral economics reading group. Readers working on information frictions or pay transparency will want to see the exact belief-elicitation protocol and robustness checks. It is worth sending to referees because the design is replicable and the question is live, even if the external-validity step will need more evidence later.","headline":"The experiment cleanly separates short-run belief shifts from persistent disclosure effects on stated tradeoffs, but external validity to real job choices remains the main open issue.","tokens_in":2281,"tokens_out":421,"would_cite":false,"duration_ms":9017,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.3","headline":"Full disclosure of pay for jobs under consideration raises the pay of chosen jobs by about 4% and aligns willingness-to-pay estimates with full-information benchmarks.","keywords":["pay beliefs","amenity-pay tradeoff","job choice","information disclosure","willingness to pay","salary expectations","workplace amenities","survey experiment"],"falsifier":"A field experiment that randomly provides or withholds pay information for actual job offers and measures whether the 4% upward shift in accepted pay and the alignment of amenity tradeoffs appear in real acceptance decisions.","tokens_in":2576,"feed_emoji":"💼","tokens_out":768,"duration_ms":18862,"temperature":0.7,"pith_summary":"The paper tests how beliefs about salaries shape workers' willingness to trade pay for amenities like workplace conditions. Respondents under-predict starting salaries by 18% and expect higher-amenity jobs to pay more than they actually do. Short-term pay information raises beliefs by 4% and cuts dispersion by 15% but leaves the perceived amenity-pay slope unchanged. Full disclosure of pay for the specific jobs being considered shifts choices toward higher-paying options by 4% and brings implied tradeoffs in line with benchmarks from fully informed hypothetical choices. This shows that the form and persistence of pay information matter for correcting belief-driven distortions in job selection.","feed_headline":"Pay disclosure shifts chosen jobs to 4% higher salaries","feed_subtitle":"Workers under-predict salaries and overstate the pay-amenity link; revealing exact pay for options under review corrects both beliefs and tr","key_machinery":"The multi-stage incentivized survey experiment combining hypothetical choice experiments with elicited beliefs about starting salaries in real jobs, with random variation in the provision of explicit pay information.","core_discovery":"Workers under-predict starting salaries by 18% and expect higher-amenity jobs to pay more, substantially over-predicting the amenity-pay gradient. A short-term pay information intervention raises mean beliefs about pay in similar jobs by 4% and reduces belief dispersion by 15%, but does not alter the perceived amenity-pay slope or the implied tradeoffs in stated choices. Full disclosure of pay for jobs under consideration raises the pay of chosen jobs by about 4% and recovers willingness-to-pay estimates closely aligned with full-information hypothetical-choice benchmarks. Short-term disclosure thus moves beliefs but not perceived tradeoffs, while persistent disclosure erases biases in pay b","pith_inferences":["Job platforms that display salary ranges for listed positions could reduce the distortion in amenity valuations documented here.","Mandates for pay transparency in job postings might produce larger effects on choices than one-time information interventions.","Similar belief biases about compensation gradients could appear in other choice settings such as housing or education, where full price disclosure might realign tradeoffs.","Testing whether the 4% pay shift persists when workers face real budget constraints or repeated decisions would clarify external validity."],"forward_implications":["Correcting pay beliefs through full disclosure increases the salary level of jobs workers select.","Short-term information updates shift average beliefs and reduce uncertainty but leave the slope of the amenity-pay gradient unchanged.","Persistent disclosure of pay for specific options nearly eliminates the gap between stated tradeoffs and those observed under full information.","Biased expectations about the amenity-pay relationship can lead to choices that forgo higher pay without corresponding amenity gains."],"fun_headline_variants":["Workers underestimate starting salaries by 18 percent","Beliefs overpredict pay in high amenity jobs","Short disclosure changes pay views without affecting choices","Full pay reveal boosts selected job salaries by 4 percent"],"cache_read_input_tokens":64,"weakest_assumption_plain":"The hypothetical choice experiments in the survey accurately reflect real-world job preferences and tradeoffs, without significant hypothetical bias.","fun_headline_variants_meta":{"raw":{"variants":["Workers underestimate starting salaries by 18 percent","Beliefs overpredict pay in high amenity jobs","Short disclosure changes pay views without affecting choices","Full pay reveal boosts selected job salaries by 4 percent"]},"model":"grok-4.3","cost_usd":0.004207,"raw_usage":{"total_tokens":2141,"prompt_tokens":701,"num_sources_used":0,"completion_tokens":57,"cost_in_usd_ticks":42074500,"prompt_tokens_details":{"text_tokens":701,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":1383,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":701,"tokens_out":57,"duration_ms":9844,"temperature":1.0,"reasoning_tokens":1383,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-02T23:13:00.551194+00:00","model_set":{"reader":"grok-4.3"},"falsifier":"A field experiment that randomly provides or withholds pay information for actual job offers and measures whether the 4% upward shift in accepted pay and the alignment of amenity tradeoffs appear in real acceptance decisions.","supporting_citations":[],"review_version":2}