{"id":"1ae15406-9167-45b0-80c5-30ecb47bb109","arxiv_id":"2501.10501","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"QSOs with higher dust extinction preferentially show steep radio spectra and fewer peaked SEDs, linking radio emission to outflow-driven shocks.","lead":"Dusty quasars are four times more likely than blue ones to have steeply falling radio spectra, which points to shocks rather than compact jets as the radio source. The result supports the idea that dusty quasars are a short-lived 'blow-out' phase before the dust is destroyed and a normal blue quasar appears.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The headline 46% vs 12% steep-fraction difference is only marginally significant and is not robust to dropping a single dusty source; the paper's own continuous correlation is p=0.11 and becomes p=0.26 after removing the most extincted QSO.","rationale":"The reader's weakest assumption targets the LoTSS-removal of four upturned sources. That is a legitimate concern, but it affects the peaked/curved comparison more than the headline steep fraction: after refitting, the three non-dusty upturned sources become peaked, and the one dusty upturned source becomes flat, so the 46% vs 12% steep claim is not an artifact of that specific cut. The steep fraction difference is instead fragile because it is a 6-vs-3 count difference in a 38-object sample, and the paper's own continuous correlation is not significant. A single leave-one-out deletion of the most extincted source, whose removal the authors already test for the continuous relation, pushes the binned Fisher p above 0.05. This is the most load-bearing weakness for the central claim; the physical interpretation and peaked-fraction comparison are secondary. The paper is otherwise careful with new data and acknowledges model degeneracies; with a required robustness test and language softening, the existing conditional verdict stands.","tokens_in":43001,"tokens_out":11236,"duration_ms":114024,"concrete_test":"Perform a leave-one-out jackknife on the 2x2 classification table (dusty E(B−V)>0.1 vs non-dusty; steep α<−0.5 vs not), computing Fisher exact p (one- and two-sided) after excluding each of the 38 objects in turn, and also a permutation test with 10^4 random relabelings of the 38 sources. Report the p after excluding 1007+2853 and the minimum p across all leave-one-out deletions. If the minimum deletion p remains below 0.05 under the permutation distribution, the fraction difference survives; if not, the abstract should state the steep-fraction excess is tentative.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Core empirical claim is the dusty/non-dusty steep fraction difference. The paper never reports a p-value for this 2x2 comparison; a Fisher exact test gives p≈0.04, which is only marginal. More importantly, the difference is not robust to the removal of a single source. If the most extincted object, 1007+2853 (E(B−V)=0.59, steep BPL), is excluded, the counts become 5/12 versus 3/25 and the Fisher exact p rises above 0.05; removing any one of the other five dusty steep sources has the same qualitative effect. The paper's own continuous test is even weaker: Spearman ρ=−0.37, p=0.11 (Fig. 5), dropping to ρ=−0.27, p=0.26 after excluding that same object (Section 3.2), yet Section 4.1 calls this 'weak but significant'. The quoted binomial errors (46±12%, 12±4%) treat model classifications as fixed and omit this jackknife sensitivity. Thus the central 'more likely steep in dusty QSOs' statement is driven by a handful of objects and is not established at the conventional level without one high-extinction source. The LoTSS upturned reclassification is a secondary concern for the peaked/curved comparison, not the principal weakness of the headline fraction.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs 4–5 band radio SEDs between 0.144 and 3 GHz for 38 intermediate-redshift QSOs, combining new uGMRT Band-3/Band-4 observations with archival LoTSS, FIRST, VLASS, TGSS, and LoLSS data. The SEDs are fit with power-law, broken power-law, and curved models, and the best-fit shapes are used to classify sources as steep, flat, peaked, or curved. The central claim is that dusty QSOs (E(B-V)>0.1 mag) are more likely to show steep radio spectral slopes (alpha<-0.5) than non-dusty QSOs, with reported fractions of 46±12% and 12±4%, and that this steepness is evidence for an outflow-driven shock origin for the enhanced radio emission in dusty QSOs.","tokens_in":43285,"tokens_out":7540,"duration_ms":69853,"significance":"If the central claim is robust, the paper makes a valuable contribution to the debate on whether red/dusty QSOs represent a blow-out phase: it uses an independent data set (new uGMRT observations), carefully documents the data reduction and SED fitting, and combines the SED shapes with existing e-MERLIN size information. The authors are also commendably transparent about data availability and about the limitations of their sample size. However, the headline statistical evidence is currently fragile: the reported p-values do not support the wording used in the abstract and Section 4.1, and the main 2x2 comparison is not accompanied by a p-value and is sensitive to the removal of single objects. The interpretation is plausible but needs to be matched to the actual statistical strength of the result.","major_comments":[{"comment":"The headline claim that dusty QSOs are more likely to be steep is based on 6/13 versus 3/25 sources, but no p-value is reported for this 2x2 comparison. A Fisher exact test gives p≈0.04, which is only marginal, and the difference is not robust to the removal of a single source: excluding the most extincted object, 1007+2853 (E(B-V)=0.59), leaves 5/12 versus 3/25 and raises the Fisher exact p to approximately 0.08, above the conventional significance threshold. Because this comparison is the load-bearing result of the paper, the authors should report a p-value, include a jackknife or bootstrap sensitivity analysis, and temper the abstract wording accordingly.","section":"Section 3.2 / Abstract"},{"comment":"Section 4.1 describes the Spearman correlation between alpha and E(B-V) as a 'weak but significant correlation (rho=-0.37, p=0.11)'. A p-value of 0.11 is not significant at the conventional 0.05 level, and the paper itself notes in Section 3.2 that removing the most extincted QSO reduces the correlation to rho=-0.27, p=0.26. The phrase 'weak but significant' is therefore internally inconsistent with the reported statistics. The authors should either use a more accurate description ('marginally significant', 'tentative', or similar) or provide a statistical threshold and justify why p=0.11 should be considered significant for this small sample.","section":"Section 4.1"},{"comment":"The treatment of the four 'upturned' QSOs has a direct effect on the claimed excess of peaked SEDs in non-dusty QSOs. For three non-dusty QSOs (1046+3427, 1203+4510, 1630+3847), the LoTSS point is removed and the sources are reclassified as peaked; the paper's peaked fraction of 48±9% for non-dusty QSOs therefore depends on the assumption that the low-frequency LoTSS emission is a separate relic component rather than part of the same compact synchrotron spectrum. If the LoTSS emission belongs to the same component, these sources would instead be steep or otherwise different, and the peaked/non-peaked contrast would shrink. The authors should at least quantify how the peaked-fraction comparison changes if the LoTSS points are retained or if the four upturned sources are treated as unclassified.","section":"Section 2.6 / Table 3"},{"comment":"The conclusion that dusty QSOs have steeper spectra is partly entangled with the choice of the E(B-V)=0.1 mag split and with the adoption of alpha_high or alpha_low for broken power-law sources. The continuous correlation is already non-significant at p=0.11, so the binned comparison carries most of the weight, but the binned comparison uses a threshold that is not varied across a range of values. A simple sensitivity test varying the split between, say, E(B-V)=0.05 and 0.2 mag would show whether the 46% versus 12% difference is stable or is an artifact of the chosen boundary.","section":"Section 3.2 / Table 3"}],"minor_comments":[{"comment":"The Figure 3 caption contains a duplicated word: 'the the boundary' should be 'the boundary'.","section":"Section 3.2 / Figure 3 caption"},{"comment":"The text refers to the 'viral relation' for black hole mass estimation; this should be 'virial relation'.","section":"Section 4.3"},{"comment":"The abstract states the redshift range as 1.0<z<1.5 while Section 2.1 states 1.0<z<1.55; these should be made consistent (Table 1 shows z=1.53 for 1630+3847).","section":"Section 1 / Section 2.1"},{"comment":"In the introductory text, the frequency range is written as '0.144MHz–3GHz'; this should be '0.144–3 GHz' for clarity, and the same notation should be used consistently throughout.","section":"Section 2.4 / Section 2.3"},{"comment":"The DOI for the uGMRT data is split across a line break in an unclear way ('https://doi.org/10. 101...'), which will not resolve correctly; please provide the full URL in a machine-readable form.","section":"Data Availability"}],"recommendation":"major_revision","confidential_remarks":"This is a useful observational paper with new data and a plausible physical interpretation, but the headline result is currently supported by marginal statistics, and the abstract oversells the significance. The main fixes are within the manuscript's scope: report a p-value for the 2x2 comparison, add a jackknife/sensitivity analysis, correct the 'weak but significant' wording, and reframe the peaked-fraction comparison after the LoTSS-removal assumption. I would not reject, but I would ask for these revisions before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper gives us something real – new uGMRT 400/650 MHz data and 4–5 band radio SEDs for 38 QSOs – but the headline \"dusty QSOs are more likely to be steep\" is statistically marginal and rests on a handful of objects. It should be published, but only after the authors recalibrate the claims.\n\nWhat's genuinely good: constructing multi-band SEDs from 0.14–3 GHz for a sample with e-MERLIN sizes is work that needed doing. Separating peaked from power-law/broken-power-law classes, and then comparing the fractions against E(B-V), is a step beyond the two-point slopes in Glikman et al. (2022) and Rosario et al. (2020). The discussion of whether jets or winds produce the radio emission is balanced and honestly acknowledges that the Nims et al. wind model cannot reproduce the most radio-loud sources. The section on LoTSS-extended sources and possible restarted activity is speculative but clearly labelled.\n\nThe soft spots are statistical and presentational. The continuous correlation between alpha and E(B-V) has p=0.11, and it drops to p=0.26 after removing the most reddened QSO. That is not \"significant\" by any conventional standard, yet Section 4.1 calls it \"weak but significant.\" The binned fraction difference (46±12% vs 12±4%) looks impressive, but the 2×2 Fisher exact p is around 0.04, and removing any one of the six dusty steep sources pushes it over 0.05. The paper does not report that robustness test. The binomial errors quoted in the abstract treat the SED classifications as fixed, which they are not. The post-hoc removal of the LoTSS point for four upturned sources is understandable, but it shifts the peaked/curved comparison and needs a sensitivity analysis, not a brief justification.\n\nIn short: the data and the SED-fitting approach are valuable, and the trend is worth reporting as tentative. But the abstract and conclusions overstate it. A referee should demand: a proper p-value for the 2×2 table, a jackknife or leave-one-out test for the fraction difference, and consistent language (tentative trend, not significant) throughout. After that, this would be a solid contribution to the red QSO/AGN feedback literature.","headline":"A careful observational study with new uGMRT data that overstates a statistically marginal and fragile steep-vs-dust trend; the data are valuable but the abstract and conclusions need major qualification.","tokens_in":43923,"tokens_out":2533,"would_cite":false,"duration_ms":25535,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Dusty quasars are four times more likely to show steep radio spectra than non-dusty ones.","keywords":["quasars","radio spectral index","dust extinction","active galactic nuclei","outflows","shocks","uGMRT","e-MERLIN"],"falsifier":"Measure the low-frequency flux of the four upturned quasars with a radio telescope that resolves the extended lobes, such as the Very Long Baseline Array; if the compact core shows a spectral turnover at LoTSS frequencies, the two-component interpretation is wrong and the steep-spectrum fractions would change.","tokens_in":1340,"feed_emoji":"🔭","tokens_out":1443,"duration_ms":62799,"temperature":0.7,"pith_summary":"The paper combines new low-frequency uGMRT observations with archival data to build four-to-five-band radio spectra for 38 quasars, split by how much dust they contain. It finds that quasars with significant dust extinction (E(B-V) > 0.1 mag) are about four times more likely to have steep radio spectra (spectral index alpha < -0.5) than non-dusty ones (46% vs 12%), while non-dusty quasars more often show peaked spectra (48% vs 23%). The authors interpret this as evidence that the radio emission in dusty quasars comes from shocks driven by outflows, not from compact jets, supporting an evolutionary picture in which dusty quasars are a young, blow-out phase that clears the dust to reveal a normal blue quasar.","feed_headline":"Dusty quasars show steeper radio spectra","feed_subtitle":"46% of dusty quasars have steep radio slopes vs 12% of non-dusty, pointing to shock-powered outflows.","key_machinery":"The paper constructs radio spectral energy distributions (SEDs) from 0.144 to 3 GHz using new uGMRT 400/650 MHz flux measurements plus archival LoTSS, FIRST, VLASS, and TGSS data, and fits each SED with three models: a single power law, a broken power law, and a curved (peaked) model. Model selection uses the Bayesian Information Criterion with thresholds that prefer the simplest model, and each source is classified as steep, flat, inverted, peaked, or curved. The classification, together with 0.2-arcsecond e-MERLIN sizes from the parent sample, allows the authors to compare the radio SEDs of dusty and non-dusty quasars and to test whether the peaked sources are consistent with GPS/CSS jetted sources.","core_discovery":"The central discovery is an empirical connection between dust obscuration and radio spectral shape: dusty quasars preferentially show steep, optically-thin synchrotron spectra, whereas non-dusty quasars preferentially show peaked spectra consistent with compact, self-absorbed jets. The steep spectra in dusty quasars are attributed to shocks generated when outflowing winds or low-power jets interact with the dusty interstellar medium, and the authors argue this supports the blow-out evolutionary scenario in which the shocks heat and destroy the surrounding dust.","pith_inferences":["The same SED-fitting method applied to a larger, dust-selected quasar sample (e.g., from DESI) could reveal whether the fraction of steep-spectrum sources increases continuously with E(B-V), which would strengthen the case for a causal connection.","If the shock origin is correct, the steep dusty quasars should show stronger ionized or molecular outflows and hotter dust (mid-IR excess) than the peaked non-dusty quasars; near-infrared spectroscopy could test this.","The upturned sources, after removing the LoTSS data, appear as peaked cores on top of steep relic emission; this is consistent with restarted radio activity, and future high-frequency observations could confirm whether the relic lobes are truly detached.","The four-times-greater fraction of steep spectra among dusty quasars is based on only 38 objects; a dedicated survey with a sample of a few hundred would determine whether the effect is as large as reported or diluted."],"forward_implications":["The radio spectral shape can be used as a fast diagnostic of whether a quasar is in a dusty blow-out phase or hosts a compact jet.","The excess of steep spectra in dusty quasars implies that the enhanced radio emission in red quasars is produced by shocks, not by star formation or orientation.","The higher fraction of peaked spectra in non-dusty quasars supports the idea that these systems host young or frustrated jets.","The wind-shock model calculations show that accretion-disc winds alone cannot power the radio emission in the most radio-loud members, suggesting low-power jets as the outflow driver in those cases."],"supporting_citations":[{"why":"Supplies the e-MERLIN 0.2-arcsecond imaging, sample selection, and radio sizes that anchor the SED comparison","marker":"Rosario et al. 2021"},{"why":"Provides the blue QSO composite template and method for measuring E(B-V) from SDSS spectra","marker":"Fawcett et al. 2022"},{"why":"Gives the wind-shock model used to predict radio luminosity and to test whether winds can explain the observed emission","marker":"Nims et al. 2015"},{"why":"Establishes the GPS/CSS turnover-frequency versus size relation used to compare peaked sources to known jetted objects","marker":"O'Dea & Baum 1997"},{"why":"Provides the curved (peaked) SED model that is fit to the data","marker":"Snellen et al. 1998"},{"why":"Defines the radio-loudness parameter and the R=-3.7 division used to split the sample","marker":"Klindt et al. 2019"},{"why":"Earlier work establishing the dust-extinction versus radio-detection correlation that motivates the present study","marker":"Fawcett et al. 2023"},{"why":"Supplies fitting methodology for peaked-spectrum sources","marker":"Callingham et al. 2017"}],"fun_headline_variants":["Shocks in dusty quasars steepen radio spectra","Steep radio slopes reveal shock-powered outflows in dusty QSOs","Dusty QSO radio steepness points to shock-driven outflows","Outflow shocks link dusty quasars to steeper radio emission"],"cache_read_input_tokens":45952,"weakest_assumption_plain":"The classification of four 'upturned' quasars as two-component sources relies on the assumption that their low-frequency LoTSS emission is a separate ancient relic and not part of the same compact synchrotron spectrum.","fun_headline_variants_meta":{"raw":{"variants":["Shocks in dusty quasars steepen radio spectra","Steep radio slopes reveal shock-powered outflows in dusty QSOs","Dusty QSO radio steepness points to shock-driven outflows","Outflow shocks link dusty quasars to steeper radio emission"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000616,"raw_usage":{"total_tokens":2901,"prompt_tokens":1025,"completion_tokens":1876,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":641,"completion_tokens_details":{"reasoning_tokens":1803}},"tokens_in":641,"tokens_out":1876,"duration_ms":14284,"temperature":1.0,"reasoning_tokens":1803,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T19:10:17.438452+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the low-frequency flux of the four upturned quasars with a radio telescope that resolves the extended lobes, such as the Very Long Baseline Array; if the compact core shows a spectral turnover at LoTSS frequencies, the two-component interpretation is wrong and the steep-spectrum fractions would change.","supporting_citations":[],"review_version":1}