[ { "trial": "T24", "eligible": true, "reason": null, "why": "Low-income Compton households were randomly assigned to two years of flat unconditional transfers (paid twice monthly or quarterly) or a no-transfer control, and the endline reports a psychological well-being index that includes the K6, plus labor market participation, hours, and earnings.", "country": "United States", "sites": [ "Compton, California (Los Angeles County)" ], "population": "Low-income households in Compton with a member aged 23 to 57 and income below 220% of the federal poverty line, excluding SSI/SSDI recipients", "years": "2021-02 to 2023-04", "arms": [ { "label": "Twice-monthly transfers", "kind": "recurring", "amount": 492, "currency": "USD", "per": "month", "months_paid": 24, "n_randomized": 345 }, { "label": "Quarterly transfers", "kind": "recurring", "amount": 492, "currency": "USD", "per": "month", "months_paid": 24, "n_randomized": 350 }, { "label": "Control", "kind": "control", "amount": 0, "currency": "USD", "per": "month", "months_paid": 0, "n_randomized": 1402 } ], "comparisons": [ { "id": "T24-1", "treated": "Twice-monthly transfers", "control": "Control", "kind": "recurring" }, { "id": "T24-2", "treated": "Quarterly transfers", "control": "Control", "kind": "recurring" } ], "randomization": "individual", "stratified_by": "participant sex", "payment_start": "2021-02", "payment_months": 24, "waves": [ { "label": "Endline survey (18 months after the start of transfers, May-September 2022)", "months_since_first_payment": 18, "during_payments": true, "measures": [ "mental_health", "employment", "hours", "earnings" ] }, { "label": "Follow-up survey (Summer 2023)", "months_since_first_payment": 28, "during_payments": false, "measures": [ "hours", "earnings" ] } ], "mental_health_instruments": [ "Psychological well-being index: first principal component of depression frequency (1-5), stress frequency (1-5), life satisfaction (1-10), happiness (1-4), and the Kessler 6 (K6); only the index is reported" ], "employment_measures": [ "participated in labor market (any hours worked)", "weekly hours worked in formal, informal, and self-employment jobs", "respondent earnings before taxes, last 30 days", "household income excluding transfers, last 30 days", "follow-up survey: self-reported increase or decrease since March 2021 in weekly work hours and weekly earnings (Likert items)" ], "evaluator": "Balakrishnan (Jain Family Institute), Chan and Morduch (NYU Wagner), Constantino (Stanford), and Haushofer (Cornell); survey data collected by the USC Center for Economic and Social Research (CESR)", "result_reports": [ "registered:AEARCTR-0007621_2", "aei:compton", "registered:AEARCTR-0007621_1" ], "notes": "Transfer size depended on the number of children ($300, $450, or $600 a month); $492 is the average among the 625 who took up; averaged over all 695 offered (70 never took up) it was $442 ($448 twice-monthly, $436 quarterly). Both arms had the same total value: the twice-monthly arm was paid on the 7th and 21st of each month and the quarterly arm a lump sum at the start of each quarter, so amounts are coded as monthly equivalents. Paid by prepaid debit card, direct deposit, PayPal, or Venmo; benefits protected by waivers and a hold-harmless fund. Everyone was paid $50 per completed survey, with extra follow-up and timeliness bonuses aimed especially at controls. Endline response was 51% (1,074: 374 treated, 700 control). Randomization ran in enrollment waves with different treatment shares, and the analysis is weighted; 253 participants were erroneously re-randomized in the second wave; 102 households paid from the program's start in December 2020-January 2021 are excluded; 38 households had more than one recipient. The Summer 2023 follow-up (942 respondents) asked retrospectively about changes since March 2021 (Table A14); its months are computed from dates. The May 2025 paper supersedes the November 2024 NBER version (aei:compton and registered:AEARCTR-0007621_1 are the same file).", "third_reader": { "decided": [ "comparisons" ], "why": "The paper randomizes 695 households between twice-monthly (345) and quarterly (350) transfers of equal total value and 1,402 controls who got no transfers, and estimates each cash arm against control. It states the average monthly transfer among the 625 households that took it up as $492 ($491 when derived from the yearly $5,892; $448 and $436 by arm counting non-takers as zero), and the same schedule by number of children applied to both arms, so both are coded at $492." } }, { "trial": "T86", "eligible": true, "why": "Two Chicago lotteries (the City Department of Housing and The Resurrection Project) gave flat one-time unconditional $1,000 grants paid directly to selected applicants against unselected applicants, and follow-up surveys report a PHQ-4 anxiety item for each lottery.", "country": "United States", "sites": [ "Chicago, Illinois" ], "population": "Chicago residents applying for COVID-19 emergency grants: DOH (renters and homeowners at risk of eviction, foreclosure or homelessness, pre-pandemic income at most 60% of area median, pandemic income loss) and TRP (income below 300% of the poverty line, not receiving CARES Act aid; undocumented and mixed-status families encouraged)", "years": "2020-04 to 2020-10", "arms": [ { "label": "Chicago DOH: selected by lottery", "kind": "lump_sum", "amount": 1000, "currency": "USD", "per": "once", "months_paid": 0, "n_randomized": 1537 }, { "label": "Chicago DOH: not selected", "kind": "control", "amount": 0, "currency": "USD", "per": "once", "months_paid": 0, "n_randomized": 73126 }, { "label": "Chicago TRP: selected by lottery", "kind": "lump_sum", "amount": 1000, "currency": "USD", "per": "once", "months_paid": 0, "n_randomized": 1765 }, { "label": "Chicago TRP: not selected", "kind": "control", "amount": 0, "currency": "USD", "per": "once", "months_paid": 0, "n_randomized": 4688 } ], "comparisons": [ { "id": "T86-1", "treated": "Chicago DOH: selected by lottery", "control": "Chicago DOH: not selected", "kind": "lump_sum" }, { "id": "T86-2", "treated": "Chicago TRP: selected by lottery", "control": "Chicago TRP: not selected", "kind": "lump_sum" } ], "randomization": "individual", "stratified_by": null, "payment_start": "2020-04", "payment_months": null, "waves": [ { "label": "Chicago DOH survey (May-June 2020)", "months_since_first_payment": 1, "during_payments": true, "measures": [ "mental_health" ] }, { "label": "Chicago TRP survey (February-April 2021)", "months_since_first_payment": 7, "during_payments": false, "measures": [ "mental_health" ] } ], "mental_health_instruments": [ "PHQ-4 item 'How often have you felt nervous, anxious, or on edge?' over the last week, coded 1 for more than half the days or nearly every day" ], "employment_measures": [], "evaluator": "University of Chicago Inclusive Economy Lab with John Eric Humphries and Winnie van Dijk; reported in the four-city study of Collinson, DeFusco, Humphries, Keys, Phillips, Reina, Turner and van Dijk (NBER Working Paper 32463)", "result_reports": [ "registered:AEARCTR-0007731_1" ], "notes": "Two separate lotteries, each with its own control group. DOH (first round; applications March 27 to April 5, 2020; about 75,700 applications): random ranks, offers in rank order until funds ran out, eligibility checked after selection; 722 grants paid directly to tenants in April-May 2020 (later DOH rounds had no usable control group). TRP Chicago Resiliency Fund (second round, applications from June 22, 2020; 10,300 applications): 1,607 grants paid directly, August-October 2020. n_randomized are the paper's balance-table counts of selected and unselected applicants (maximum counts; the registry gives 1,685 DOH and 1,761 TRP selected). Estimates are TSLS of grant receipt instrumented by selection, with survey weights for unequal response (DOH 41% selected vs 28% not; TRP 31% vs 18%). DOH survey: the month after payment, while some grants were still being paid. TRP survey months computed from the first TRP payments (August 2020). Both programs took renters and homeowners (75-80% renters). Eviction moratoria were in force. Administrative outcomes (Experian credit panel, Infutor address histories, Cook County eviction records, Chicago HMIS homelessness records) have no mental health or employment measures and are not coded as waves. Employment appears only as a pre-lottery subgroup and baseline 'any paid work'. Only the Chicago columns of the four-city paper belong to this trial.", "third_reader": { "decided": [ "waves" ], "why": "Of the coded measures the paper presents only the PHQ-4 'feeling anxious' item, from the Chicago DOH survey (May-June 2020, which it says came the month after the April-May grants) and the TRP survey (February-April 2021, 6 to 8 months after TRP's first grants in August 2020, so 7 at its midpoint). The credit, address, eviction, and homelessness records hold none of the coded measures, so they are not waves." } }, { "trial": "T92", "eligible": true, "why": "Matched-tuple randomized trial of a flat unconditional $200 a month for 24 months against a $10-a-month control in Anderson, South Carolina, reporting quarterly employment and labor earnings from state records.", "country": "United States", "sites": [ "Anderson, South Carolina (Anderson County)" ], "population": "Low-income adults (household income at most $60,000) served by Anderson Interfaith Ministries, mostly living in non-urban areas; SSI recipients screened out", "years": "2023-08 to 2025-07", "arms": [ { "label": "Treatment group", "kind": "recurring", "amount": 200, "currency": "USD", "per": "month", "months_paid": 24, "n_randomized": 40 }, { "label": "Control group", "kind": "control", "amount": 10, "currency": "USD", "per": "month", "months_paid": 24, "n_randomized": 53 } ], "comparisons": [ { "id": "T92-1", "treated": "Treatment group", "control": "Control group", "kind": "recurring" } ], "randomization": "individual", "stratified_by": [ "matched four-tuples (Bai 2022) formed on baseline demographics (sex, race, Hispanic, high school degree, home ownership, age, household size) and baseline outcomes (labor force participation, income, poor or fair health, smoking, poverty)" ], "payment_start": "2023-08", "payment_months": 24, "waves": [ { "label": "Quarter 1 after first deposit (2023 Q4), state records", "months_since_first_payment": 3, "during_payments": true, "measures": [ "employment", "earnings" ] }, { "label": "Quarter 2 (2024 Q1), shorter term, state records", "months_since_first_payment": 6, "during_payments": true, "measures": [ "employment", "earnings" ] }, { "label": "Quarter 3 (2024 Q2), state records", "months_since_first_payment": 9, "during_payments": true, "measures": [ "employment", "earnings" ] }, { "label": "Quarter 4 (2024 Q3), state records", "months_since_first_payment": 12, "during_payments": true, "measures": [ "employment", "earnings" ] }, { "label": "Phone follow-up (October 2024, about 2024 Q3)", "months_since_first_payment": 12, "during_payments": true, "measures": [ "employment" ] }, { "label": "Quarter 5 (2024 Q4), longer term, state records", "months_since_first_payment": 15, "during_payments": true, "measures": [ "employment", "earnings" ] } ], "mental_health_instruments": [], "employment_measures": [ "employed in the quarter (any taxed wages or tips in South Carolina Department of Employment and Workforce records)", "monthly earnings from labor income (state records)", "employment and income from all sources, taxed and untaxed (phone interview about 2024 Q3)", "labor force participation (baseline survey)" ], "evaluator": "Jorge Luis García (Texas A&M University), Patrick L. Warren and L. Reed Watson (Clemson University, John E. Walker Department of Economics), with Anderson Interfaith Ministries and Spero Financial", "result_reports": [ "registered:AEARCTR-0014466_1" ], "notes": "Both arms got free Spero credit-union accounts with debit cards, into which deposits were made monthly from August 25, 2023 ($200 treatment, $10 control), so the control's $10 is a token and account access is common to both arms. Existing benefits were not protected (no discernible effect on SNAP or TANF). 100 enrolled at AIM in May-June 2023; 7 who did not activate their debit card were excluded before randomization; 93 randomized within 23 matched tuples (40 treatment, 53 control). Quarter 0 (2023 Q3) is used as baseline although it includes about one month of payments. Results so far run to five quarters after the first deposit (bank data to 19 months); payments continue to July 2025 and later quarters are promised. Employment means any taxed wages or tips in the quarter (unemployment and non-participation not distinguished); results are plotted for every quarter, with tests reported at quarters 2 and 5 and p-values for each quarter's employment difference. Phone interviews in October 2024 reached 42 (40 with data) about 2024 Q3 employment and all-source income, as a check on the records. Hospital visits, medical spending, SNAP/TANF, criminal records and bank data are not coded as waves. No mental health outcome is reported. An earlier registration (AEARCTR-0008509, 2021) planned a 40/60 design with a $10 monthly honorarium for controls and mental health questions in the baseline survey.", "third_reader": { "decided": [ "waves" ], "why": "Employment and labor earnings from state records are plotted for each of quarters 1 to 5 after the first deposit of August 25, 2023, which the paper counts as 3 to 15 months (two quarters is six months, five is fifteen), all during payments. The October 2024 phone interviews asked about employment in Quarter 3 of 2024 and the paper calls them a follow-up four quarters after baseline, so 12 months." } }, { "trial": "T111", "eligible": true, "why": "Randomized trial among people experiencing homelessness of $750 a month for 12 months plus a phone-buddy program against the phone-buddy program alone (and a waitlist), with employment status reported at follow-up.", "country": "United States", "sites": [ "San Francisco Bay Area (San Francisco, Oakland)", "Los Angeles County" ], "population": "English-speaking adults experiencing homelessness (HUD definition) who expressed interest in the Miracle Friends phone-buddy program", "years": "2022-05 to 2024-10", "arms": [ { "label": "Miracle Money", "kind": "recurring", "amount": 750, "currency": "USD", "per": "month", "months_paid": 12, "n_randomized": 256 }, { "label": "Miracle Friends", "kind": "control", "amount": 0, "currency": "USD", "per": "month", "months_paid": 0, "n_randomized": 267 }, { "label": "Waitlist", "kind": "control", "amount": 0, "currency": "USD", "per": "month", "months_paid": 0, "n_randomized": 237 } ], "comparisons": [ { "id": "T111-1", "treated": "Miracle Money", "control": "Miracle Friends", "kind": "recurring" } ], "randomization": "individual", "stratified_by": null, "payment_start": null, "payment_months": 12, "waves": [ { "label": "3-month survey", "months_since_first_payment": 3, "during_payments": true, "measures": [] }, { "label": "6-month survey", "months_since_first_payment": 6, "during_payments": true, "measures": [] }, { "label": "9-month survey", "months_since_first_payment": 9, "during_payments": true, "measures": [] }, { "label": "12-month survey", "months_since_first_payment": 12, "during_payments": true, "measures": [] }, { "label": "15-month survey (final follow-up)", "months_since_first_payment": 15, "during_payments": false, "measures": [ "employment" ] } ], "mental_health_instruments": [ "PROMIS Scale 1.2 global health item (self-rated mental health, excellent to poor)" ], "employment_measures": [ "employment status (employed full-time, 35+ hours a week; part-time; unemployed looking or not looking for work; retired), from Los Angeles Homeless Count questions" ], "evaluator": "University of Southern California, Suzanne Dworak-Peck School of Social Work, Center for Homelessness, Housing and Health Equity Research (Benjamin Henwood), with Miracle Messages", "result_reports": [ "registered:NCT05408884_1", "registered:NCT05408884_3" ], "notes": "People were recruited for the Miracle Friends phone-buddy program without being told about cash; those assigned to Miracle Money were told only after they began engaging with their phone buddy, so 103 of the 256 assigned received payments (about $1 million in 2022-2024; no ongoing conditions are described once payments began). Two-stage randomization: 523 to Miracle Friends vs 237 to the waitlist, then 256 of the 523 to Miracle Money as well (registry enrollment 759). The reports' main contrast, Miracle Money vs waitlist, confounds cash with the phone-buddy program, so the comparison coded is Miracle Money vs Miracle Friends, whose arms differ only by the cash; results are shown for all three groups. Reported analyses are not ITT: the final report compares participants who completed the study and received an intervention (Miracle Money n=88, Miracle Friends n=40, waitlist n=107), with chart numbers garbled by OCR, and the SSWR 2026 abstract reports 25% employed at follow-up with no differences between groups. The follow-up at which employment is reported is not named; it is coded as the last (15-month) survey. The November 2023 interim report (registered:NCT05408884_2) compares 69 recipients with at least 6 payments to 86 waitlisted at 6 months on housing and unmet needs only. Survey months count from baseline; payment start dates are not reported beyond recruitment beginning in May 2022. $30 per survey for all. Self-rated mental health was measured but no result is reported; loneliness (UCLA 3-item) and substance use are reported. Qualitative interviews with recipients near notification and at the end of the year are not coded as waves. Payment dates are not reported, so years run from the start of recruitment (May 2022) to the planned end of data collection (October 2024).", "third_reader": { "decided": [ "comparisons", "years" ], "why": "The Miracle Friends arm got the same phone-buddy program as Miracle Money without the $750 a month, so the cash arm is compared with it; the waitlist also went without the phone buddy, so Miracle Money against the waitlist, the contrast the final report emphasizes, differs by more than cash and is not a comparison. Years are put in the codebook's format, bounded by recruitment's start and the planned end of data collection, since payment dates are not reported." } } ]