The Effects of Exposure to Better Neighborhoods on Children: New Evidence from the Moving to Opportunity Experiment

• p. 19 Section III.B

  mean non-W-2 earnings for tax filers in the control group – which is a plausible upper bound for non-filers – is only $1,721

• p. 22 Footnote 25

  mean spousal income of $802 for tax filers in the control group

• p. 76 Appendix Table 3c, Control Group Mean

  1721.7

Only the authors can say what is right here.

Judge's note: Appendix Table 3c shows control means of 1721.7 and 802.1 for all 8,420 Panel A observations. Its notes and the Section II definitions say non-W-2 earnings is 0 for non-filers and spouse income is 0 for singles. The quoted figures are therefore full-sample means labelled as filer means. The authors need to say which statistic was meant.

Second check: another model agreed

• p. 24 Section III.C opening

  Appendix Tables with the same number provide ITT estimates with the full set of controls and TOT estimates

• p. 89 Appendix Table 10a notes

  This table replicates Table 12 in the text

Judge's note: I checked the Appendix Table 10a image and its notes. The 10b notes also say they replicate Table 12. Appendix Tables 3a/3b, 4a/4b, 8a/8b and 9a/9b follow the stated rule; 10a/10b do not. The main text never points readers to Appendix Table 10 for the Table 12 versions. This is a numbering inconsistency only.

Second check: another model agreed

• p. 72 Appendix Table 1c, Notes

  Panel A lists neighborhoods based on the most common housing projects

Judge's note: The p.72 image shows no Panel A/B/C labels. The note's descriptions match the three sections in order, so the mapping can be recovered.

Second check: another model agreed

• p. 23 Section III.B Summary

  with a TOT impact on individual earnings of approximately 35%

• p. 20 Section III.B

  This is a 31% increase relative to the control group mean earnings of $11,270 and a 34% increase relative to the “control complier mean”

• p. 23 Section III.B Summary

  earnings of approximately 35%

• p. 2 Abstract

  $3,477 (31%) higher on average relative to a mean of $11,270

Judge's note: I checked the p.20 and p.23 crops. 35% matches neither the 31% control-mean figure nor the 34% CCM figure, and the summary does not say which base it uses. It may be a loose rounding of 34%. This is an inconsistency in how the headline effect is summarised, not a data error.

Second check: another model agreed

• p. 7 Introduction

  we find robust, statistically significant evidence that the experimental voucher improved long-term outcomes for (young) boys

• p. 28 Section III.F

  an estimate that is significantly different from 0 with p = 0.085

• p. 61 Table 7, Panel A, Col 1

  1679.2⁺

• p. 28 Section III.F

  significantly different from 0 with p = 0.085

Only the authors can say what is right here.

Judge's note: Table 7 Panel A col 1 shows 1679.2⁺ (973.3), 522.3⁺ (290.1) and ZIP poverty -1.498*. Elsewhere (p.19) the paper calls p=0.101 'marginally significant', and the intro uses p<0.05 as its example threshold. Calling p=0.085 'significantly different from 0' is inconsistent with that usage. Worth a query on the wording; the numbers themselves are not in dispute.

Second check: another model agreed

• p. 29 Section III.F

  in every racial group (Black, Hispanic, and White)

• p. 83 Appendix Table 7a, column header

  Non-Black Non-Hisp.

• p. 65 Table 11 notes

  racial groups (Black, Hispanic, and Other)

• p. 7 Introduction, para 2

  for Whites, Blacks, and Hispanics

Judge's note: I checked the p.29 and p.7 text crops, the Table 11 notes crop ("racial groups (Black, Hispanic, and Other)") and Appendix Table 7a on p.83. The table's columns are Hispanic, Non-Black Non-Hisp. and Black Non-Hisp. The text's 'White' group is therefore the residual non-Black, non-Hispanic group, and the paper never defines it as White. This is a minor labelling inconsistency.

Second check: another model agreed

• p. 12 Section II.B, Marital Status and Fertility

  having a teenage birth if she had a child between the ages of 13 and 19

• p. 59 Table 5 notes

  In Column 5, it is an indicator for having a child before the age of 19.

Only the authors can say what is right here.

Judge's note: Both images confirm the wording. The two definitions differ on births at age 19; only the authors can say which was used.

Second check: another model agreed

• p. 20 Section III.B

  the Section 8 voucher has smaller positive, marginally significant effects on younger children

• p. 74 Appendix Table 3a, Panel A, Sec. 8 row

  908.6

Only the authors can say what is right here.

Judge's note: The Appendix Table 3a image shows no ⁺ on any Section 8 row entry (492.4, 908.6, 1426.3, 219.4, 1075.9, 0.736, 1140.1, 638.2). Without controls, Table 3 col 2 gives 1109.3 (p=0.101, which the paper calls marginal). The sentence may be meant to cover both specifications, but as written it does not fit the controlled estimates.

Second check: another model agreed

• p. 22 Section III.B, Earnings Trajectories

  the treatment effects fall significantly with the age at which income is measured for the older children

• p. 49 Figure 1 notes

  p = 0.06 for the age 13-18 series

Judge's note: I checked both crops. The test is the age-interaction test described in footnote 27. For the older series p = 0.06, which is not below 0.05. Elsewhere the paper calls p = 0.101 "marginally significant" (p.19). Calling a p = 0.06 result "significantly" therefore overstates it unless a 10% level is meant, and the text does not say so.

Second check: another model agreed

• p. 59 Table 5 header Col 6 and notes

  Father on Birth Cert.

• p. 26 Section III.D

  indicator for having a father listed on the first- born child’s SSN application

Judge's note: Table 5 header and notes checked on the p.59 image; the data definition is in the main.txt lines 426-435 (Kidlink SSN application data). The label describes the measure as a birth-certificate field, but the source is the SSN application. The paper treats this as a proxy for the father being present at birth, so either wording could be defended. Recorded as a labelling inconsistency.

Second check: another model agreed

• p. 28 III.F

  we find adverse long-term treatment effects for both boys and girls who were above age 13 at RA (Panel B)

• p. 61 Table 7 Panel B, ZIP Poverty, Female

  -0.627

Judge's note: In the Table 7 image, the Panel B earnings and college quality estimates are negative for both genders. ZIP poverty is adverse in sign for boys (0.281, 0.844) but favourable in sign for girls (-0.627, -1.241); all are insignificant. The preceding sentence covers neighbourhood quality, so the blanket "adverse" statement overstates things for that outcome among girls. This is a minor wording point.

Second check: another model agreed

• p. 30 Equation (3)

  controlling for randomization site indicators (si)

• p. 62 Table 8 notes

  and interactions between randomization site indicators and age at RA

Only the authors can say what is right here.

Judge's note: On p.28 (PDF p.30), Equation (3) contains only s_i γ. The Table 8 notes (p.62 crop) describe site × age-at-RA interactions and do not mention site main effects. The specification behind Table 8 is ambiguous; only the authors can say which one is right.

Second check: another model agreed

• p. 32 Section III.G

  Panel B shows significant positive effects on college quality for children who move at young ages

• p. 50 Figure 2, Panel B

  B. College Quality, Ages 18-20 ($)

Only the authors can say what is right here.

Judge's note: In Figure 2B (p.50), the gridlines are at +1000 and -1000, so zero is midway between them. Point estimates at ages 4–10 are about +700 to +800. Read off the image, the lower dashed CI line sits below the zero level at ages 4, 8 and 10 and roughly at zero at age 6. This is a visual reading only. The text may be referring to the pooled under-13 estimate, but the sentence attributes significance to the panel, so this is a query to the authors.

Second check: another model agreed

• p. 55 Table 1 notes

  Columns 4- 6 include those above age 13 at random assignment

• p. 55 Table 1 column header

  Age 13-18 at Random Assignment

Judge's note: On p.55, the header gives 'Age 13-18' and the columns 1–3 header gives '< Age 13'. 'Above age 13' would exclude 13-year-olds, which conflicts with the header.

Second check: another model agreed

• p. 56 Table 2 notes

  all children born before 1991 in the MTO data

• p. 56 Table 2 notes

  includes all children born before 1991 in the MTO data

• p. 55 Table 1 notes

  all children in the MTO data who were born in or before 1991

Only the authors can say what is right here.

Judge's note: I compared the Table 1 notes (p.55) with the Table 2 notes (p.56). Table 2 calls its sample 'nearly identical' to the linked sample, so the different cutoff wording is probably a slip, but only the authors can confirm it.

Second check: another model agreed

• p. 1 Cover disclaimer

  National Bureau of Economic Research..

Judge's note: The image on p.1 shows two full stops after 'Research'.

Second check: none (judge only)

• p. 2 Title block

  May 2015, Revised Septemberr 2015

• p. 2 Abstract page header

  Revised Septemberr 2015

• p. 6 Introduction, end of paragraph

  yields produces larger benefits for younger children

• p. 13 III.C

  linked MTO-Tax data data

Judge's note: The p.2 image shows 'Septemberr'. The p.29 image shows 'below age at 13'. The footnote 41 image shows 'we focus children'. 'yields produces' and 'data data' are judged under their own clusters, C005 and C009.

Second check: none (judge only)

• p. 6 Introduction

  lower-poverty areas yields produces larger benefits for younger children

• p. 6 Introduction

  lower-poverty areas yields produces larger benefits

• p. 6 Introduction, paragraph 'Putting the results together'

  yields produces larger benefits for younger

• p. 6

  yields produces larger benefits for younger children

Judge's note: Confirmed on the p.6 image.

Second check: none (judge only)

• p. 38 Footnote 43

  0.308 × 0.56 × $1.74m = $302, 000.

• p. 38 Footnote 43

  0.308 × 0.56 × $1.74m = $302, 000

• p. 7 Introduction, last para

  total lifetime earnings by about $302, 000

Judge's note: The image shows the footnote as quoted. The headline $302,000 on p.5 probably comes from an unrounded control-earnings share (about 56.4%). The same share gives the $99,000 PDV: with 0.56 it would be about $98,300. The gap is a rounding presentation issue.

Second check: none (judge only)

• p. 9 Section I

  indefinitely as long their income was below 50 percent

Judge's note: Confirmed on the p.9 image.

Second check: none (judge only)

• p. 13 Section II.C

  balanced in the linked MTO-Tax data data

• p. 13 Section II.C

  linked MTO-Tax data data

• p. 13

  linked MTO-Tax data data.

• p. 13

  alanced in the linked MTO-Tax data data. 13 of the 196 diffe

Judge's note: The p.13 image shows both words on the same line.

Second check: code check

• p. 15 Section III.A, first paragraph

  include no controls other that randomization site indicators

Judge's note: Confirmed on the p.15 image.

Second check: none (judge only)

• p. 21 Section III.B, employment

  appear to be driven primary by higher wage rates

Judge's note: The wrong word form is visible in the page crop.

Second check: none (judge only)

• p. 25 Footnote 29

  larger than the $686 impact on projected earnings at age 31

• p. 58 Table 4 Panel A Col 6

  686.7**

• p. 25 Footnote 29

  larger than the $686 impact on projected earnings

• p. 25 Section III.C text

  by $687

Judge's note: I read the p.25 page image and Table 4 Panel A Col 6 (686.7). The value 686.7 rounds to 687. The footnote is inconsistent with the rounding used elsewhere on the same page.

Second check: none (judge only)

• p. 76 Appendix Table 3c notes

  Column 9 and 10 replicate the specifications

• p. 25 Section III.C

  shown in Column 7-10 of Table 4

Judge's note: Both instances confirmed on the p.76 page image and the p.25 crop. Purely cosmetic.

Second check: none (judge only)

• p. 26 Section III.D

  which we infer from applications from social security numbers for children

Judge's note: The wording is visible in the page image. Section II.B uses "applications for social security numbers".

Second check: none (judge only)

• p. 29 Section III.F

  for children below age at 13 at RA

• p. 29 Section III.F

  for children below age at 13

• p. 29 Section III.F, heterogeneity by race

  children below age at 13 at RA

Judge's note: The p.29 crop shows "below age at 13 / at RA" as quoted.

Second check: none (judge only)

• p. 33 Section IV.A

  is -$354 (2.4% of the control group mean)

• p. 63 Table 9, Col 1

  14381.0

• p. 63 Table 9, Column 1

  -354.1

Judge's note: Table 9, column 1 (p.63) shows -354.1 and a control mean of 14381.0. The TOT share in the same paragraph (734/14381 = 5.1%) is rounded correctly, so 2.4% is a rounding slip.

Second check: none (judge only)

• p. 35 Footnote 41

  because we focus children who moved at young ages

Judge's note: The footnote 41 crop shows the text exactly as quoted.

Second check: none (judge only)

• p. 44 References

  Chetty, Raj and Nathan Hendren. 2015.

Judge's note: The reference crop shows 'Nathan Hendren'. The title page and the 'Hendren, Nathaniel. 2013' reference spell the name Nathaniel.

Second check: none (judge only)

• p. 51 Figure 3 notes

  See notes to Tables 9 for the definition of individual earnings

• p. 51 Figure 3 notes, last sentence

  See notes to Tables 9 for the definition of individual earnings.

Judge's note: Checked in the Figure 3 notes on p.51.

Second check: none (judge only)

• p. 56 Table 2 notes

  nearly identical our linked analysis sample

• p. 56

  This sample is nearly identical our linked analysis sample

Judge's note: Checked in the Table 2 notes on p.56.

Second check: none (judge only)

• p. 56 Table 2 notes

  at random assignment The estimates in Panels A and

Judge's note: Checked in the Table 2 notes on p.56: there is no period before 'The'.

Second check: none (judge only)

• p. 60 Table 6 notes

  The dependent variable in Columns 1 is the poverty share

Judge's note: Confirmed on the p.60 page image.

Second check: none (judge only)

• p. 66 Table 12 notes

  The dependent variable in columns 1 is an indicator for filing a 1040 tax return

Judge's note: Confirmed on the p.66 page image.

Second check: none (judge only)

• p. 68 Appendix Table 1a row label

  Disabed Person in Household (%)

Judge's note: Confirmed on the p.68 page image.

Second check: none (judge only)