Investigate the institution
Compare the offer of a housing voucher with its use and examine how age and follow-up change the question. Is the effect of an offer the same as the effect of moving?
Neighborhood conditions link institutions, transport, safety, schools and social networks. A change in address can alter several of these together.
A randomized offer identifies an offer comparison under appropriate design and follow-up conditions. Actual use is a subsequent choice and cannot simply inherit randomization.
Age at exposure, duration and follow-up outcome define different questions. A result for one group at one time does not settle every neighborhood-effect claim.
Distinguish authentic evidence from the practice ledger
Moving to Opportunity program design
Families in high-poverty public housing in five US cities were offered housing vouchers through a lottery. The experiment used an offer to study changes in neighborhood circumstances and later outcomes.
Research-project summary of the assignment design. Selection into the participating population and use of an offered voucher are distinct from lottery assignment.
NBER Moving to Opportunity research project — Historical experiment, launched in the 1990s. This is an attributed summary, not reproduced article text.
Age and long-term outcomes
Using linked tax data, this analysis finds improved later earnings and college attendance for children who were younger than 13 when their families moved using the experimental voucher. The reported $3,477 annual-income gain is about 31% of a $11,270 control mean for the stated young-child comparison.
The reported gain is an estimated effect of voucher use, not the raw effect of offering a voucher to every eligible family. Preserve subgroup, outcome, follow-up age and identification assumptions; do not generalize it to all ages or every mover.
Raj Chetty, Nathaniel Hendren and Lawrence F. Katz — NBER working paper 21156, revised September 2015. This is an attributed summary, not reproduced article text.
Invented encouragement trial
A fictional lottery assigns 200 families to an offer and 200 to no offer. In the offer group 100 use the program; nobody in the no-offer group can access it. The average measured outcome is 54 in the offer group and 50 in the no-offer group.
These are invented outcome points, not dollars or MTO estimates. The four-point offer contrast and 50% participation difference allow an assumption-dependent instrumental-variable illustration.
Inspect the fictional ledger
Fictional practice data, not the study results. Scroll horizontally on narrow screens.| Fictional lottery group | Families assigned | Program users | Mean outcome points |
|---|
| Offer | 200 | 100 | 54 |
|---|
| No offer | 200 | 0 | 50 |
|---|
Keep lottery assignment → program use → neighborhood exposure → measured outcome as separate columns. Add a possible arrow from an offer directly to an outcome. Explain how such a direct channel would affect the assumption used to interpret the participation estimate.
Work the comparison
In the invented ledger, the offer contrast is 54−50=4 outcome points. The participation difference is 100/200−0/200=0.5. Dividing 4/0.5 gives 8 points only under additional assumptions: the offer affects outcomes through participation, no one participates less because of being offered, assignment is valid, and interference is absent. It targets people whose participation changes with the offer, not everyone.
Make an inspectable artifact
Build separate rows for assignment, participation, neighborhood exposure and outcome. Calculate the fictional offer contrast and explain the assumptions needed to translate it into a participant effect. Audit the actual income comparison’s denominator and subgroup.
Record your diagram or table as a written trace in the claim field. Number each link, identify the relevant source ID and distinguish measured evidence from a proposed mechanism. Save a capture before revising your interpretation.
Test a stronger claim
Write a housing-evidence brief separating offer, take-up, exposure and outcome. Cite the real study accurately, show the fictional calculation, identify assumptions and propose what evidence would help distinguish neighborhood mechanisms.
Independent extension
Explain the exclusion restriction, monotonicity and interference assumptions in words. Compare the offer estimand with the local effect for people induced to participate, then discuss age-defined subgroups and the limits of attributing the result to any one neighborhood feature.
Use the supplied records and public research descriptions. Do not conduct deceptive hiring or housing applications, collect classmates’ sensitive information, or treat the fictional ledger as a real effect estimate.
