What the evidence says
The direct experimental evidence on eligibility-outreach nudges for the Earned Income Tax Credit (EITC) comes from Linos (2020), a large-scale randomized trial (N≈1 million) that varied the content, design, messenger, and delivery mode of outreach messages. Across all of these variations, the study found no measured effect on households' likelihood of filing a tax return or claiming the credit. A companion factorial design (Studies 2 and 4) tested individual message features in isolation — formality, messenger credibility, information about the credit amount, and information about VITA filing sites — and found no measurable effect from any single feature either. This is not a case of underpowered null results: the sample sizes and point-estimate precision let researchers rule out effects as small as 0.5 to 1.5 percentage points, depending on the specific trial.
That null result sits alongside a separate body of associational evidence, cited within the same source, describing positive effects on take-up from earlier interventions — Bhargava and Manoli (2015), Bettinger et al. (2012), and Finkelstein and Notowidigdo (2019). These studies used varied designs and are reported here only as prior findings rather than as independently sourced primary claims.
A third strand, Guyton et al. (2017), measured a positive effect of outreach specifically on tax filing rates, with an effect size of about 0.5 percentage points — described as positive but much smaller than the effect found by Bhargava and Manoli (2015).
Taken together, the claims here describe two different kinds of eligibility-outreach effects: outreach aimed at people with no measured effect on EITC filing/claiming behavior (Linos 2020), and outreach aimed at populations already engaged with tax filing, where positive effects on take-up or filing rates are reported (Bhargava and Manoli 2015, Bettinger et al. 2012, Finkelstein and Notowidigdo 2019, Guyton et al. 2017). No claim in this set addresses replication of a Gerber (2014) eligibility-outreach RCT specifically, in other states, other time periods, or otherwise.
Where the evidence splits
Linos (2020) found no measured effect of behavioral nudges on EITC filing or claiming, while Bhargava and Manoli (2015), Bettinger et al. (2012), and Finkelstein and Notowidigdo (2019) are cited as finding positive effects on take-up. The source states a design difference behind this split: the Bhargava and Manoli (2015) and Bettinger et al. (2012) interventions were conducted among taxpayers who had already filed a return and only needed to be nudged to complete additional forms covering similar material, whereas the Linos (2020) trial targeted the filing/claiming decision itself.
Guyton et al. (2017) found a positive effect of outreach on tax filing rates, but at about 0.5 percentage points — described as much smaller than the effect found by Bhargava and Manoli (2015). No difference in study design, population, or period is given in the claims to explain this gap in magnitude; it is unexplained by the sources on this page.