Church, D., Hawk, C., Brooks, A.J., Toukolehto, O., Wren, M., Dinter, I. et al. · 2013
The primary paper (Church, Hawk, Brooks, Toukolehto, Wren, Dinter & Stein, 2013, Journal of Nervous and Mental Disease, 201(2):153-160, PMID 23364126) does not report a Cohen's d anywhere for any outcome, including depression — it reports F-statistics and p-values only. Verbatim from its Table 2 (SA-45 depression subscale): EFT group pretest 71.10 (SE 1.4) to post-6-sessions 57.71 (SE 1.9), F(1,51)=32.16, p<.0001, versus a standard-care/waitlist group that changed minimally (71.49 to 69.77 over the same period); 'EFT posttest lower than SOC/WL posttest, p<0.008' (between-group) and 'EFT posttest lower than EFT pretest, p<0.003' (within-group). A rough recalculation from these means/SEs suggests a within-group effect in the range of d≈1.3-1.8, consistent with the independently-reported Nelms & Castel (2016) meta-analysis figure of d=1.31 for this literature — far below the previously recorded d=8.02, which does not appear in the primary text in any form and is not plausibly derivable from the paper's own reported statistics. The trial randomized N=59 total (30 EFT, 29 standard-care/waitlist); N=49 (as previously recorded) was actually the number of participants who completed all six EFT sessions across both arms after crossover, not the total randomized sample.
This is a big deal because it's the earliest published peer-reviewed randomized trial of tapping for veteran PTSD and depression — the study that a long chain of later reviews and meta-analyses draws its numbers from. That makes getting it exactly right unusually important: correcting an inflated, unsupported effect size back to what the paper actually showed doesn't make the underlying finding less real, it makes it more trustworthy. Honest, verifiable evidence is what earns skeptical clinicians' attention — inflated numbers just invite dismissal.
If the corrected, more modest effect size still holds up in replication, picture a veteran with PTSD-linked depression given six sessions of a technique that produced genuine, meaningful improvement, and that the veteran can then continue administering to themselves at no cost afterward, not the implausibly large effect once cited, but a real one nonetheless. Getting the number right matters because it's what lets clinicians and veterans make honest, well-calibrated decisions about trying it.
With the numbers now correctly reconciled to the paper's actual reported statistics, the compelling next step is re-running a larger version of this original design with the biomarker tools that weren't available at the time — cortisol, inflammatory panels, or heart-rate variability — to see whether this foundational finding, honestly reported, replicates and deepens with today's measurement toolkit.
| Design | Randomized trial |
|---|---|
| Participants | 59 people |
| Population | veterans with PTSD |
| Comparison group | treatment as usual |
| Outcome measures | SA-45 |
| Journal | Original publication venue not confirmed (indexed via Nelms & Castel 2016 Table 4) |
| Year | 2013 |
| Country | United States |
| Language | English |
| Method | EFT / tapping |
| Publication type | Study / trial |
| Verification | ✓ Confirmed against the primary source |
Church, D., Hawk, C., Brooks, A.J., Toukolehto, O., Wren, M., Dinter, I., & Stein, P. (2013). Veterans PTSD trial — depression outcome (as tabulated in Nelms & Castel 2016). https://doi.org/10.1097/NMD.0b013e31827f6351
This record is part of the Tapping Evidence Base — an openly-sourced, fully-referenced directory of the research on EFT/tapping. Explore more studies on Depression · PTSD & Trauma
A ready-made graphic — right-click or long-press to save the image.