Longitudinal comparison of psychological outcomes of professional content moderators engaged in resilience training program
Open accessThe DOI matches its Crossref record; 32 of the 32 references it cites were found in a registry or an index; 9 of the 31 sources its claims cite were checked for support.
Abstract
Introduction: Content moderators safeguard the ever-growing online realm by reviewing and filtering out harms, yet are themselves subject to the risk of psychological concerns. Method: This study explored the psychological outcomes of content moderators (N = 311) who were offered a resilience training program called The TaskUs Method 2.0 from baseline to 18 months. Participants completed psychometric screeners for resilience, burnout, secondary traumatic stress, compassion satisfaction, and perceived stress. For the analyses, three timepoints were identified to follow participants across the study period based on their tenure-T0 (0-3 months tenure), T1 (5-10 months), and T2 (11-16 months tenure). Results: Findings from the entire sample (regardless of whether participants had data for all three timepoints) revealed that secondary traumatic stress and perceived stress did not differ significantly between timepoints. However, burnout, compassion satisfaction and resilience showed small yet significant decline between T0 and T2. With a reduced sample of only those having data for all three timepoints, similar trends remained although the rate of change from T0 to T1 did not differ significantly from that of T1 to T2 for any of the variables. Discussion: Results indicated that participants in the program showed no significant increases in secondary traumatic stress, the main target of the program. Across the different psychometrics, effect sizes were small suggesting that content moderators using the program were not at significant risk for increased psychological distress from content moderation during the study duration. These findings provide a valuable foundation for further refining interventions aimed at content moderators. Future studies may consider experimental and multifactorial designs to delve into therapeutic mechanisms and different types of psychological outcomes.
Overview
CoChat Summary
## TL;DR
This longitudinal study followed 311 full-time content moderators in TaskUs Method 2.0 for up to 18 months and found no significant worsening in secondary traumatic stress or perceived stress. Burnout, compassion satisfaction, and resilience changed modestly, with small effect sizes and overall minimal clinical worsening.
## Key findings
- In the full sample (N = 311), secondary traumatic stress did not differ significantly between timepoints: T1 vs T0, T2 vs T1, and T2 vs T0 were all non-significant.
- Perceived stress also showed no significant change between T0 and T1 or between T1 and T2; no T0-to-T2 comparison was possible because no participants had data at all three timepoints for this measure.
- Burnout increased modestly: T1 was higher than T0 by 1.15 points (p < .001, d = .22), and T2 was higher than T0 by 1.16 points (p = .022, d = .20), while T1 vs T2 was not significant.
- Compassion satisfaction and resilience declined slightly: compassion satisfaction fell by 1.23 points from T0 to T1, 0.93 from T1 to T2, and 1.55 from T0 to T2; resilience fell by 0.65 from T0 to T1 and 0.87 from T0 to T2, with small effect sizes.
## Methods
- Secondary analysis of repeated psychometric survey data from full-time content moderators employed by TaskUs; surveys were administered at 6-month intervals from May 2022 to November 2023.
- Sample size was 311 moderators overall; 131 had data at all three tenure-based timepoints (T0: 0–3 months, T1: 5–10 months, T2: 11–16 months).
- Outcomes assessed were resilience (CD-RISC 10), burnout, secondary traumatic stress, compassion satisfaction (ProQOL), and perceived stress (PSS); analyses used paired t-tests in the full sample and follow-up comparisons among those with complete data at all three timepoints.
## Limitations
- The paper states there was no randomization, no experimental control group, and no internal comparator, which prevents definitive causal conclusions about the intervention.
- The paper states the outcomes were self-reported and drawn from a single organization, introducing reporting and desirability bias.
- The paper states the sample was grouped by tenure rather than calendar time, and that dropout reasons could not be confirmed; evident additional limitation: this raises possible self-selection/survival bias and limits control over seasonality and exposure differences.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures






Verification
Claims the paper makes
6 Valid · 2 Not valid · 17 Not checked · 91 Own results and interpretations
At least one cited source doesn’t support this claim or points to a different paper, and none confirmed it — see the verdicts below.
In the paper
“In comparison with previous studies on content moderators wherein psychological distress was at clinically moderate to high levels (e.g., 47, 48), in the current sample, lower scores of secondary traumatic stress and burnout for instance across time were observed.”
§ Longitudinal comparison of psychological outcomes of professional content moderators engaged in resilience training program > Discussion
What the sources say
- Not checked — no full text of the sourcecites [47] Spence, 2024, Content moderator mental health, secondary trauma, and well-being: A cross-sectional study, Cyberpsychol Behav Soc Network source
- Disputedcites [48] Spence, 2025, Content moderator mental health and associations with coping styles: replication and extension of previous studies, Behav Sci sourceExperimental — accuracy not yet measured
Split into 2 atoms — 1 confirmed, 0 partially supported, 1 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
Checked by a reasoning model — may be wrong.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“On one hand some studies (13–15), evaluating technological interventions such as image modification through blurring and filters, found positive effects like reduced negative feelings while retaining operational efficiency.”
§ Longitudinal comparison of psychological outcomes of professional content moderators engaged in resilience training program > Introduction
What the sources say
- Supportedcites [13] Das, 2020, Fast, accurate, and healthier: interactive blurring helps moderators reduce exposure to harmful content, Proc AAAI Conf Hum Comput Crowdsourcing source
The Conclusion (Section 6) and Introduction state the paper's own finding verbatim in the claim's terms: interactive blur interfaces reduce emotional impact on moderators without sacrificing accuracy or speed, while static blur lowers accuracy (Sec. 4 Evaluation, Fig. 5a). Section 4.3 reports lower mean negative emotion and significantly lower emotional exhaustion for hover vs. the unblurred baseline, plus increased positive affect with blur (I-PANAS-SF). The lit review also credits Karunakaran and Ramakrishan (2019) for grayscale/blur transformations, covering the 'filters' part of the paraphrase. (LLM, conf=0.88)
- Partly supportedcites [14] Karunakaran, 2019, Testing stylistic interventions to reduce emotional impact of content moderation workers, Seventh AAAI Conf Hum Comput Crowdsourcing (HCOMP-19) sourceExperimental — accuracy not yet measured
Karunakaran et al. (2019) is indeed a technological image-modification intervention (grayscale and blur) tested on live content-review queues, and it supports the 'retaining operational efficiency' half: review accuracy showed no significant difference under grayscale (F = 0.071, p = .791) and handling time actually fell (1.88 → 1.55 min), and under blurring neither accuracy (p = .664) nor AHT (p = .168) differed significantly (Grayscale section, Tables 2 and 4). However, the 'reduced negative feelings' part is not what this paper reports: for grayscale it states 'We did not observe significant differences in negative affect (Z = -1.233, p < 0.217)', the benefit being a significant increase in POSITIVE affect, and for blurring positive affect fell significantly and 'Irritable' increased significantly (p = 0.003), with 78% of reviewers unlikely to continue using blur. So the source supports improved reviewer affect plus preserved efficiency, but not a reduction in negative feelings, and it does not support blurring as having positive emotional effects. (LLM, conf=0.82)
- Supportedcites [15] Sarridis, 2024, Mitigating viewer impact from disturbing imagery using AI filters: A user-study, Int J Human–Comp Interac source
The source is exactly the kind of study the citing sentence describes: a 107-participant user study evaluating technological image-modification interventions (conventional Blurring and Partial Blurring plus AI filters) on both negative-feelings mitigation and information retention. Its Abstract (p. 1) and Section 4.2 (p. 7) report reduced negative feelings (-30.38% for Drawing) while preserving interpretability (97.19%), and Section 4.2 (p. 8) likewise reports Partial Blurring retaining 96.26% interpretability with 25.54% mitigation; the Introduction (p. 2) frames that retained detail as 'necessary for various professional purposes', matching 'operational efficiency'. The only caveat is that full Blurring in this study caused severe information loss (6.54% interpretability), but the claim only names blurring as an example of the intervention class, not as individually retaining efficiency. (LLM, conf=0.85)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, burnout, compassion satisfaction and resilience showed small yet significant decline between T0 and T2.”
§ Longitudinal comparison of psychological outcomes of professional content moderators engaged in resilience training program > Abstract > Results
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“and T2 was not significantly different from T1 [t(130) = 1.24, p=.216, d = .11].”
§ Longitudinal comparison of psychological outcomes of professional content moderators engaged in resilience training program > Results > Full sample analyses
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“With a reduced sample of only those having data for all three timepoints, similar trends remained although the rate of change from T0 to T1 did not differ significantly from that of T1 to T2 for any of the variables.”
§ Longitudinal comparison of psychological outcomes of professional content moderators engaged in resilience training program > Abstract > Results
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
58 more claims
Create a free account to see all 116 claims and to verify the 4 references still open.
- Probably not a substantial claim. 31
- It may state two separate claims as one. 35
- Borderline: one of the tests for a substantial claim is unclear. 1
Statistics recomputed
Every reported test result, recomputed from its test statistic and degrees of freedom. A result shows as not adding up only once an independent check has confirmed it.
- Not confirmedt(130) = 0.93reported p = .351 · recomputed p = .354
“t(130) = .93, p = .351”
- Consistentt(310) = 0.895reported p = .371 · recomputed p = .371
“t (310) = .895, p = .371”
- Consistentt(130) = 1.24reported p = .216 · recomputed p = .217
“t(130) = 1.24, p = .216”
- Consistentt(130) = 1.14reported p = .255 · recomputed p = .256
“t(130) = 1.14, p = .255”
- Consistentt(109) = 1.39reported p = .169 · recomputed p = .167
“t(109) = 1.39, p = .169”
17 more recomputed tests
Create a free account to see every test the Stats check recomputed.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Self-check results appear here once an independent check has confirmed them.
Citations
Cites
A Global Measure of Perceived Stress
Cohen, S., Kamarck, T., & Mermelstein, R. (1983) · Journal of Health and Social Behavior · doi:10.2307/2136404 · Cited by 32,807
Development of a new resilience scale: The Connor-Davidson Resilience Scale (CD-RISC)
Connor, K. M., & Davidson, J. R. T. (2003) · Depression and Anxiety · doi:10.1002/da.10113 · Cited by 11,793
Campbell‐Sills, L., & Stein, M. B. (2007) · Journal of Traumatic Stress · doi:10.1002/jts.20271 · Cited by 3,575
Cusack, K., Jonas, D. E., Forneris, C. A., Wines, C., Sonis, J., Middleton, J. C., Feltner, C., Brownley, K. A., Olmsted, K. R., Greenblatt, A., et al. (2016) · Clinical Psychology Review · doi:10.1016/j.cpr.2015.10.003 · Cited by 865
Dialectical behavior therapy is effective for the treatment of suicidal behavior: A meta-analysis
DeCou, C. R., Comtois, K. A., & Landes, S. J. (2019) · Behavior Therapy · doi:10.1016/j.beth.2018.03.009 · Cited by 439
Bajaj, B., & Pande, N. (2016) · Personality and Individual Differences · doi:10.1016/j.paid.2015.09.005 · Cited by 422 · cited 1×
“Research suggests that these activities and topics support resiliency and psychological functioning, especially in those exposed to graphic materials or who have experienced a traumatic event (42 – 44).”
Neacsiu, A. D., Eberle, J. W., Kramer, R., Wiesmann, T., & Linehan, M. M. (2014) · Behaviour Research and Therapy · doi:10.1016/j.brat.2014.05.005 · Cited by 394
Somatic experiencing: using interoception and proprioception as core elements of trauma therapy
Payne, P., Levine, P. A., & Crane-Godreau, M. A. (2015) · Frontiers in Psychology · doi:10.3389/fpsyg.2015.00093 · Cited by 344 · cited 1×
“…on brain-based and mind-body interventions (26, 28) namely, Eye-Movement Desensitization and Reprocessing (30, 31), Somatic Experiencing (32 – 34), and Rapid Resolution Therapy (35, 36), coupled with cognitive modalities to include Cognitive Processing Therapy (27, 37 – 39) and Dialectical…”
Adult neuroplasticity: more than 40 years of research
Fuchs, E., & Flügge, G. (2014) · Neural Plasticity · doi:10.1155/2014/541870 · Cited by 325 · cited 1×
“The goal of the sessions was to foster neuroplasticity and improve moderators ’ stress tolerance as well as healthy coping techniques (29).”
Haglund, M. E. M., Nestadt, P. S., Cooper, N. S., Southwick, S. M., & Charney, D. S. (2007) · Development and Psychopathology · doi:10.1017/s0954579407000430 · Cited by 314
Research works printed but not linked: 9
- Cook, 2018, Awe versus aww: the effectiveness of two kinds of positive emotional stimulation on stress reduction for online content moderators, 19
- Stamm, 2010, The concise proQOL manual
- Shapiro, 2001, Eye movement desensitization and reprocessing: Basic principles, protocols, and procedures
- Levine, 2015, Trauma and memory: Brain and body in a search for the living past: A practical guide for understanding and working with traumatic memory
- Bowles, 2019, An integrated rapid memory reconsolidation approach: rapid resolution therapy, Neuropsychotherapist, 7, 43
- Rivera, 2016, Reconnecting the mind and body: using bodywork to help improve mental health
- Kouba, 2017, Group cognitive processing therapy for veterans experiencing trauma:A systematic review
- Linehan, 1993, Cognitive-behavioral treatment of borderline personality disorder
- Schiraldi, 2017, The resilience workbook essential skills to recover from stress, trauma, and adversity
Cited by
Commercial content moderation: Trauma exposure on the digital frontline
DeMarco, J. N., Watson, P., & Norman, S. B. (2026) · Journal of Traumatic Stress · doi:10.1002/jts.70087 · Cited by 1
Metadata
- Authors
- Savio, M. T., Balammal, A., de Villa, J. C., Perez, J., Huang, X., & Guevara, R. L.
- Corresponding author
- Balammal, Aruna
- Venue
- Frontiers in Psychiatry
- Publisher
- Frontiers Media SA
- Published
- 2025-12-04
- Volume
- 16
- Pages
- 1666543-1666543
- ISSN
- 1664-0640 (ISSN-L)
- PMC
- PMC12711784
- OpenAlex
- W4416976801
- Cited by
- 1
- References
- 31
- Citations by year
- 2026: 1
- FWCI
- 1.10
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Frontiers in Psychiatry (publishedVersion, cc-by)
- PubMed (publishedVersion)
- PubMed Central (publishedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Topics
- Social Sciences › Psychology › Social Psychology › Stress and Burnout Research
- Health Sciences › Medicine › Public Health, Environmental and Occupational Health › Occupational Health and Burnout
- Social Sciences › Psychology › Clinical Psychology › Resilience and Mental Health
- Keywords
- resilience training, burnout, secondary traumatic stress, compassion satisfaction, perceived stress, psychological distress, online content moderation
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Longitudinal comparison of psychological outcomes of professional content moderators engaged in resilience training program. https://app.cochat.ai/papers/10.3389/fpsyt.2025.1666543
- The paper
- Savio, M. T., Balammal, A., de Villa, J. C., Perez, J., Huang, X., & Guevara, R. L. (2025). Longitudinal comparison of psychological outcomes of professional content moderators engaged in resilience training program. Frontiers in Psychiatry. https://doi.org/10.3389/fpsyt.2025.1666543
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.