pith. sign in

arxiv: 2605.20438 · v1 · pith:WGSWKXDHnew · submitted 2026-05-19 · 💻 cs.HC

Closing the Motivation Gap: Incentives Enhance Visual Misinformation Discernment and Verification

Sijia Qian , Cuihua Shen , Jingwen Zhang , Magdalena Wojcieszak This is my paper

Pith reviewed 2026-05-21 07:02 UTC · model grok-4.3

classification 💻 cs.HC
keywords visual misinformationcheapfakesincentivesmedia literacyreverse image searchdiscernment accuracysocial mediaverification behavior
0
0 comments X

The pith

Task-based monetary incentives best prompt reverse image searches and short-term discernment of misleading images, while result-based incentives better sustain accuracy over time.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

People often fail to verify images even when they know how, because motivation is missing. This study added different incentives to a media literacy intervention on a simulated social media platform and tracked behavior across two waves. Task-based monetary rewards got participants to actually perform reverse image searches right away and raised immediate accuracy at spotting cheapfakes. Result-based incentives, by contrast, helped keep discernment accuracy higher in the later wave. The pattern suggests that incentive design can close the gap between knowing verification steps and using them.

Core claim

In a pre-registered two-wave between-subjects experiment with 1,421 participants, task-based incentives, especially when monetary, most effectively started image verification behaviors such as reverse image search and improved short-term discernment of cheapfakes presented in misleading contexts, whereas result-based incentives proved more effective at sustaining discernment accuracy across the second wave.

What carries the argument

A 2 (incentive type: symbolic versus monetary) by 2 (incentive mechanism: task-based versus result-based) factorial design with control groups, delivered inside a professionally designed social media platform simulation.

If this is right

  • Task-based monetary incentives increase immediate use of reverse image search compared with no-incentive controls.
  • Result-based incentives produce higher discernment accuracy in a follow-up session than task-based or control conditions.
  • Short-term gains from task-based monetary rewards fade without follow-up mechanisms.
  • Media literacy programs may need separate phases: one to initiate verification and another to maintain it.
  • Both incentive type and mechanism matter; neither alone explains the full pattern of short- and long-term effects.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Real platforms might test small result-based rewards to encourage repeated verification without constant external prompts.
  • The same incentive split could be examined for text or video misinformation to see whether the pattern generalizes beyond images.
  • Future designs could combine both mechanisms in sequence within one intervention to capture both initiation and maintenance benefits.
  • Demand effects remain a concern; studies that hide the incentive purpose from participants would strengthen causal claims.

Load-bearing premise

Observed changes in verification behavior and accuracy are caused by the incentive conditions themselves rather than by participants guessing the study goals or by different dropout rates between the two waves.

What would settle it

A replication that finds no differences in reverse-image-search rates or discernment accuracy after balancing attrition and removing obvious demand cues would show the incentive effects are not robust.

Figures

Figures reproduced from arXiv: 2605.20438 by Cuihua Shen, Jingwen Zhang, Magdalena Wojcieszak, Sijia Qian.

Figure 3
Figure 3. Figure 3: Digital media literacy infographic teaching reverse image search [PITH_FULL_IMAGE:figures/full_fig_p017_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Mean discernment and verification intention by incentive strategies. Note. This figure presents group means for discernment and verification intention across five experimental conditions: task-based symbolic, task-based monetary, result-based symbolic, result-based monetary, and no incentive. Measures were taken immediately after the intervention (Wave 1) and one week later (Wave 2). (a) Mean discernment s… view at source ↗
Figure 5
Figure 5. Figure 5: Effects of incentive strategies on discernment. Note. This figure displays the estimated coefficients and 95% confidence intervals for each incentive condition (vs. the no-incentive control group) on discernment (left) and verification intention (right) across two waves. Horizontal lines represent coefficient estimates from Wave 1 (upper point) and Wave 2 (lower point) for each condition. Models control fo… view at source ↗
read the original abstract

Cheapfakes, or real images presented misleadingly or in unrelated contexts, are an increasingly prominent form of visual misinformation. While media literacy interventions can enhance individuals' ability to detect such content, motivational barriers often hinder the adoption of image verification. This study examines whether incorporating different mechanisms and types of incentives into a digital media literacy intervention improves visual misinformation discernment and image verification behavior, both immediately and over time. We conducted a pre-registered two-wave between-subjects online experiment (N = 1,421) on a professionally designed social media platform. The study used a 2 (Incentive Type: symbolic vs. monetary) x 2 (Incentive Mechanism: task- vs. result-based) factorial design with additional control groups. Results show that task-based incentives, particularly monetary ones, were most effective at initiating image verification behaviors, namely reverse image search, and boosting short-term discernment, whereas result-based incentives were more effective in sustaining discernment accuracy. These findings suggest that both the mechanism and the type of incentives play a critical role in shaping the short- and long-term effectiveness of media literacy interventions, highlighting the value of multi-phased incentive strategies for combating visual misinformation in digital environments.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 3 minor

Summary. The manuscript reports a pre-registered 2×2 factorial between-subjects online experiment (N=1,421) on a simulated social media platform testing incentive type (symbolic vs. monetary) crossed with incentive mechanism (task-based vs. result-based), plus control groups. A two-wave design measures immediate effects on image verification (e.g., reverse image search) and discernment accuracy as well as sustained accuracy in the follow-up wave. The central claim is that task-based monetary incentives most effectively initiate verification behaviors and short-term discernment gains, whereas result-based incentives better maintain accuracy over time.

Significance. If the attribution of effects to the incentive manipulations holds after addressing potential confounds, the study would offer useful empirical guidance on designing motivational supports for media literacy interventions against cheapfakes. The pre-registration, large sample, and two-wave structure are clear strengths that allow credible short- versus long-term comparisons and could inform platform-level incentive strategies in HCI and misinformation research.

major comments (2)
  1. [Experimental procedure] Experimental procedure: The manuscript provides no description of suspicion probes, debriefing questions, or other checks for demand characteristics. Because monetary incentives are especially salient in a between-subjects design, participants may have inferred the study purpose and altered verification or discernment behavior accordingly; without such checks the directional differences cannot be confidently attributed to the incentive conditions rather than demand effects.
  2. [Two-wave design and results] Two-wave design and results: No attrition tables by condition, differential dropout rates, or sensitivity analyses for missing data are reported. The claim that result-based incentives sustain discernment accuracy rests on the second wave; non-random attrition across arms could produce the observed pattern without any true incentive effect, making this a load-bearing concern for the longitudinal conclusions.
minor comments (3)
  1. [Abstract] Abstract: The abstract states the sample size and pre-registration but does not briefly define 'cheapfakes' for readers outside the subfield.
  2. [Results] Results: Exact statistical models (e.g., regression specifications, covariates, handling of multiple comparisons) and effect sizes are not fully detailed, which limits assessment of practical significance.
  3. [Discussion] Discussion: The simulated platform's ecological validity relative to real social media interfaces could be addressed more explicitly to strengthen external validity claims.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive feedback on our pre-registered experiment. The comments highlight important issues for strengthening causal attribution and the longitudinal claims. We respond to each major comment below and have prepared revisions to address them.

read point-by-point responses

  1. Referee: Experimental procedure: The manuscript provides no description of suspicion probes, debriefing questions, or other checks for demand characteristics. Because monetary incentives are especially salient in a between-subjects design, participants may have inferred the study purpose and altered verification or discernment behavior accordingly; without such checks the directional differences cannot be confidently attributed to the incentive conditions rather than demand effects.

    Authors: We agree that explicit checks for demand characteristics are necessary to support attribution of effects to the incentive manipulations. Our procedure included a post-task debriefing that asked participants to describe what they believed the study was about and whether they suspected any connection between incentives and verification tasks. We will add a dedicated subsection in the Methods describing these questions and report the distribution of responses (which showed low rates of accurate suspicion) in the revised manuscript to rule out demand effects as an alternative explanation. revision: yes

  2. Referee: Two-wave design and results: No attrition tables by condition, differential dropout rates, or sensitivity analyses for missing data are reported. The claim that result-based incentives sustain discernment accuracy rests on the second wave; non-random attrition across arms could produce the observed pattern without any true incentive effect, making this a load-bearing concern for the longitudinal conclusions.

    Authors: We concur that transparent reporting of attrition is essential for the credibility of the two-wave results. We have now generated condition-specific attrition tables and conducted sensitivity analyses (including inverse probability weighting and multiple imputation). These will be added to the Results section of the revised manuscript, with explicit discussion of whether differential dropout alters the interpretation of sustained accuracy under result-based incentives. revision: yes

Circularity Check

0 steps flagged

No circularity: purely empirical experiment with independent data

full rationale

This paper reports a pre-registered two-wave between-subjects factorial experiment (N=1,421) testing incentive mechanisms and types on visual misinformation discernment and verification behaviors. No equations, derivations, or self-citations reduce any reported outcome to fitted parameters or prior results by construction. All central claims rest on direct experimental manipulations, measured behaviors (e.g., reverse image search uptake), and statistical comparisons across conditions, with the design and pre-registration providing external grounding independent of the present data. The study is self-contained against its own empirical benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard experimental assumptions about random assignment, incentive perception, and measurement validity rather than new free parameters or invented entities.

axioms (2)
  • standard math Random assignment to incentive conditions produces equivalent groups at baseline
    Invoked by the between-subjects factorial design described in the abstract.
  • domain assumption Participants interpret and respond to the incentive manipulations as intended without substantial demand characteristics
    Required for attributing behavioral changes to the incentive types and mechanisms.

pith-pipeline@v0.9.0 · 5750 in / 1257 out tokens · 85998 ms · 2026-05-21T07:02:29.933341+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

39 extracted references · 39 canonical work pages

  1. [1]

    Dan, et al., Visual Mis- and Disinformation, Social Media, and Democracy

    V. Dan, et al., Visual Mis- and Disinformation, Social Media, and Democracy. Journalism & Mass Communication Quarterly 98, 641–664 (2021)

    work page 2021

  2. [2]

    Hameleers, T

    M. Hameleers, T. E. Powell, T. G. L. A. V. D. Meer, L. Bos, A Picture Paints a Thousand Lies? The Effects and Mechanisms of Multimodal Disinformation and Rebuttals Disseminated via Social Media. Political Communication 37, 281–301 (2020)

    work page 2020

  3. [3]

    Tucker, et al., Social Media, Political Polarization, and Political Disinformation: A Review of the Scientific Literature

    J. Tucker, et al., Social Media, Political Polarization, and Political Disinformation: A Review of the Scientific Literature. SSRN Journal (2018). https://doi.org/10.2139/ssrn.3144139

    work page doi:10.2139/ssrn.3144139 2018

  4. [4]

    J. S. Brennen, F. M. Simon, R. K. Nielsen, Beyond (Mis)Representation: Visuals in COVID-19 Misinformation. Int J Press Polit (2020). https://doi.org/10.1177/1940161220964780

    work page doi:10.1177/1940161220964780 2020

  5. [5]

    Garimella, D

    K. Garimella, D. Eckles, Images and misinformation in political groups: Evidence from WhatsApp in India. Harvard Kennedy School Misinformation Review (2020). https://doi.org/10.37016/mr-2020-030

    work page doi:10.37016/mr-2020-030 2020

  6. [6]

    Hameleers, Cheap versus deep manipulation: the effects of cheapfakes versus deepfakes in a political setting

    M. Hameleers, Cheap versus deep manipulation: the effects of cheapfakes versus deepfakes in a political setting. International Journal of Public Opinion Research 36, edae004 (2024)

    work page 2024

  7. [7]

    Deepfakes and cheap fakes

    B. Paris, J. Donovan, “Deepfakes and cheap fakes” (Data & Society Research Institute, 2019)

    work page 2019

  8. [8]

    A. M. Guess, et al., A digital media literacy intervention increases discernment between mainstream and false news in the United States and India. PNAS 117, 15536–15545 (2020)

    work page 2020

  9. [9]

    E. Hoes, B. Aitken, J. Zhang, T. Gackowski, M. Wojcieszak, Prominent misinformation interventions reduce misperceptions but increase scepticism. Nat Hum Behav 8, 1545–1553 (2024)

    work page 2024

  10. [10]

    S. Qian, C. Shen, J. Zhang, Fighting cheapfakes: using a digital media literacy intervention to motivate reverse search of out-of-context visual misinformation. Journal of Computer-Mediated Communication 28, zmac024 (2023)

    work page 2023

  11. [11]

    Hertwig, T

    R. Hertwig, T. Grüne-Yanoff, Nudging and Boosting: Steering or Empowering Good Decisions. Perspect Psychol Sci 12, 973–986 (2017)

    work page 2017

  12. [12]

    Johansson, et al., How can we combat online misinformation? A systematic overview of current interventions and their efficacy

    P. Johansson, et al., How can we combat online misinformation? A systematic overview of current interventions and their efficacy. [Preprint] (2022). Available at: http://arxiv.org/abs/2212.11864 [Accessed 18 February 2023]

    work page arXiv 2022

  13. [13]

    W. Yu, F. Shen, Mapping verification behaviors in the post-truth era: A systematic review. New Media & Society 14614448231191138 (2023). https://doi.org/10.1177/14614448231191138. 13

    work page doi:10.1177/14614448231191138 2023

  14. [14]

    R. C. Moore, J. T. Hancock, A digital media literacy intervention for older adults improves resilience to fake news. Sci Rep 12, 6008 (2022)

    work page 2022

  15. [15]

    J. E. Brodsky, et al., Improving college students’ fact-checking strategies through lateral reading instruction in a general education civics course. Cogn. Research 6, 23 (2021)

    work page 2021

  16. [16]

    C. P. Cerasoli, J. M. Nicklin, M. T. Ford, Intrinsic motivation and extrinsic incentives jointly predict performance: A 40-year meta-analysis. Psychological Bulletin 140, 980–1008 (2014)

    work page 2014

  17. [17]

    F. H. Schneider, et al., Financial incentives for vaccination do not have negative unintended consequences. Nature 613, 526–533 (2023)

    work page 2023

  18. [18]

    C. E. Wollbrant, M. Knutsson, P. Martinsson, Extrinsic rewards and crowding-out of prosocial behaviour. Nat Hum Behav 1–8 (2022). https://doi.org/10.1038/s41562-022-01293-y

    work page doi:10.1038/s41562-022-01293-y 2022

  19. [19]

    Hechter, S

    M. Hechter, S. Kanazawa, Sociological Rational Choice Theory. Annual Review of Sociology 23, 191–214 (1997)

    work page 1997

  20. [20]

    Panizza, et al., Lateral reading and monetary incentives to spot disinformation about science

    F. Panizza, et al., Lateral reading and monetary incentives to spot disinformation about science. Sci Rep 12, 5678 (2022)

    work page 2022

  21. [21]

    Rathje, J

    S. Rathje, J. Roozenbeek, J. J. Van Bavel, S. van der Linden, Accuracy and social motivations shape judgements of (mis)information. Nat Hum Behav 1–12 (2023). https://doi.org/10.1038/s41562-023-01540-w

    work page doi:10.1038/s41562-023-01540-w 2023

  22. [22]

    Kapoor, et al., Does incentivization promote sharing “true” content online? Harvard Kennedy School Misinformation Review (2023)

    H. Kapoor, et al., Does incentivization promote sharing “true” content online? Harvard Kennedy School Misinformation Review (2023). https://doi.org/10.37016/mr-2020-120

    work page doi:10.37016/mr-2020-120 2023

  23. [23]

    Speckmann, C

    F. Speckmann, C. Unkelbach, Monetary incentives do not reduce the repetition-induced truth effect. Psychonomic Bulletin & Review 29, 1045–1052 (2022)

    work page 2022

  24. [24]

    Steinhorst, C

    J. Steinhorst, C. A. Klöckner, Effects of monetary versus environmental information framing: Implications for long-term pro-environmental behavior and intrinsic motivation. Environment and Behavior 50, 997–1031 (2018)

    work page 2018

  25. [25]

    Gallus, Fostering public good contributions with symbolic awards: A large-scale natural field experiment at Wikipedia

    J. Gallus, Fostering public good contributions with symbolic awards: A large-scale natural field experiment at Wikipedia. Management Science 63, 3999–4015 (2017)

    work page 2017

  26. [26]

    Gneezy, S

    U. Gneezy, S. Meier, P. Rey-Biel, When and Why Incentives (Don’t) Work to Modify Behavior. Journal of Economic Perspectives 25, 191–210 (2011)

    work page 2011

  27. [27]

    Gallus, Fostering Public Good Contributions with Symbolic Awards: A Large-Scale Natural Field Experiment at Wikipedia

    J. Gallus, Fostering Public Good Contributions with Symbolic Awards: A Large-Scale Natural Field Experiment at Wikipedia. [Preprint] (2017). Available at: https://papers.ssrn.com/abstract=2579118 [Accessed 29 January 2026]

    work page 2017

  28. [28]

    E. L. Deci, R. Koestner, R. M. Ryan, A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin 125, 627–668 (1999). 14

    work page 1999

  29. [29]

    A. K. Boggiano, J. M. Harackiewicz, J. M. Bessette, D. S. Main, Increasing Children’s Interest through Performance-Contingent Reward. Social Cognition 3, 400–411 (1985)

    work page 1985

  30. [30]

    M. S. Gazzaniga, Who’s in charge?: free will and the science of the brain (Robinson, 2016)

    work page 2016

  31. [31]

    What” and “Why

    E. L. Deci, R. M. Ryan, The “What” and “Why” of Goal Pursuits: Human Needs and the Self-Determination of Behavior. Psychological Inquiry 11, 227–268 (2000)

    work page 2000

  32. [32]

    S. M. Noar, C. N. Benac, M. S. Harris, Does tailoring matter? Meta-analytic review of tailored print health behavior change interventions. Psychological Bulletin 133, 673–693 (2007)

    work page 2007

  33. [33]

    R. L. Fu Tom Jones, Angela, Fact-checkers are out. The internet gets to vote on the truth now. Poynter (2025). Available at: https://www.poynter.org/commentary/2025/fact-checkers- out-community-notes-in/ [Accessed 6 November 2025]

    work page 2025

  34. [34]

    Chuai, H

    Y. Chuai, H. Tian, N. Pröllochs, G. Lenzini, Did the Roll-Out of Community Notes Reduce Engagement With Misinformation on X/Twitter? Proc. ACM Hum.-Comput. Interact. 8, 428:1-428:52 (2024)

    work page 2024

  35. [35]

    Bouchaud, P

    P. Bouchaud, P. Ramaciotti, Algorithmic resolution of crowd-sourced moderation on X in polarized settings across countries. [Preprint] (2025). Available at: http://arxiv.org/abs/2506.15168 [Accessed 6 November 2025]

    work page arXiv 2025

  36. [36]

    Ajzen, The theory of planned behavior

    I. Ajzen, The theory of planned behavior. Organizational Behavior and Human Decision Processes 50, 179–211 (1991)

    work page 1991

  37. [37]

    Tardáguila, How to use your phone to spot fake images surrounding the U.S.-Iran conflict

    C. Tardáguila, How to use your phone to spot fake images surrounding the U.S.-Iran conflict. Poynter (2020). Available at: https://www.poynter.org/fact-checking/2020/how-to-use- your-phone-to-spot-fake-images-surrounding-the-u-s-iran-conflict/ [Accessed 8 June 2021]

    work page 2020

  38. [38]

    College graduate (B.S., B.A., or other 4-year degree)

    B. Guay, A. J. Berinsky, G. Pennycook, D. Rand, How to think about whether misinformation interventions work. Nat Hum Behav 7, 1231–1233 (2023). 15 Figures and Tables Figure 1. Overall design of the experiment. 16 Figure 2. Experimental platform shown to the task-based monetary incentive group 17 Figure 3. Digital media literacy infographic teaching rever...

    work page 2023

  39. [39]

    How confident are you in your answer to the previous question about the credibility of the post?

    Age 42.20 14.12 2. Income 6.64 3.19 .04 [-.02, .09] 3. Educati on 5.38 1.24 .12** .38** [.07, .17] [.33, .42] 4. Media use 4.54 1.78 -.22** .14** .05 [-.27, - .17] [.09, .19] [-.01, .10] 5. Politica l leaning 3.24 1.82 .19** .08** -.09** -.03 [.14, .24] [.03, .13] [-.14, - .04] [-.08, .03] 6. Digital media literacy 3.99 0.83 .02 .05 .03 .14** -.06* [-.03,...

    work page 2014