Extraction report — Shurakov unpublished Experiment 3

Source: papers/shurakov2023trainflightretraction/shurakov2023trainflightretraction.yaml · Generated: 2026-05-12 19:23 UTC
1 studies2 effects

Paper

paper_idshurakov2023trainflightretraction
short_labelShurakov unpublished Experiment 3
citationShurakov, N. (unpublished). Train and Flight scenarios from an earlier draft of The Stakes Effect: New Evidence from a Retraction-Based Experimental Design.
doi
year2023
publishedNo
publication_languageEnglish
publication_language_other
research_objectiveTest whether the retraction-based stakes effect observed in Shurakov's Bank experiments generalizes to two additional third-person scenarios: Train and Flight.
data_available_onlineNo
data_url
notesRelated published article: Shurakov, N. (2025). The Stakes Effect: New Evidence from a Retraction-Based Experimental Design. Episteme. https://doi.org/10.1017/epi.2025.10060. The Train/Flight experiment was not included in the published version.

Experiment 3 (unpublished Train/Flight extension)

study_id: 1

Study

study_id1
labelExperiment 3 (unpublished Train/Flight extension)
objectiveAssess whether the stakes effect from the retraction-based Bank experiments generalizes to two additional third-person scenarios using the modified design with an initial knowledge-ascription screen.
study_languageEnglish
study_language_other
designBetween-Subjects
design_other3 x 2 between-subjects design: story type (Neutral, Stakes, Evidence) by scenario (Train, Flight).
manipulated_factorsstory type: Neutral vs Stakes vs Evidence; scenario: Train vs Flight
paradigmRetraction of knowledge attribution
paradigm_other
notesThis unpublished study is Experiment 3 from an earlier draft of the published Shurakov paper. The focal effects are the two Stakes-vs-Neutral contrasts; Evidence is retained in the QMD as a manipulation/defeater comparison but is not extracted as a stakes effect.

Sample

n_final338
recruitmentProlific
recruitment_other
compensationmoney
compensation_otherGBP 0.40 for approximately 3 minutes.
characteristicsDraft reports 361 native English speakers from the US, UK, and Australia via Prolific, 181 female, mean age 41. Recomputing from the supplied Finished and Unfinished Qualtrics exports yields 338 usable composite-score responses after consent, initial knowledge endorsement, attention-check, and usable follow-up response filters (Train=161, Flight=177).
mean_age41
Provenance
page
table_refout/fulltext.md
tei_id
I recruited 361 native speakers of English from the US, UK, and Australia via Prolific.

Scale

labelcomposite-score
points
anchorsBinary retraction response ('I do'=1, 'I don't'=-1) multiplied by confidence from 1 to 7, yielding scores from -7 to 7.
directionHigher = more confident standing by the initial knowledge attribution; lower/negative = more confident retraction.
Provenance
page
table_refout/fulltext.md
tei_id
Composite scores were calculated by multiplying the retraction response by the participant's confidence level... resulting in scores between -7 and 7.

Measures

knowledge_question_textInitial knowledge-ascription question followed by a stand-by/retraction question: 'do you stand by your previous claim that [the target] knows...?'
knowledge_question_firstYes
additional_question_textConfidence rating after the binary retraction response, using a 7-point confidence item.

Scenarios

Scenarios (2)
train · Third-person Train scenario: a stranger with a printed timetable says the train stops at Kensington.
scenario_codetrain
scenario_typeThird-person Train scenario: a stranger with a printed timetable says the train stops at Kensington.
High stakes text
STAKES: the partner needs the participant to reach Kensington quickly to buy medicine after a shellfish reaction and may need resuscitation.
Low stakes text
NEUTRAL: the partner mentions a small party next week and asks whether the participant would like to join.
Provenance
page
table_refout/fulltext.md; data/Finished_Responses.csv
tei_id
You are at the railway station and want to take a train to Kensington... The guy takes out a printed train timetable.
flight · Third-person Flight scenario: a woman in an airport queue says a direct flight to Tokyo takes 14 hours.
scenario_codeflight
scenario_typeThird-person Flight scenario: a woman in an airport queue says a direct flight to Tokyo takes 14 hours.
High stakes text
STAKES: a close friend wants to plan a Star Wars movie marathon during the flight and says it is very important that the 14-hour claim is right.
Low stakes text
NEUTRAL: an acquaintance on the same flight clearly does not care how long the flight takes and asks whether the participant stands by the knowledge claim.
Provenance
page
table_refout/fulltext.md; data/Finished_Responses.csv
tei_id
You are at the airport to take a direct flight to Tokyo... She says: 'I've been to Japan a couple of times; it takes 14 hours to get to Tokyo if you have a direct flight'.

Effects

s1_e1 · Train (retraction-based): Stakes vs Neutral · Between-Subjects · d=0.150717614261 · v=0.037492591641

Effect

effect_ids1_e1
subgroupTrain (retraction-based): Stakes vs Neutral
subgroup_descComposite retraction score in the Train scenario, high-stakes follow-up vs neutral follow-up, after excluding initial knowledge non-endorsers.
designBetween-Subjects
design_other
quality_flagsunpublished_author_dataset; manuscript_dataset_n_discrepancy
notesThe manuscript reports the same rounded Neutral and Stakes descriptives for Train when the Finished and Unfinished Qualtrics exports are combined.

Effect Size

metricSMD
d0.150717614261
v0.037492591641
computed_fromgroups
needs_reviewfalse
notesComputed from author-sent Finished and Unfinished Qualtrics exports in analysis/effect_sizes.qmd using esc::esc_mean_sd; sign follows d = mean(low) - mean(high).

Moderators

scenarioother
skeptical_pressureNo
awarenessNo
evidenceExternal
attribution_personOther
evidence_reliabilityHigh

Contrast

group_hightrain_stakes
group_lowtrain_neutral
sign_conventiond = mean(low) - mean(high)
other_notesEffect uses Stakes vs Neutral; Evidence is a defeater condition and is not extracted as the primary stakes effect.

Moderator Coding

moderatorvaluereasonevidence
scenariootherThe vignette is a railway/train case, which is outside the fixed scenario vocabulary.
Provenance
page
table_refout/fulltext.md
tei_id
You are at the railway station and want to take a train to Kensington.
skeptical_pressureNoThe Stakes condition raises practical consequences but does not introduce an explicit doubt or alternative about whether the train stops at Kensington.
Provenance
page
table_refdata/Finished_Responses.csv
tei_id
STAKES: Could you buy Drugpills for me?... I may need resuscitation if you don't arrive in an hour.
awarenessNoThe target knowledge subject is the guy at the railway station. The high-stakes consequence is conveyed to the participant-attributor by the partner after the guy's timetable-based claim; the vignette does not state that the guy is aware of those stakes.
Provenance
page
table_refdata/Finished_Responses.csv
tei_id
The guy takes out a printed train timetable, and says 'Yeah, it stops at Kensington'. STAKES follow-up: your partner asks whether you could buy Drugpills.
evidenceExternalThe knowledge attribution rests on testimony supported by an external timetable.
Provenance
page
table_refout/fulltext.md
tei_id
The guy takes out a printed train timetable, and says 'Yeah, it stops at Kensington'.
attribution_personOtherParticipants attribute knowledge to the guy, not to themselves.
Provenance
page
table_refdata/Finished_Responses.csv
tei_id
The guy knows that the train stops at Kensington.
evidence_reliabilityHighThe explicit basis includes a printed timetable, an external record presented without unreliability cues, so it is coded High.
Provenance
page
table_refout/fulltext.md
tei_id
The guy takes out a printed train timetable.

Groups

group_idlabelnmeansdseprovenance
train_neutralTrain / Neutral545.333333333333332.17186121381535
Provenance
page
table_refanalysis/effect_sizes.qmd
tei_id
Computed from Finished_Responses.csv and Unfinished_Responses.csv after consent, initial knowledge endorsement, attention-check, and usable response filters.
train_stakesTrain / Stakes534.943396226415092.95097828850303
Provenance
page
table_refanalysis/effect_sizes.qmd
tei_id
Computed from Finished_Responses.csv and Unfinished_Responses.csv after consent, initial knowledge endorsement, attention-check, and usable response filters.

Reported Test

testTukey HSD pairwise comparison
t
f
chi2
z
df1
df2
p0.841
reported_d
reported_r
notesDraft reports no significant Stakes-vs-Neutral composite-score difference for Train; omnibus ANOVA F(2,158)=54.61, p<.001 because Evidence differs strongly.
Provenance
page
table_refout/fulltext.md
tei_id
No statistically significant difference was found between STAKES and NEUTRAL (p = .841).

Quality Flags

unpublished_author_dataset; manuscript_dataset_n_discrepancy
s1_e2 · Flight (retraction-based): Stakes vs Neutral · Between-Subjects · d=0.181474047555 · v=0.034037850974

Effect

effect_ids1_e2
subgroupFlight (retraction-based): Stakes vs Neutral
subgroup_descComposite retraction score in the Flight scenario, high-stakes follow-up vs neutral follow-up, after excluding initial knowledge non-endorsers.
designBetween-Subjects
design_other
quality_flagsunpublished_author_dataset; manuscript_dataset_n_discrepancy
notesThe manuscript reports the same rounded Neutral and Stakes descriptives for Flight when the Finished and Unfinished Qualtrics exports are combined.

Effect Size

metricSMD
d0.181474047555
v0.034037850974
computed_fromgroups
needs_reviewfalse
notesComputed from author-sent Finished and Unfinished Qualtrics exports in analysis/effect_sizes.qmd using esc::esc_mean_sd; sign follows d = mean(low) - mean(high).

Moderators

scenarioairport
skeptical_pressureNo
awarenessNo
evidenceFirst Person
attribution_personOther
evidence_reliabilityMedium

Contrast

group_highflight_stakes
group_lowflight_neutral
sign_conventiond = mean(low) - mean(high)
other_notesEffect uses Stakes vs Neutral; Evidence is a defeater condition and is not extracted as the primary stakes effect.

Moderator Coding

moderatorvaluereasonevidence
scenarioairportThe vignette is an airport/flight case.
Provenance
page
table_refout/fulltext.md
tei_id
You are at the airport to take a direct flight to Tokyo.
skeptical_pressureNoThe Stakes condition raises social/practical consequences but does not introduce evidence against the 14-hour claim.
Provenance
page
table_refdata/Finished_Responses.csv
tei_id
It is very important to me, and I would feel betrayed if you fooled me.
awarenessNoThe target knowledge subject is the woman in the queue. The high-stakes consequence is conveyed to the participant-attributor by the friend after the woman's travel-experience claim; the vignette does not state that the woman is aware of those stakes.
Provenance
page
table_refdata/Finished_Responses.csv
tei_id
She says: 'I've been to Japan a couple of times; it takes 14 hours to get to Tokyo if you have a direct flight'. STAKES follow-up: your friend says it is very important.
evidenceFirst PersonThe target knowledge subject's basis is her own prior travel experience.
Provenance
page
table_refout/fulltext.md
tei_id
She says: 'I've been to Japan a couple of times; it takes 14 hours to get to Tokyo if you have a direct flight'.
attribution_personOtherParticipants attribute knowledge to the woman, not to themselves.
Provenance
page
table_refdata/Finished_Responses.csv
tei_id
The woman knows that the flight takes 14 hours.
evidence_reliabilityMediumThe target knowledge subject's basis is ordinary first-person travel experience without official or independent verification, so it is coded Medium.
Provenance
page
table_refout/fulltext.md
tei_id
I've been to Japan a couple of times; it takes 14 hours to get to Tokyo if you have a direct flight.

Groups

group_idlabelnmeansdseprovenance
flight_neutralFlight / Neutral594.915254237288142.20726180495842
Provenance
page
table_refanalysis/effect_sizes.qmd
tei_id
Computed from Finished_Responses.csv and Unfinished_Responses.csv after consent, initial knowledge endorsement, attention-check, and usable response filters.
flight_stakesFlight / Stakes594.423728813559323.13051383812741
Provenance
page
table_refanalysis/effect_sizes.qmd
tei_id
Computed from Finished_Responses.csv and Unfinished_Responses.csv after consent, initial knowledge endorsement, attention-check, and usable response filters.

Reported Test

testTukey HSD pairwise comparison
t
f
chi2
z
df1
df2
p0.645
reported_d
reported_r
notesDraft reports no significant Stakes-vs-Neutral composite-score difference for Flight; omnibus ANOVA F(2,173)=123.36, p<.001 because Evidence differs strongly.
Provenance
page
table_refout/fulltext.md
tei_id
No statistically significant difference was found between STAKES and NEUTRAL (p = .645).

Quality Flags

unpublished_author_dataset; manuscript_dataset_n_discrepancy
Raw YAML
schema_version: "1.2"
extraction_run:
  created_at: "2026-05-06T00:00:00+02:00"
  created_by: Codex
  model: gpt-5
  source_files:
    - papers/shurakov2023trainflightretraction/source/RetractionBased_PaperDraft_18_11_N.docx
    - papers/shurakov2023trainflightretraction/source/email.eml
    - papers/shurakov2023trainflightretraction/data/Finished_Responses.csv
    - papers/shurakov2023trainflightretraction/data/Unfinished_Responses.csv
    - papers/shurakov2023trainflightretraction/data/Retraction_Third_person_April_14_2023.csv
    - papers/shurakov2023trainflightretraction/data/Retraction_experiment.py
  notes: "Unpublished Experiment 3 removed from an earlier draft of Shurakov 2025; source materials supplied by the author by email."
paper:
  paper_id: shurakov2023trainflightretraction
  citation: "Shurakov, N. (unpublished). Train and Flight scenarios from an earlier draft of The Stakes Effect: New Evidence from a Retraction-Based Experimental Design."
  short_label: "Shurakov unpublished Experiment 3"
  doi: null
  published: "No"
  year: 2023
  language: English
  language_other: null
  research_objective: "Test whether the retraction-based stakes effect observed in Shurakov's Bank experiments generalizes to two additional third-person scenarios: Train and Flight."
  data_availability:
    data_available_online: "No"
    url: null
    notes: "Dataset, draft manuscript, and email were supplied privately by Nikolai Shurakov; the email also linked to the author's Dropbox copy of the Python analysis script."
  notes: "Related published article: Shurakov, N. (2025). The Stakes Effect: New Evidence from a Retraction-Based Experimental Design. Episteme. https://doi.org/10.1017/epi.2025.10060. The Train/Flight experiment was not included in the published version."
studies:
  - study_id: 1
    label: "Experiment 3 (unpublished Train/Flight extension)"
    language: English
    language_other: null
    objective: "Assess whether the stakes effect from the retraction-based Bank experiments generalizes to two additional third-person scenarios using the modified design with an initial knowledge-ascription screen."
    sample:
      n_final: 338
      recruitment: Prolific
      recruitment_other: null
      compensation: money
      compensation_other: "GBP 0.40 for approximately 3 minutes."
      characteristics: "Draft reports 361 native English speakers from the US, UK, and Australia via Prolific, 181 female, mean age 41. Recomputing from the supplied Finished and Unfinished Qualtrics exports yields 338 usable composite-score responses after consent, initial knowledge endorsement, attention-check, and usable follow-up response filters (Train=161, Flight=177)."
      mean_age: 41
      mean_age_prov:
        page: null
        quote: "For the final analysis, I included all valid responses per condition, resulting in a total sample of 361 participants (181 female; mean age 41 years)."
        tei_id: null
        table_ref: out/fulltext.md
      provenance:
        page: null
        quote: "I recruited 361 native speakers of English from the US, UK, and Australia via Prolific."
        tei_id: null
        table_ref: out/fulltext.md
    design: Between-Subjects
    design_other: "3 x 2 between-subjects design: story type (Neutral, Stakes, Evidence) by scenario (Train, Flight)."
    manipulated_factors:
      - "story type: Neutral vs Stakes vs Evidence"
      - "scenario: Train vs Flight"
    paradigm: Retraction of knowledge attribution
    paradigm_other: null
    scale:
      label: composite-score
      points: null
      anchors: "Binary retraction response ('I do'=1, 'I don't'=-1) multiplied by confidence from 1 to 7, yielding scores from -7 to 7."
      direction: "Higher = more confident standing by the initial knowledge attribution; lower/negative = more confident retraction."
      provenance:
        page: null
        quote: "Composite scores were calculated by multiplying the retraction response by the participant's confidence level... resulting in scores between -7 and 7."
        tei_id: null
        table_ref: out/fulltext.md
    measures:
      knowledge_question_text: "Initial knowledge-ascription question followed by a stand-by/retraction question: 'do you stand by your previous claim that [the target] knows...?'"
      knowledge_question_first: "Yes"
      additional_question_text: "Confidence rating after the binary retraction response, using a 7-point confidence item."
    scenarios:
      - scenario_code: train
        scenario_type: "Third-person Train scenario: a stranger with a printed timetable says the train stops at Kensington."
        high_stakes_text: "STAKES: the partner needs the participant to reach Kensington quickly to buy medicine after a shellfish reaction and may need resuscitation."
        low_stakes_text: "NEUTRAL: the partner mentions a small party next week and asks whether the participant would like to join."
        provenance:
          page: null
          quote: "You are at the railway station and want to take a train to Kensington... The guy takes out a printed train timetable."
          tei_id: null
          table_ref: out/fulltext.md; data/Finished_Responses.csv
      - scenario_code: flight
        scenario_type: "Third-person Flight scenario: a woman in an airport queue says a direct flight to Tokyo takes 14 hours."
        high_stakes_text: "STAKES: a close friend wants to plan a Star Wars movie marathon during the flight and says it is very important that the 14-hour claim is right."
        low_stakes_text: "NEUTRAL: an acquaintance on the same flight clearly does not care how long the flight takes and asks whether the participant stands by the knowledge claim."
        provenance:
          page: null
          quote: "You are at the airport to take a direct flight to Tokyo... She says: 'I've been to Japan a couple of times; it takes 14 hours to get to Tokyo if you have a direct flight'."
          tei_id: null
          table_ref: out/fulltext.md; data/Finished_Responses.csv
    effects:
      - effect_id: s1_e1
        subgroup: "Train (retraction-based): Stakes vs Neutral"
        subgroup_desc: "Composite retraction score in the Train scenario, high-stakes follow-up vs neutral follow-up, after excluding initial knowledge non-endorsers."
        design: Between-Subjects
        design_other: null
        paradigm: Retraction of knowledge attribution
        paradigm_other: null
        moderators:
          scenario: other
          skeptical_pressure: "No"
          awareness: "No"
          evidence: External
          attribution_person: Other
          evidence_reliability: High
        moderators_coding:
          scenario:
            provenance:
              page: null
              quote: "You are at the railway station and want to take a train to Kensington."
              tei_id: null
              table_ref: out/fulltext.md
            reason: "The vignette is a railway/train case, which is outside the fixed scenario vocabulary."
          skeptical_pressure:
            provenance:
              page: null
              quote: "STAKES: Could you buy Drugpills for me?... I may need resuscitation if you don't arrive in an hour."
              tei_id: null
              table_ref: data/Finished_Responses.csv
            reason: "The Stakes condition raises practical consequences but does not introduce an explicit doubt or alternative about whether the train stops at Kensington."
          awareness:
            provenance:
              page: null
              quote: "The guy takes out a printed train timetable, and says 'Yeah, it stops at Kensington'. STAKES follow-up: your partner asks whether you could buy Drugpills."
              tei_id: null
              table_ref: data/Finished_Responses.csv
            reason: "The target knowledge subject is the guy at the railway station. The high-stakes consequence is conveyed to the participant-attributor by the partner after the guy's timetable-based claim; the vignette does not state that the guy is aware of those stakes."
          evidence:
            provenance:
              page: null
              quote: "The guy takes out a printed train timetable, and says 'Yeah, it stops at Kensington'."
              tei_id: null
              table_ref: out/fulltext.md
            reason: "The knowledge attribution rests on testimony supported by an external timetable."
          attribution_person:
            provenance:
              page: null
              quote: "The guy knows that the train stops at Kensington."
              tei_id: null
              table_ref: data/Finished_Responses.csv
            reason: "Participants attribute knowledge to the guy, not to themselves."
          evidence_reliability:
            provenance:
              page: null
              quote: "The guy takes out a printed train timetable."
              tei_id: null
              table_ref: out/fulltext.md
            reason: "The explicit basis includes a printed timetable, an external record presented without unreliability cues, so it is coded High."
        contrast:
          group_high: train_stakes
          group_low: train_neutral
          sign_convention: "d = mean(low) - mean(high)"
          other_notes: "Effect uses Stakes vs Neutral; Evidence is a defeater condition and is not extracted as the primary stakes effect."
        groups:
          - group_id: train_neutral
            label: "Train / Neutral"
            n: 54
            mean: 5.33333333333333
            sd: 2.17186121381535
            se: null
            provenance:
              page: null
              quote: "Computed from Finished_Responses.csv and Unfinished_Responses.csv after consent, initial knowledge endorsement, attention-check, and usable response filters."
              tei_id: null
              table_ref: analysis/effect_sizes.qmd
          - group_id: train_stakes
            label: "Train / Stakes"
            n: 53
            mean: 4.94339622641509
            sd: 2.95097828850303
            se: null
            provenance:
              page: null
              quote: "Computed from Finished_Responses.csv and Unfinished_Responses.csv after consent, initial knowledge endorsement, attention-check, and usable response filters."
              tei_id: null
              table_ref: analysis/effect_sizes.qmd
        reported_test:
          test: "Tukey HSD pairwise comparison"
          p: 0.841
          notes: "Draft reports no significant Stakes-vs-Neutral composite-score difference for Train; omnibus ANOVA F(2,158)=54.61, p<.001 because Evidence differs strongly."
          provenance:
            page: null
            quote: "No statistically significant difference was found between STAKES and NEUTRAL (p = .841)."
            tei_id: null
            table_ref: out/fulltext.md
        effect_size:
          metric: SMD
          d: 0.150717614261
          v: 0.037492591641
          computed_from: groups
          needs_review: false
          notes: "Computed from author-sent Finished and Unfinished Qualtrics exports in analysis/effect_sizes.qmd using esc::esc_mean_sd; sign follows d = mean(low) - mean(high)."
        quality_flags:
          - unpublished_author_dataset
          - manuscript_dataset_n_discrepancy
        notes: "The manuscript reports the same rounded Neutral and Stakes descriptives for Train when the Finished and Unfinished Qualtrics exports are combined."
      - effect_id: s1_e2
        subgroup: "Flight (retraction-based): Stakes vs Neutral"
        subgroup_desc: "Composite retraction score in the Flight scenario, high-stakes follow-up vs neutral follow-up, after excluding initial knowledge non-endorsers."
        design: Between-Subjects
        design_other: null
        paradigm: Retraction of knowledge attribution
        paradigm_other: null
        moderators:
          scenario: airport
          skeptical_pressure: "No"
          awareness: "No"
          evidence: First Person
          attribution_person: Other
          evidence_reliability: Medium
        moderators_coding:
          scenario:
            provenance:
              page: null
              quote: "You are at the airport to take a direct flight to Tokyo."
              tei_id: null
              table_ref: out/fulltext.md
            reason: "The vignette is an airport/flight case."
          skeptical_pressure:
            provenance:
              page: null
              quote: "It is very important to me, and I would feel betrayed if you fooled me."
              tei_id: null
              table_ref: data/Finished_Responses.csv
            reason: "The Stakes condition raises social/practical consequences but does not introduce evidence against the 14-hour claim."
          awareness:
            provenance:
              page: null
              quote: "She says: 'I've been to Japan a couple of times; it takes 14 hours to get to Tokyo if you have a direct flight'. STAKES follow-up: your friend says it is very important."
              tei_id: null
              table_ref: data/Finished_Responses.csv
            reason: "The target knowledge subject is the woman in the queue. The high-stakes consequence is conveyed to the participant-attributor by the friend after the woman's travel-experience claim; the vignette does not state that the woman is aware of those stakes."
          evidence:
            provenance:
              page: null
              quote: "She says: 'I've been to Japan a couple of times; it takes 14 hours to get to Tokyo if you have a direct flight'."
              tei_id: null
              table_ref: out/fulltext.md
            reason: "The target knowledge subject's basis is her own prior travel experience."
          attribution_person:
            provenance:
              page: null
              quote: "The woman knows that the flight takes 14 hours."
              tei_id: null
              table_ref: data/Finished_Responses.csv
            reason: "Participants attribute knowledge to the woman, not to themselves."
          evidence_reliability:
            provenance:
              page: null
              quote: "I've been to Japan a couple of times; it takes 14 hours to get to Tokyo if you have a direct flight."
              tei_id: null
              table_ref: out/fulltext.md
            reason: "The target knowledge subject's basis is ordinary first-person travel experience without official or independent verification, so it is coded Medium."
        contrast:
          group_high: flight_stakes
          group_low: flight_neutral
          sign_convention: "d = mean(low) - mean(high)"
          other_notes: "Effect uses Stakes vs Neutral; Evidence is a defeater condition and is not extracted as the primary stakes effect."
        groups:
          - group_id: flight_neutral
            label: "Flight / Neutral"
            n: 59
            mean: 4.91525423728814
            sd: 2.20726180495842
            se: null
            provenance:
              page: null
              quote: "Computed from Finished_Responses.csv and Unfinished_Responses.csv after consent, initial knowledge endorsement, attention-check, and usable response filters."
              tei_id: null
              table_ref: analysis/effect_sizes.qmd
          - group_id: flight_stakes
            label: "Flight / Stakes"
            n: 59
            mean: 4.42372881355932
            sd: 3.13051383812741
            se: null
            provenance:
              page: null
              quote: "Computed from Finished_Responses.csv and Unfinished_Responses.csv after consent, initial knowledge endorsement, attention-check, and usable response filters."
              tei_id: null
              table_ref: analysis/effect_sizes.qmd
        reported_test:
          test: "Tukey HSD pairwise comparison"
          p: 0.645
          notes: "Draft reports no significant Stakes-vs-Neutral composite-score difference for Flight; omnibus ANOVA F(2,173)=123.36, p<.001 because Evidence differs strongly."
          provenance:
            page: null
            quote: "No statistically significant difference was found between STAKES and NEUTRAL (p = .645)."
            tei_id: null
            table_ref: out/fulltext.md
        effect_size:
          metric: SMD
          d: 0.181474047555
          v: 0.034037850974
          computed_from: groups
          needs_review: false
          notes: "Computed from author-sent Finished and Unfinished Qualtrics exports in analysis/effect_sizes.qmd using esc::esc_mean_sd; sign follows d = mean(low) - mean(high)."
        quality_flags:
          - unpublished_author_dataset
          - manuscript_dataset_n_discrepancy
        notes: "The manuscript reports the same rounded Neutral and Stakes descriptives for Flight when the Finished and Unfinished Qualtrics exports are combined."
    notes: "This unpublished study is Experiment 3 from an earlier draft of the published Shurakov paper. The focal effects are the two Stakes-vs-Neutral contrasts; Evidence is retained in the QMD as a manipulation/defeater comparison but is not extracted as a stakes effect."