Observational Aggression, Positive Reinforcement Training, and Disgust-Based Phobias – Comprehensive Study Notes

Bandura, Ross & Ross (1961) – “Transmission of Aggression Through Imitation of Aggressive Models”

Historical Context & Purpose

  • Built on prior incidental-learning & modelling work (e.g., Bandura & Huston 1961; Miller & Dollard 1941)
  • Key issue: Will children reproduce aggression in new settings without the model present?
  • Additional aims:
    • Examine inhibitory effect of witnessing subdued/non-aggressive models
    • Test Sex of Model × Sex of Child predictions
    • Explore aggression as a “masculine-typed” behaviour

Hypotheses

  • Children observing aggressive models will later show more imitative aggression vs. non-aggressive & control groups
  • Observation of subdued models will generalise inhibition, yielding less aggression than controls
  • Children will imitate same-sex models more; boys > girls on aggressive imitation, especially after viewing a male model

Participants & Setting

  • N = 72 (36 boys, 36 girls); age M = 52 months (range 37–69)
  • Recruited from Stanford University Nursery School
  • Two adult models (1 ♂ , 1 ♀); one female experimenter ran all sessions

Experimental Design

  • 8 experimental cells (2 Behavioural Conditions × 2 Model Sex × 2 Child Sex) + Control Group (n=24)
    • Behavioural conditions: Aggressive vs Non-Aggressive (Subdued)
    • Half of each experimental group saw same-sex, half opposite-sex model
  • Matched-triplet assignment on baseline nursery-school aggression (4 five-point scales → composite, r=.89 inter-rater)

Procedure Overview

  1. Exposure Room (10 min)

    • Child engaged with attractive potato-print & sticker activity (high intrinsic interest)
    • Opposite corner: Model + tinker toys, mallet, 5 ft inflated Bobo doll
    • Aggressive model: 1 min toy play → ~9 min scripted novel aggression:
      • Physically: lay doll, sit & punch nose; strike with mallet; toss; kick; repeat ×3
      • Verbal: “Sock him in the nose”, “Pow”, etc.; plus 2 neutral comments
    • Non-aggressive model: quiet tinker-toy assembly; ignored Bobo doll
    • No instructions to observe/imitate → ensures observational (covert) learning only

  2. Aggression Arousal (Frustration Phase)

    • Short play with high-value toys (fire engine, doll-set, spinning top, etc.)
    • After ~2 min toys removed (“reserved for other children”)
    • Purpose:
      • Equalise instigation; counter potential catharsis in Aggressive-Model group
      • Provide test of inhibition in Non-Aggressive group

  3. Test Room (20 min, 240 × 5-s intervals)

    • Mixed aggressive & non-aggressive toys (3 ft Bobo, mallet, dart guns, tether ball; tea set, crayons, bears, cars…)
    • Behaviour scored through one-way mirror by naïve observer (except when scorer had been model) → reliability r\approx .90

Behavioural Categories Scored

  • Direct Imitation
    • Physical Aggression (mallet hits to Bobo, sit-punch, kick, toss)
    • Verbal Aggression (exact phrases)
    • Non-Aggressive Verbal (“He keeps coming back…”)
  • Partial Imitation
    • Mallet aggression to other objects
    • Sit on Bobo without further attack
  • Non-Imitative Aggression
    • Punch Bobo, other physical & verbal hostility, aggressive gun play
  • Other: Non-aggressive play, quiet sitting

Major Results (Key Statistics)

  • Imitative Physical Aggression: Friedman ANOVA significant; Aggressive-Model group ≫ Non-Aggressive ≈ Control
  • Imitative Verbal Aggression: same pattern; Q significant at p<.001
  • Partial Imitation (Mallet): \chi^2 significant; Aggressive & Control > Non-Aggressive (girls particularly)
  • Sit-on-Bobo: Aggressive > Non-Aggressive (p=.018) & > Control (p=.059)
  • No-Imitative Physical & Verbal Aggression: Aggressive > Non-Aggressive (\chi^2_r = 8.96, p<.02)
  • Gun play & Punch-Bobo: no treatment effect
Sex-Related Findings
  • Boys > Girls imitative physical aggression after aggressive model (t=2.50, p<.01); no sex difference for verbal
  • Model Sex Interaction:
    • Boys showed more imitation after male aggressive model (physical t=2.07, p<.05; verbal t=2.51, p<.05; gun play t=2.12)
    • Girls imitated female model for verbal aggression but ns.
    • Non-aggressive male model produced marked inhibitory effect vs controls on multiple indices (sign-test p=.002 to .07)
Non-Aggressive Play Patterns
  • Sex effects: Girls spent more units with dolls, tea set, colouring (p<.05 – .001); Boys ↑ gun exploration (p<.01)
  • Treatment effects: Non-Aggressive group ↑ doll play and ↑ quiet sitting compared with Aggressive & Control groups

Discussion & Theoretical Implications

  • Demonstrates pure observational learning of novel aggression (no rehearsal, no external reinforcement)
  • Challenges Miller–Dollard view (discriminative stimuli + reinforcement) & Skinnerian shaping; shows vicarious acquisition
  • Aggressive models may lower inhibition and supply specific response scripts
  • Evidence of sex-typing: aggression more masculine; male models more potent, especially for physical acts
  • Findings question psychoanalytic “identification-with-the-aggressor” as sole mechanism; neutral models sufficed
  • Suggest future work on model attributes (fearful, nurturant, neutral) and on broader moral/ethical impacts

Fagen, Acharya & Kaufman (2014) – “Positive Reinforcement Training for a Trunk Wash in Nepal’s Working Elephants”

Rationale & Welfare Context

  • Free-contact Nepalese elephants trained traditionally with kocha (sharp stick), negative reinforcement & punishment
  • Global zoo trend: Protected-contact + Secondary Positive Reinforcement (SPR) → welfare & keeper-safety gains
  • Regular tuberculosis (TB) monitoring requires cooperative trunk-wash sample; traditional methods unreliable / stressful

Objectives

  • Test feasibility & efficiency of teaching traditionally trained, free-contact elephants to voluntarily perform trunk wash using SPR (clicker/whistle + food) only

Subjects & Husbandry

  • n=5 captive Asian females at Chitwan stable: 4 juveniles (5–7 y), 1 adult (~50 y)
  • Daily routine: jungle grazing, leg-chained stakes; no changes except water offered before sessions

Training Methodology

  • Primary reinforcer: chopped banana; Secondary: brief whistle (“click”)
  • Sessions: 7:30–10:00 am & 4:00–7:00 pm; mean \approx 12 min; max gap ≤2 days
  • Foundational operant tools:
    • Capture (reinforce spontaneous desirable act)
    • Lure (guide with treat)
    • Shaping (successive approximations)
  • Core learned behaviours (with bespoke Nepali-neutral cues):
    1. Trunk-Here – tip gently in trainer’s palm
    2. Trunk-Up – lift & hold fluid
    3. Bucket – place tip inside bucket
    4. Blow – strong exhale; later → Blow-in-Bucket
    5. Steady – maintain any prior position
    6. Syringe Desensitisation – tolerate 1–60 mL saline/water instillation
  • Additional (later dropped) tasks: targeting, trunk-down/out (control)
  • Chaining: assembled into full sequence: Here → Steady → Up (hold) → Bucket → Blow

Data Collection & Metrics

  • Session length (min); Offers (# of cues per task)
  • Periodic Performance Tests (≈ every 5 sessions) – 10 offers per task; pass ≥80 % correct
  • Relative task difficulty = mean offers needed pre-first pass
  • Statistical tests: one-way ANOVA across tasks

Key Findings

  • Juveniles (4/5) achieved full trunk wash within ≤35 sessions
    • Elephant 2 fastest: 25 sessions, \bar{t}=10.29 min (≈257 min total)
    • Mean total training time (successful): \approx 367 min (≃ 6 h)
  • Adult elephant (Elephant 5) did not complete within study window; distractions (nearby calf), possible visual impairment, foot abscess & age implicated
  • Performance Growth: Group mean success from 39\% (after 10 sessions) → 89.3\% (after 35)
  • Relative Difficulty (avg offers pre-pass):
    • Trunk-Here \approx 295\pm62 (hardest)
    • Trunk-Up \approx 166\pm25
    • Desensitisation \approx 108
    • Bucket \approx 61; Blow-in-Bucket \approx 54 (easiest)
    • ANOVA significant p = .017

Practical & Ethical Implications

  • Demonstrates efficient, low-stress alternative to aversive kocha training
  • Voluntary compliance aids regular TB surveillance; reduces need for sedation; improves keeper safety
  • SPR fosters animal choice, control & welfare; aligns with modern zoo standards
  • Future research: larger samples, males, cross-cultural settings, welfare impact metrics (stress hormones, etc.)

Saavedra & Silverman (2002) – “Disgust, Evaluative Learning & Childhood Specific Phobias”

Theoretical Background

  • Disgust labelled the “forgotten emotion of psychiatry” (Phillips et al., 1998)
  • Two distinct classical-conditioning routes:
    1. Expectancy Learning – CS predicts aversive event → fear
    2. Evaluative Learning – CS acquires negative valence independent of expectancy → disgust
  • Interaction may heighten avoidance in phobias (Woody & Teachman 2000)

Case Overview: 9-yr-old Boy with Button Phobia

  • Onset (age 5): bowl of buttons spilled on him during kindergarten art project
  • DSM-IV Specific Phobia (Buttons) confirmed via ADIS-C/P; 4-year duration; no OCD features; no trauma history
  • Impairment: clothing, dressing, school concentration, social contact

Treatment Phases & Observations

  1. In Vivo Graduated Exposures (4 sessions)
    • Hierarchy: jean buttons (large → small) → plastic buttons (large-coloured → small-clear) → hugging mother wearing buttons
    • Behavioural approach success but Subjective Units of Distress (SUDs) ↑ across sessions (from 2→8), indicating persistent disgust
  2. Shift to Disgust-Targeted Imagery & Cognitive Restructuring (7 sessions)
    • Child prompted to imagine buttons falling on body, describe look/feel/smell (“gross”, “emit bad odour”)
    • Included “hundreds of buttons” scenario & hugging mother with “shirt full of buttons”
    • SUDs during imagery dropped: e.g., 8→5→3 within session (Fig 2); maintained across repeat sessions (Fig 3)

Outcomes

  • Post-Treatment & 6- & 12-Month Follow-up: No DSM-IV phobia; minimal distress; wore uniform with clear plastic buttons daily

Implications for Clinical Practice

  • Standard exposure may reduce avoidance yet leave disgust evaluation intact → risk of relapse or subjective distress
  • Incorporating disgust imagery, cognitive re-evaluation, evaluative conditioning techniques can complete symptom remission
  • Need randomised studies to isolate disgust-targeting components; identify phobia subtypes where disgust is primary driver (e.g., BII, contamination, certain animal phobias)

Cross-Lecture Connections & Broader Themes

  • Observational vs Operant vs Evaluative Learning: Bandura shows acquisition of aggressive scripts without reinforcement; Fagen applies operant SPR for husbandry; Saavedra illustrates evaluative conditioning in emotional disorders.
  • Ethics & Welfare: Bandura cautions against media aggression; Fagen advocates humane training; Saavedra highlights comprehensive emotion-focused therapy.
  • Methodological Parallels: All studies emphasise precise behavioural coding, reliability, and matched controls; statistical tests (e.g., t, \chi^2, Q, ANOVA) underpin validity.