Psychological Effects of Insufficient Rewards — Detailed Study Notes

Key Concepts & Definitions

• Reward-based learning tradition:

  • Historically emphasized that presence of reward strengthens behavior.

  • Led to post-hoc assumptions that something must be rewarding whenever behavior persists (e.g., “removal of pain is rewarding”).

• Leon Festinger’s proposal:

  • Absence or inadequacy of reward can itself create predictable psychological consequences that explain persistence, valuation, and resistance to extinction of behavior.

• Cognition (as used by Festinger): any stored information that can be inferred from differential behavior under different stimulus conditions; applies to humans and animals.

• Cognitive Dissonance:

  • A state produced when two cognitions are in a non-fitting (dissonant) relation.

  • Example structure:

  • Cognition 1: “I exerted \text{Effort} / endured \text{Pain}.”

  • Cognition 2: “Outcome/reward was \text{Inadequate}.”

  • Reduction methods (when behavioral change is blocked):

  1. Re-evaluate effort/pain (“It wasn’t so bad.”)

  2. Re-evaluate goal (“The goal is actually more valuable.”)

• Insufficient Reward Effect: When individuals/animals voluntarily perform an act for little/no reward, they later value the act or its consequences more highly than those who received ample reward.

Foundational Human Experiments

• Preparatory Action & Probability (Yaryan & Festinger, unpublished)

  • Setup: High-school girls study symbol–definition list for a possible test.

  • Easy prep group: list available during test → minimal effort.

  • Effortful prep group: list absent → memorization required.

  • Both told selection for test already decided (50 % chance).

  • Result: Effortful group judged probability of actually taking test to be higher—illustrates reverse causality of effort → belief.

• Embarrassing Initiation Study (Aronson & Mills, 1959)

  • Women volunteers screened to join discussion group.

  • Mild test vs. severe, embarrassing test.

  • All hear identical, boring taped discussion.

  • Measures: Interest in discussion, liking of members.

  • Result: Severe-test subjects rated discussion & members significantly higher.

  • Interpretation: Painful initiation created dissonance (pain vs. low value) → inflated value to reduce dissonance.

Rat & Comparative Research

General Logic

• Rats experience cognitions (info about environment).

• Though “less intelligent,” they can still reduce dissonance, albeit with fewer mechanisms.

Partial Reward & Resistance to Extinction

• Observed law: behaviors reinforced intermittently resist extinction longer than those reinforced every trial.

• Festinger & Lawrence explanation:

  • Each non-rewarded trial introduces dissonance:
    \text{Effort} \; \rightarrow \; \text{No Food}.

  • Animal cannot simply stop (because some trials are rewarded) → reduces dissonance by developing an “extra preference” for the runway, goal box, or act.

  • During extinction (no food at all), partially-rewarded rat must now overcome both the disappearance of food and the acquired extra preference → slower extinction.

Delay of Reinforcement

• Choice test prediction: given immediate vs. delayed food, rats prefer immediate → delay is aversive.

• Non-choice training with enforced delays therefore should mimic partial reward effect.

• Empirical support:

  • Crum, Brown, & Bitterman (1951); Scott & Wike (1956); Wike & McNemara (1957); Fehrer (1956)

  • Longer delays or higher proportion of delayed trials → greater resistance to extinction.

Effort Manipulations

• Prediction: more work during acquisition (pressing stiff panel, swimming, long climbs) → stronger extra preference → higher resistance to extinction, provided extinction effort is held constant.

• Aiken (1957) confirms: high-effort training groups required more extinction trials than low-effort groups.

Operational Variables in Partial Reward

• Traditional view: ratio of reward (e.g., 33\%, 50\%) governs persistence.

• Festinger & Lawrence hypothesis: absolute number of non-rewarded trials is the critical variable (total dissonance magnitude).

Theios Study (Festinger, Lawrence, & Theios)

• 16 conditions manipulating:

  • Non-rewarded trials: 0,16,27,72.

  • Reward ratios: 33\%,50\%,67\%,100\%.

• Result summary (see figure in paper):

  • Holding non-reward count constant → negligible effect of ratio.

  • Increasing non-reward count → monotonic increase in extinction resistance.

“Pure” Non-Reward Experiments (Delay-Box Series)

Single Mid-Box Design

• Mid-box A in runway.

  • Group 100 %: Fed in mid-box and end-box.

  • Group 0 %: Delayed (no food) in mid-box; fed only in end-box.

• Extinction: only run Start → Mid-box (no food).

• Results:

  • 0 % group slower during acquisition, but extinguished slower than 100 % group; curves crossed after ~12–15 trials.

Satiation Replication

• Added 3-day free-feeding period; removed all food from apparatus.

• Outcome: pattern replicated; 0 % group still showed no extinction through 8 days (24 trials), indicating preference independent of hunger motivation.

Double Mid-Box (A & B) Counter-Preference Test

• Group A: Delayed in Box A, free through Box B.

• Group B: Free through Box A, delayed in Box B.

• Extinction path: Box A → Box B.

  • Thus Group A runs away from its delay site; Group B runs toward its delay site.

• While satiated:

  • Group B showed zero extinction across 30 trials; Group A extinguished normally.

• Interpretation: extra preference is localized to place where dissonance occurred (supports theoretical mechanism).

Mathematical & Graphical Insights

• Resistance to extinction (RTE) grows with cumulative dissonance D, where
  D \propto N{\text{nonreward}} \times I{d}
  N{\text{nonreward}} = count of unrewarded/effort/delay trials; I{d} = individual dissonance intensity per trial (function of effort, delay length, pain, etc.).

• Qualitative prediction: \frac{d\;\text{RTE}}{dD} > 0.

Theoretical Implications & Connections

• Challenges simple reinforcement accounts: non-rewarded effort can strengthen subsequent persistence.

• Aligns with broader cognitive-consistency frameworks in social psychology (e.g., post-decision dissonance, effort justification in fraternities, military hazing).

• Explains “sink cost” behaviors: people/animals value goals proportionally to sacrifices already made.

Ethical, Philosophical & Practical Considerations

• Highlighted “magical” thinking: effort shapes belief about likelihood or value, even when logically unrelated.

• Practical applications:

  • Education/training: mild challenges can increase course valuation but risk backfiring if too severe.

  • Organizational initiation rites: rigorous entry → stronger group cohesion.

  • Therapy & habit change: ensuring some dissonance without escape may consolidate new healthy preferences.

• Ethical red flags: Deliberately imposing pain/effort to induce attachment (e.g., hazing) raises welfare concerns.

Key References (chronological)

• Sheffield (1949); Crum, Brown, & Bitterman (1951); Festinger (1957); Weinstock (1954); Scott & Wike (1956); Lewis (1956); Fehrer (1956); Wike & McNemara (1957); Aiken (1957); Aronson & Mills (1959).

Study Tips & Exam Connections

• Be able to reproduce core dissonance equation: \text{Dissonance} = f(\text{Cognition}1,\text{Cognition}2) and list three reduction methods.

• Compare/contrast operant conditioning vs. cognitive dissonance accounts of persistence.

• Memorize classic paradigms: Severe Initiation, Partial Reward, Delay of Reinforcement.

• Practice deriving predictions: “If effort ↑ and reward = constant, what happens to valuation?”

• Connect to real-world scenarios (gym memberships, sunk-cost fallacy, recruitment rituals) for essay responses.

Key Concepts & Definitions

• Reward-based learning tradition:

  • Historically emphasized that presence of reward strengthens behavior.

  • Led to post-hoc assumptions that something must be rewarding whenever behavior persists (e.g., “removal of pain is rewarding”).

• Leon Festinger’s proposal:

  • Absence or inadequacy of reward can itself create predictable psychological consequences that explain persistence, valuation, and resistance to extinction of behavior.

• Cognition (as used by Festinger):

  • Any stored information that can be inferred from differential behavior under different stimulus conditions; applies to humans and animals.

• Cognitive Dissonance:

  • A state produced when two cognitions are in a non-fitting (dissonant) relation.

  • Example structure:

  • Cognition 1: “I exerted \text{Effort} / endured \text{Pain}.”

  • Cognition 2: “Outcome/reward was \text{Inadequate}.”

  • Reduction methods (when behavioral change is blocked):

  1. Re-evaluate effort/pain (“It wasn’t so bad.”)

  2. Re-evaluate goal (“The goal is actually more valuable.”)

• Insufficient Reward Effect:

  • When individuals/animals voluntarily perform an act for little/no reward, they later value the act or its consequences more highly than those who received ample reward.

Foundational Human Experiments

• Preparatory Action & Probability (Yaryan & Festinger, unpublished)

  • Setup: High-school girls study symbol–definition list for a possible test.

  • Easy prep group: list available during test → minimal effort.

  • Effortful prep group: list absent → memorization required.

  • Both told selection for test already decided (50\% chance).

  • Result: Effortful group judged probability of actually taking test to be higher—illustrates reverse causality of effort → belief.

• Embarrassing Initiation Study (Aronson & Mills, 1959)

  • Women volunteers screened to join discussion group.

  • Mild test vs. severe, embarrassing test.

  • All hear identical, boring taped discussion.

  • Measures: Interest in discussion, liking of members.

  • Result: Severe-test subjects rated discussion & members significantly higher.

  • Interpretation: Painful initiation created dissonance (pain vs. low value) → inflated value to reduce dissonance.

Rat & Comparative Research

General Logic

• Rats experience cognitions (info about environment).

• Though “less intelligent,” they can still reduce dissonance, albeit with fewer mechanisms.

Partial Reward & Resistance to Extinction

• Observed law: behaviors reinforced intermittently resist extinction longer than those reinforced every trial.

• Festinger & Lawrence explanation:

  • Each non-rewarded trial introduces dissonance:
    
    \text{Effort} \; \rightarrow \; \text{No Food}.

  • Animal cannot simply stop (because some trials are rewarded) → reduces dissonance by developing an “extra preference” for the runway, goal box, or act.

  • During extinction (no food at all), partially-rewarded rat must now overcome both the disappearance of food and the acquired extra preference → slower extinction.

Delay of Reinforcement

• Choice test prediction: given immediate vs. delayed food, rats prefer immediate → delay is aversive.

• Non-choice training with enforced delays therefore should mimic partial reward effect.

• Empirical support:

  • Crum, Brown, & Bitterman (1951); Scott & Wike (1956); Wike & McNemara (1957); Fehrer (1956)

  • Longer delays or higher proportion of delayed trials → greater resistance to extinction.

Effort Manipulations

• Prediction: more work during acquisition (pressing stiff panel, swimming, long climbs) → stronger extra preference → higher resistance to extinction, provided extinction effort is held constant.

• Aiken (1957) confirms: high-effort training groups required more extinction trials than low-effort groups.

Operational Variables in Partial Reward

• Traditional view: ratio of reward (e.g., 33\% , 50\% ) governs persistence.

• Festinger & Lawrence hypothesis: absolute number of non-rewarded trials is the critical variable (total dissonance magnitude).

Theios Study (Festinger, Lawrence, & Theios)

• 16 conditions manipulating:

  • Non-rewarded trials: 0,16,27,72.

  • Reward ratios: 33\%,50\%,67\%,100\%.

• Result summary (see figure in paper):

  • Holding non-reward count constant → negligible effect of ratio.

  • Increasing non-reward count → monotonic increase in extinction resistance.

“Pure” Non-Reward Experiments (Delay-Box Series)

Single Mid-Box Design

• Mid-box A in runway.

  • Group 100\%: Fed in mid-box and end-box.

  • Group 0\%: Delayed (no food) in mid-box; fed only in end-box.

• Extinction: only run Start → Mid-box (no food).

• Results:

  • 0\% group slower during acquisition, but extinguished slower than 100\% group; curves crossed after ~$12–15~$ trials.

Satiation Replication

• Added 3-day free-feeding period; removed all food from apparatus.

• Outcome: pattern replicated; 0\% group still showed no extinction through 8 days (24 trials), indicating preference independent of hunger motivation.

Double Mid-Box (A & B) Counter-Preference Test

• Group A: Delayed in Box A, free through Box B.

• Group B: Free through Box A, delayed in Box B.

• Extinction path: Box A → Box B.

  • Thus Group A runs away from its delay site; Group B runs toward its delay site.

• While satiated:

  • Group B showed zero extinction across 30 trials; Group A extinguished normally.

• Interpretation: extra preference is localized to place where dissonance occurred (supports theoretical mechanism).

Mathematical & Graphical Insights

• Resistance to extinction (RTE) grows with cumulative dissonance D, where
  
  D \propto N{\text{nonreward}} \times Id
  
  N{\text{nonreward}} = count of unrewarded/effort/delay trials; Id = individual dissonance intensity per trial (function of effort, delay length, pain, etc.).

• Qualitative prediction: \frac{d\;\text{RTE}}{dD} > 0.

Theoretical Implications & Connections

• Challenges simple reinforcement accounts: non
  -rewarded effort can strengthen subsequent persistence.

• Aligns with broader cognitive-consistency frameworks in social psychology (e.g., post-decision dissonance, effort justification in fraternities, military hazing).

• Explains “sink cost” behaviors: people/animals value goals proportionally to sacrifices already made.

Ethical, Philosophical & Practical Considerations

• Highlighted “magical” thinking: effort shapes belief about likelihood or value, even when logically unrelated.

• Practical applications:

  • Education/training: mild challenges can increase course valuation but risk backfiring if too severe.

  • Organizational initiation rites: rigorous entry → stronger group cohesion.

  • Therapy & habit change: ensuring some dissonance without escape may consolidate new healthy preferences.

• Ethical red flags: Deliberately imposing pain/effort to induce attachment (e.g., hazing) raises welfare concerns.

Key References (chronological)

• Sheffield (1949); Crum, Brown, & Bitterman (1951); Festinger (1957); Weinstock (1954); Scott & Wike (1956); Lewis (1956); Fehrer (1956); Wike & McNemara (1957); Aiken (1957); Aronson & Mills (1959).

Study Tips & Exam Connections

• Be able to reproduce core dissonance equation: \text{Dissonance} = f(\text{Cognition}1,\text{Cognition}2) and list three reduction methods.

• Compare/contrast operant conditioning vs. cognitive dissonance accounts of persistence.

• Memorize classic paradigms: Severe Initiation, Partial Reward, Delay of Reinforcement.

• Practice deriving predictions: “If effort ↑ and reward = constant, what happens to valuation?”

• Connect to real-world scenarios (gym memberships, sunk-cost fallacy, recruitment rituals) for essay responses.