Senior Theses: 1998-1999

Memory and the Misinformation Effect in Relation to Eyewitness Testimony
F. Coker Gamble & Hilary A. Wayne
Faculty advisor: Alliston K. Reid, Ph.D.

(Submitted for publication to Legal and Criminological Psychology)

In order to examine the effects that misinformation has on existing memory, 130 Wofford College students participated in a study arranged in a 4 x 3 factorial design. Participants were exposed to original information in either the audio, video, audio-video, or picture modality. Then some participants were exposed to affidavits that contained misinformation about either peripheral details or central details, while others were exposed to the correct information. A misinformation effect was observed with peripheral details but not with central details. The modality influenced the degree of the misinformation effect. The audio-video and video modalities produced the effect, but the audio and picture modalities did not show a significant effect. Therefore, memory that is visual and audio-visual can be manipulated by the presence of misinformation in affidavits.

Response Strength within Response Sequences
Heather Bratt, Aaron Bliley, Louise Buchholz, & Jennifer Jones
Faculty advisor: Alliston K. Reid, Ph.D.
Submitted for publication to the Journal of the Experimental Analysis of Behavior. 

Traditional accounts of learned response sequences, such as chaining theory and the delay of reinforcement gradient, assume that the strength of each response is determined by an operant contingency. Behavioral momentum theory proposes that response strength has two possible sources: one produced by operant contingencies and measured by relative response rates, and another produced by Pavlovian contingencies and measured by relative resistance to change. The theory has found support in multiple and chain schedules, but has not been studied in relation to response strength within simple response sequences. We first trained pigeons to complete a sequence composed of a single peck to either a red or white key followed by a single peck to the alternate key. A delay of three seconds between key pecks ensured a greater difference in the strengths of red and white. Subjects were subsequently shifted to a multiple schedule containing a component requiring this sequence, a blackout component, and other components with identical VI schedules for pecking red and for pecking white. After performance stabilized, resistance to change was measured by presenting response-independent food, by an extinction schedule, or both. Thus, the differences in strength of responding to red and white could be measured by relative response rates, relative resistance to change, or a combination of the two, reflecting operant or Pavlovian contingencies. The results show that there was no reliable difference in relative response rates or resistance to change for any of the subjects. This contradicts our prediction of higher response strength for the stimulus that is more contiguous with reinforcement.

Effects of Delay in Response Sequence on Choice
Jane Ferguson, Amanda Floyd, & Sara Longfellow
Faculty advisor: Alliston K. Reid, Ph.D. 

The purpose of this experiment is to determine the effect of delay in response sequence on choice. The subjects were four homing pigeons which were previous exposed to experimental conditioning. The training sessions consisted of four different conditions, varying in ratio of delay to reinforcement. Pigeons advanced to each condition based on their behavior. Currently, pigeons are still in the training phase. After training, one sequence will be maintained on a 1-second delay to reinforcement. In successive experimental conditions, the alternate sequence will be placed on a 3, 5, or 7 second delay to reinforcement. Current data are inconclusive, and further research is being conducted to complete the experiment and answer questions concerning its validity and preference.

The Discriminative Stimulus in Learned Response Sequences
Susan Kelly
Faculty advisor: Alliston K. Reid, Ph.D. 

The present study concentrated on alternative roles of a discriminative stimulus in a response sequence. Rats were exposed to two conditions involving a presentation of a tone at different locations. The subjects were required to respond correctly to a specific three-response sequence consisting of left and right lever presses. Delivery of reinforcement occurred after an accurately completed sequence of responses. Errors were measured to establish the effects of the tone which served as the discriminative stimulus. In the first condition, which consisted of a tone sounding after the first response, analysis of errors established that the subjects relied on the discriminative stimulus in difficult sequences to improve accuracy in both the second and third response. In a second condition, the tone was moved to the position following the second response during probe trials. Analysis of errors demonstrated that the subjects associated the presentation of the tone after the first response with the knowledge of their position in the sequence. These findings suggest the limitations of traditional chaining theory.

Functional Response Units on VI Schedules
Kevin L. Morton and Harriet Willimon
Faculty advisor: Alliston K. Reid, Ph.D.
These data have been published in the Journal of the Experimental Analysis of Behavior. 

Rats were exposed to a free-operant schedule in which sequences of two lever presses were reinforced on a concurrent variable interval schedule. Even lever presses, not odd, had the possibility of reinforcement. Reinforcement for heterogeneous sequences was three times that for homogeneous sequences. Analysis of matching and other analyses allowed the experimenters to determine that, with extended training, response sequences acted as functional response units that influence choice. Although there was a noticeable difference between response sequences, the rats were unable to determine that the highest rate of reinforcement was produced with heterogeneous sequences. Also, the interresponse times suggest that the rats were unable to determine that only even lever presses produced reward. the relative response rates and the relative reinforcement rates for this experiment did not support a matching account for complex functional response units.

The Development of Functional Response Units when Exposed to Non-Differential Reinforcement
Myila Dunham and Angela Miller
Faculty advisor: Alliston K. Reid, Ph.D.
These data were published in the Journal of the Experimental Analysis of Behavior. 

In order to study the development of functional response units, rats were exposed to a non-differential reinforcement schedule which reinforced all possible two-response sequences with equal probability. Homogeneous and heterogeneous sequences of left and right lever presses were reinforced with the same probability of 0.4. An imaginary coin was flipped after every even, but not odd, lever press to determine whether or not to reinforce the response sequence. Subjects strongly preferred homogeneous sequences over heterogeneous sequences, even though both had the same reinforcement probability. Analysis of interresponse times demonstrated that all subjects paused briefly after most two-response sequences, but the pauses were very short compared to those observed in studies that differentially reinforced heterogeneous sequences with higher probability than homogeneous sequences. We postulate that the strong tendency to persevere on the same lever, producing rapid bursts of responses, may have retarded the process of learning complex functional response units. 

The Development of Functional Response Units: A Control Condition
Amanda Schaekel and Jill Warren
Faculty advisor: Alliston K. Reid, Ph.D.
These data were published in the Journal of the Experimental Analysis of Behavior. 

In studying the factors that affect the acquisition of a functional response unit, subjects were exposed to a probabilistic free-operant schedule in a BAB design, to serve as a control group for the ABA design in the experiment by Reid and Chadwick. The computer flipped an imaginary coin after each even lever press. In the demarcated condition (Condition B), a 0.3 second tone sounded with each coin flip, but no tone was produced in the undemarcated condition (Condition A). Heterogeneous sequences (LR & RL) were reinforced with a greater probability than persisting on one lever (LL or RR). After 60 sessions, the data were examined for the effect that the presence of the tone had on the interresponse times, transition probabilities, and other measures. All analyses suggested that the demarcation of the response sequence supported the rapid acquisition of the behavioral unit and supported the findings of Reid and Chadwick.