Teaching Classical and Operant Conditioning in a Laboratory-Based Course: Eight Effective Experiments

Transcription

1 Teaching Classical and Operant Conditioning in a Laboratory-Based Course: Eight Effective Experiments Eric S. Murphy & Robert J. Madigan University of Alaska Anchorage We teach a junior-level experimental psychology course entitled Learning and Cognition. In the laboratory portion of the course, students conduct eight experiments on the topics of classical and operant conditioning. The classical conditioning experiments use Virtual Pavlov 2000 (Benedict, 2000), a commercially-available computer program that simulates classical conditioning in dogs and rats using the Rescorla-Wagner Theory (Rescorla & Wagner, 1972) to generate the data. The students conduct four experiments on classical conditioning: 1) acquisition and extinction, 2) conditioned stimulus (CS)-unconditioned stimulus (US) temporal arrangements/cs-us intensities, 3) the conditioned emotional response procedure, and 4) the blocking effect. The operant conditioning experiments use real laboratory rats to investigate conditioned reinforcement, shaping, stimulus control, and backwards chaining. At the conclusion of each experiment, students are required to analyze the data and submit their results in an APA-style lab report. We describe each experiment so that a laboratory instructor will be able to replicate the activity at their home institution. Based on teaching evaluations, the students enjoy the laboratory activities, particularly working with the rats in the operant conditioning experiments. Additionally, graduates from our program perform well (e.g., approximately 65 th percentile) on the Learning and Cognition subtest of the standardized PACAT (Psychology Area Concentration Achievement Test), a national test that assesses knowledge in psychology. We suspect that the laboratory activities have a lot to do with our students success. Keywords: Learning, classical conditioning, operant conditioning, laboratory activities We teach a junior-level course entitled Learning and Cognition. The purpose of the course is to survey the behavioral and cognitive processes by which organisms adapt to their environments. When students enroll in our course, they are also required to enroll in one of the three laboratory sections taught before or after the lecture. Laboratory work involves both computer exercises and work with laboratory animals. The lab provides an opportunity to observe phenomena relevant to the class under relatively controlled conditions. It also helps the students gain experience applying basic behavioral principles. Although the students sometimes approach the lab with trepidation, this soon disappears. It has been our observation that the laboratory experience contributes significantly to the course. The computer simulations are engaging, and the rat experiments increase the students appreciation for the use of animals in behavioral research. The lecture component provides a theoretical foundation for understanding basic learning principles along with practical examples of how these concepts are applied to human affairs (see Table 1). To complement the lecture topics, we have designed eight laboratory exercises to illustrate the key principles of classical and operant conditioning. These activities were designed for laboratory sections that meet twice a week for 75 minutes each during the first eight weeks of a sixteen week semester. The students are typically psychology majors who have taken sophomore-level courses in statistics and research methods. Each laboratory section has an enrollment cap of 16 students and the course is offered twice per year. The textbooks for the course are Terry (2009) and Miller (2006). Terry is used for the material on classical conditioning and cognition (covered in the final eight weeks of the semester). Miller is used exclusively for the operant conditioning material. We also provide a custommade laboratory manual that covers the laboratory exercises. The entire course is taught by one faculty member and each laboratory section has a teaching 120

2 Table 1. Lecture Topics on Classical and Operant Conditioning Week Lecture Topic Readings 1 Introduction to the Course Terry: 1-25 Pavlovian Conditioning: Basic Principles Terry: Pavlovian Conditioning: Examples Terry: & Pavlovian Conditioning: Applications Terry: The Nature of the Pavlovian Association Terry: & Theories of Pavlovian Conditioning Terry: Quiz 1 All of the above Introduction to Operant Conditioning Miller: Reinforcement and Punishment Miller: Shaping Miller: Stimulus Control Miller: Chaining and Conditioned Reinforcement Miller: Schedules of Reinforcement Miller: Theories of Reinforcement Miller: Observational Learning Miller: Exam 1 All of the above Note. The topics of classical and operant conditioning are covered during the first eight weeks of a sixteen week semester. The textbooks for the lecture are Terry (2009) and Miller (2006). assistant. The teaching assistants are advanced undergraduates who excelled in Learning and Cognition during a previous semester. The first four experiments investigate classical conditioning using a computer program called Virtual Pavlov 2000 (Benedict, 2000). The students conduct four experiments on classical conditioning: 1) acquisition and extinction, 2) conditioned stimulus (CS)-unconditioned stimulus (US) temporal arrangements/cs-us intensities, 3) the conditioned emotional response procedure, and 4) the blocking effect. The four operant conditioning experiments use real laboratory rats to investigate conditioned reinforcement, shaping, stimulus control, and backwards chaining. For each laboratory exercise, we describe the learning objectives, principles, materials, procedures, and laboratory report requirements. Classical Conditioning Laboratory Activities Classical conditioning is the most basic type of associative learning and is found at all levels of the phylogenetic chain. Therefore, it is crucial to cover classical conditioning in the laboratory exercises. We use four computer-simulated experiments to explore important aspects of classical conditioning. By working through the experiments, students gain familiarity with the technical terms of conditioning and improve their understanding of classical conditioning processes presented in the lecture portion of the course. The experiments also provide opportunities to apply principles of research design and to prepare written reports following the publication guidelines of the American Psychological Association (APA, 2010). The four experiments we assign should be considered minimum requirements. The computer simulations allow various features of classical conditioning to be easily investigated. Students are encouraged to use the program creatively to examine how classical conditioning works. An Introduction to Virtual Pavlov 2000 Because of the limited time dedicated to classical conditioning in our course, we use a computer program, rather than real animals, to investigate classical conditioning phenomena. If an instructor wanted to use real animals, we recommend consulting Abramson (1990). Abramson describes several simple experiments on invertebrates, including honeybees, planarians, and earthworms. We use Virtual Pavlov 2000 (Benedict, 2000), a program that simulates classical conditioning in dog and rat subjects. The computer model used by Virtual Pavlov 2000 is based on a theory of classical conditioning proposed by Rescorla and Wagner (1972). Therefore, the subjects simulated by Virtual Pavlov 2000 produce results similar to real animals under similar laboratory conditions. At the beginning of each simulation, the learning rate parameter is automatically set to a value chosen from a table of random numbers. Each time the simulation is run, the 121

3 results are somewhat different, just as they would be with different subjects. Virtual Pavlov 2000 (Benedict, 2000) has four simulations, including salivary conditioning, suppression ratio, fear conditioning, and taste aversion learning. The salivary conditioning simulator is modeled on Pavlov s (1927) great contributions to the study of learning and closely reproduces data obtained in his laboratory. The suppression ratio is a demonstration of the conditioned emotional response procedure (CER; Estes & Skinner, 1941). The simulation begins with a video showing the procedure with a real rat. Next, the students participate in a mock CER procedure and learn how to calculate the amount of fear they showed to a tone. The suppression ratio demonstration is important for understanding the mechanics of fear conditioning, a topic explored in Experiment 4. The fear conditioning simulator allows the user to pair various auditory and visual stimuli with a shock. We use it to replicate Kamin s (1969) blocking effect. Due to time constraints, we do not use the taste aversion simulator, so we will not be discussing it here. However, we believe it is ideally suited for a laboratory-based course that has more time to explore classical conditioning phenomena. Virtual Pavlov 2000 (Benedict, 2000) is distributed by Insight Media (New York, NY; and runs on versions of Windows 95 or a later. Each copy costs $ The full installation requires a minimum of 165 megabytes. If the videos are read from the compact disk, then only 30 megabytes is required. The videos use Quicktime version 4.0 or later. We have been using Virtual Pavlov 2000 since it was released. Thus far, we have not had any installation issues on our Windows-based computers. General Laboratory Preparation At the beginning of a classical conditioning exercise, we usually give a minute lecture on the topic to be investigated. Although the students have learned the material in the lecture-portion of the course, it is a good time to review the material and invite questions. Before the students collect data, we give a demonstration how to conduct the experiment using Virtual Pavlov 2000 (Benedict, 2000). Following the demonstration, the students join their lab partner at one of the 12 computer stations in our teaching laboratory. Each computer runs on a Windows XP platform. The main preparation for the classical conditioning laboratories is to ensure that the computers are working properly and that Virtual Pavlov 2000 is displayed on the desktop of each computer. We also advise our teaching assistants to conduct the experiment before each lab. This recent practice will ensure they are able to help guide the students through the exercise. Experiment 1: Acquisition and Extinction of a Classically Conditioned Response Learning objectives. Once students have completed this experiment, they should be able to demonstrate the following skills: a) define unconditioned stimulus (US), unconditioned response (UR), conditioned stimulus (CS), conditioned response (CR), acquisition, and extinction; b) conduct a classical conditioning experiment using the Virtual Pavlov 2000 (Benedict, 2000) software; c) replicate a basic classical conditioning experiment similar to those originally performed by Pavlov (1927); and d) summarize data on the acquisition and extinction of a classically conditioned response in an APA (2010) style figure and describe them in a laboratory report. Principle. There are four components of classical conditioning. Two involve behavior (responses), and the other two involve the environment (stimuli). The unconditioned stimulus (US) is a stimulus that naturally elicits a response. For example, food is a US for the unconditioned response (UR) of salivation. The conditioned stimulus (CS) is a stimulus, such as a metronome, that elicits a response after being paired with the US. After several CS-US pairings, the CS will begin to elicit a response of its own called the conditioned response (CR). Experiment 1 explores two fundamental processes in classical conditioning: acquisition and extinction. Acquisition is the initial stage of learning where the CR is established and strengthened with repeated pairings of the CS and US. Extinction is a procedure used to decrease the CR by repeatedly presenting the CS alone. In the computer simulation, students specify details about CS and US and the timing relations between them. The simulation demonstrates how these variables influence the acquisition and extinction of the CR. Materials. The materials for conducting the experiment are Virtual Pavlov 2000 (Benedict, 2000) and the Experiment 1 Worksheet (see Appendix). The worksheet contains spaces for the student to include all of the parameters of the experiment, including the CS and US intensities, CS-US temporal arrangement, the number of acquisition and extinction trials, the dependent variable (i.e., drops of 122

4 Figure 1. Drops of saliva (CR) as a function of 20 acquisition and 20 extinction trials in Experiment 1 (Acquisition & Extinction of a Classically Conditioned Response). The CS (metronome) intensity was 75 db during all trials. The CS-US temporal arrangement was a 3-s short delay and the US (food) intensity was 10 g during the acquisition phase of the experiment. The function represents the mean of five dogs simulated by Virtual Pavlov 2000 (Benedict, 2000). saliva on every fifth trial), and an area to write down the data for each subject. Procedure. We recommend students and instructors use the following procedure for conducting the experiment: When the student arrives to their computer station, the computer should be turned on and the program menu appearing on the desktop. The students should select the Salivary Conditioning subprogram from the menu and follow the instructions on the screen to set up the experiment. We encourage each student team to explore the program s options to get an understanding of how it works. Each lab team should run a few subjects to observe both conditioning and extinction. To do this, the students should specify the details of the experiment they wish to conduct. The first task is to set the acquisition trials which will specify the number of pairings between the CS (a metronome) and the US (food) to produce classical conditioning. The second task is to set the number of extinction trials in which the CS is presented without the US to study what occurs when the CS and US are no longer paired. The third task is for the students to specify the loudness of the metronome using values in the range from db. The amount of food, or US intensity should then be specified. Virtual Pavlov 2000 (Benedict, 2000) allows a range between 0 and 20 g of food to be presented with each CS. The final parameter to be specified is the CS-US interval. The number of seconds between the CS and US onset can be set between 0 (simultaneous) to 25 s (trace conditioning). Once the parameters are entered, the students can simulate a subject by clicking on the Run button. The drops of saliva for every fifth trial will be presented on the screen. To run another dog, the students will need press the Run button again. A great feature of Virtual Pavlov 2000 is a video of a dog experiencing a mock acquisition and extinction trial. We recommend students watch the video at least once to get an idea of the experimental procedures. When students are familiar with the program, we ask them to choose reasonable values for all parameters of the program and collect data for 20 acquisition and 20 extinction trials using five subjects. Laboratory report requirements. We recommend instructors use the following laboratory requirements: 1. The students should complete the experiment and turn in the Experiment 1 Worksheet. 2. The students should compute the average responses of the five dogs for each test trial during acquisition and extinction and plot the averaged data (not individual data) in a line graph. The figure and figure caption should conform to APA (2010) guidelines. Figure 1 presents drops of saliva as a function of 20 acquisition and 20 extinction trials. The CS (metronome) intensity was 75 db during all trials. The CS-US temporal arrangement was a 3-s short delay and the US (food) intensity was 10 g during the acquisition phase of the experiment. The function represents the mean of five dogs simulated by Virtual Pavlov 2000 (Benedict, 2000). 3. We also require the data to be described in a one paragraph summary. We do not require the description to conform to APA (2010) style. 4. Finally, the students should submit the lab report that includes a) an APA (2010) style cover page, b) the written description of the results, c) the figure with a figure caption, and d) the Experiment 1 Worksheet. Experiment 2: Investigating Classical Conditioning Learning objectives. Once students have completed this experiment, they should be able to demonstrate the following skills: a) design and carry out a classical conditioning experiment to determine how the CS intensity, US intensity, and the temporal 123

5 relationships between them influence the final level of the CR; b) summarize results in APA (2010) format and present them in a table. Principle. Several variables are known to influence classical conditioning. Conditioned stimulus intensity refers to the strength of the CS. In general, the CR is acquired faster when the CS intensity (e.g., loudness of the metronome) is high. Unconditioned stimulus intensity refers to the strength of the US. The US intensity (e.g., amount of food) has two effects on the CR: it determines the rate of acquisition and the maximal amount of the CR that is possible. The temporal arrangement of the CS and US refers to the timing between them. In general, the better the CS predicts the delivery of the US, the more conditioning that will occur. In this experiment, each lab team investigates one of the following questions about these variables: 1. Does the intensity of the CS influence the final level of the CR? 2. Does the intensity of the US influence the final level of the CR? 3. Which temporal relationship appears to produce the strongest final CRs? 4. Which temporal relationship appears to produce the weakest final CRs? We randomly assign a question to each lab team. The team is required to design an experiment and carry out the appropriate simulations to provide the answer. Materials. The materials for conducting the experiment are Virtual Pavlov 2000 (Benedict, 2000) and the Experiment 2 Worksheet. The worksheet contains spaces for the student to include all of the parameters of the experiment, including the CS and US intensities, CS-US temporal arrangement, the number of acquisition trials, the levels of the independent variable, the dependent variable (i.e., drops of saliva on the last trial), and an instructor sign off area. Procedure. We recommend students and instructors use the following procedure for conducting the experiment: When the students arrive to their computer station, the computer should be turned on and the program menu appearing on the desktop. The students should select the Salivary Conditioning subprogram from the menu and follow the instructions on the screen to set up the experiment. Each question should be answered by running a simple between-subjects experiment to compare the effectiveness of classical conditioning under different values of the independent variable. The experiment should be set up so that two groups of subjects are run under each different value. The mean and standard deviation for each group should be calculated after the simulations and used to answer the question. We recommend the students start by identifying the independent variable in the question they will be investigating. The best way to identify relevant values of the independent variable is for the students to run the simulation program and note what different values can be investigated on the computer. Once the students have decided which values are appropriate to answering the question, they are ready to move on to the next step. We recommend the students use at least 10 dogs per group. They should also identify the dependent variable, such as the drops of saliva on trial 20, to assess the effect of the independent variable. A common problem we have noticed is students will not specify the dependent variable in a way that can be counted. For example, a student will state the amount of conditioning present on trial 20 as the dependent variable on their worksheet. We suggest the dependent variable be expressed in the units used to measure it, such as drops of salivation. Another area of concern is some students will not include the units for other parameters, such as the decibel level of the metronome or the amount of food presented on each trial. Once the students have provided all of the parameters with the correct units, they are ready to be signed off so they can conduct the experiment. Once the students are signed off on their experimental design, they can conduct the experiment and record the data for each dog in both groups. Data collection follows the same procedure as outlined in Experiment 1. Laboratory report requirements. We recommend instructors use the following laboratory requirements: 1. The students should complete the experiment and turn in the Experiment 2 Worksheet. 2. The students should prepare an APA (2010) style table to present the mean and standard deviation for each group and an APA style results section. Table 2 presents data collected to answer Question 1. The CS (metronome) intensity was 50 db and the CS-US temporal arrangement was a 3-s 124

6 Table 2. Drops of Saliva (CR) as a Function of US Intensity and Acquisition Trial 20 Group M SD 5 g g Note. These data were collected to answer Question 1 in Experiment 2 (Investigating Classical Conditioning). short delay for both groups. Each group consisted of 10 dogs simulated by Virtual Pavlov 2000 (Benedict, 2000). We do not require the students to conduct an independent-samples t-test to analyze the data. However, it certainly could be made a requirement depending upon the level of your students. 3. Finally, the students should submit a lab report that includes a) an APA (2010) style cover page, b) Results section, c) the table, and d) the Experiment 2 Worksheet. Experiment 3: The Conditioned Emotional Response Procedure Learning objectives. Once students have completed this experiment, they should be able to demonstrate the following skills: a) define operant conditioning, variable-interval schedule, baseline, conditioned emotional response (CER) procedure, conditioned suppression, suppression ratio, and cumulative record; b) create and interpret a cumulative record; c) calculate and interpret a suppression ratio; and d) simulate the CER procedure using Virtual Pavlov 2000 (Benedict, 2000). Principle. Experiment 3 illustrates an important, indirect way of measuring a classically conditioned response: the CER procedure (e.g., Estes & Skinner, 1941). The technique has been used to investigate several important theoretical phenomena in classical conditioning, including the blocking effect simulated in Experiment 4. The objective of Experiment 3 is for the students to understand the CER procedure. In the first two lab experiments, the dependent variable was the number of drops of salivation elicited by the CS. However, not all reflexes can be so directly observed. Consider conditioned fear responses. A fearful animal may experience an increase in heart rate, a change in respiratory patterns, dilation of pupils, an alteration in digestive processes, or other reflexes. Which one of these responses should be selected as the dependent variable? How should the selected reflexes be observed and recorded in a study of conditioned fear? The CER Procedure is an alternative way to measure the fear response in which the intensity of conditioned fear is measured through its effects on another, unrelated behavior. The degree of conditioned fear is inferred by observing the extent to which the fear disrupts the unrelated behavior. The unrelated behavior is chosen to be one that is easy to observe. A high level of disruption is understood as representing intense fear. In the usual implementation of the CER procedure, an animal is trained to press a lever for food reinforcement using standard operant conditioning techniques. The reinforcer for lever pressing is delivered on a variable interval (VI) reinforcement schedule, which means that the animal is reinforced for the first response that occurs after a randomly determined amount of time (Ferster & Skinner, 1957). A common VI schedule is one where the amount of time that must pass between reinforcers varies between 10 s and 2 minutes with an average interval of 1 minute. After a reinforcer is delivered, the experimenter (or a computer) randomly picks a new interval between 10 s and 2 minutes to be the time that must pass before another response can be reinforced. For example, if the experimenter picked 45 s as the inter-reinforcer interval, no reinforcers would be given for any lever press until the 45 s elapsed. The next press would be reinforced and a new random time picked for the next interval. Since the intervals are chosen so that the average interval is 1 minute, this is called a VI 1 minute reinforcement schedule. The advantage of placing the animal on a VI 1 minute schedule is that lever pressing behavior will stabilize into a well established pattern characterized by a steady rate of responding. This steady rate of bar pressing provides an excellent baseline for judging the disruptive effects of conditioned fear. After the animal s behavior appears to have adapted to the VI 1 minute schedule, the operant conditioning portion of the CER Procedure is now ready to indirectly assess the intensity of a conditioned fear reaction. The next phase of the CER Procedure is to condition the fear response itself. The lever is typically removed from the cage for this step. The animal now experiences a stimulus such as a tone (CS) paired with an unconditioned stimulus (US) for fear, such as a short electric shock. The tone (CS) is presented before the US so classical conditioning can take place. Not too many pairings will be necessary 125

7 because conditioned fear responses are rapidly learned. In the last phase of the CER procedure, the intensity of the conditioned fear response to the tone is assessed by reintroducing the bar back into the cage. When the animal is again pressing steadily at the lever, the tone is sounded and the amount of lever pressing recorded. The decrease in lever pressing that occurs during the CS is called conditioned suppression and is typically expressed as a suppression ratio [(B)/(A + B)]. In the equation, B is the number of responses produced during the tone (CS) and A is the number of responses produced in the time period before the CS was sounded. If the tone completely eliminated the lever pressing, the suppression ratio would be zero (i.e. B = 0); if the bar pressing went on undisturbed so that A = B, the suppression ratio would be 0.5. The suppression ratio, then, is a number between 0.00 and 0.50 that expresses the degree of disruption a CS produces to an ongoing operant behavior. The suppression ratio is used as an indirect measure of the intensity of the conditioned fear reaction. The smaller the suppression ratio, the stronger the conditioned fear response. The suppression ratio and the CER procedure are important because a number of significant classical conditioning studies have used them and they illustrate an interesting case where apples are used to measure oranges. An operant conditioning setting is used to assess the strength of a classically conditioned response. The suppression ratio involves the number of lever presses reinforced by food (operant conditioning), but the ratio itself is taken as a dependent variable expressing the effectiveness of the classical conditioning procedure in which the tone (CS) was paired with shock (US). Materials. The materials for conducting the experiment are Virtual Pavlov 2000 (Benedict, 2000) and the Experiment 3 Worksheet. The worksheet requires the students to a) explain how a VI 1 minute schedule works, b) give an example to illustrate the calculation of a suppression ratio, and c) explain how the degree of classical conditioning is reflected in the magnitude of the suppression ratio. Procedure. As in the previous experiments, the computer should be turned on and the program menu appearing on the desktop. The students should select the Suppression Ratio subprogram from the menu. The students should follow the instructions on the screen to begin the experiment. Experiment 3 is a demonstration of how the CER Procedure works and it illustrates the key concepts described above. Before students begin the experiment we ask them to make sure they understand a) the VI schedule, b) a baseline, and (c) the suppression ratio. If they do not understand the concepts, they are invited to ask the teaching assistant or lab instructor for clarification. The experiment has nine parts to it, and each helps illustrate the key concepts in the CER Procedure. As the student performs each part, we ask them to make sure they understand the point that is being illustrated. We recommend students and instructors use the following procedure for conducting the experiment: The first activity (Part 1) is a demonstration of a rat responding on a VI schedule. It represents the rat s responses as a cumulative record. We emphasize to the students that it is an important kind of graph that represents each response the rat makes as a function of time. We point out how each response moves the cumulative total up one unit on the graph, how each reinforcer delivered is represented on the graph as a vertical slash, and how time is represented on the horizontal axis. The students should be able to read the details of a subject s behavior from moment to moment, seeing when the behavior increased, when a reinforcer was delivered, and when behavior ceased. Next (Part 2), the program asks members of the lab team to generate a cumulative record of their own by pressing on the space bar. The students must press the enter key to start this part. A VI schedule will be in effect, so reinforcement will occur periodically. As instructors, we use this as an opportunity to check their understanding of how to interpret a cumulative record. In Parts 3 and 4, the program demonstrates the way fear is shown in the CER procedure. In Part 3, a rat with no fear of a tone presses the lever. A 15 second interval is noted on the cumulative record right before a tone is sounded. This is used as a baseline. Since this rat is not afraid of the tone, its rate of pressing is about the same in the baseline segment and during the tone itself. In Part 4, the rat has associated the tone with fear. Students can see what happens when the tone turns on -- the rat stops pressing. By comparing the rat s pressing in the baseline segment with its pressing during the tone, the degree of fear can be quantified. The next four parts (Parts 5 9) ask the students to simulate the lever pressing of a rat that either has or has not become afraid of the tone and then compute the suppression ratio. To obtain the data for computing the suppression ratio, the student should move the mouse pointer over the desired part of the response line on the cumulative record. When the pointer is properly placed on the line, the number of responses will appear on the screen. 126

8 Figure 2. Cumulative lever presses emitted by a hypothetical rat as a function of minutes. Annotations A, B, and C indicate fast, slow, and reinforcer delivery, respectively. Annotation D indicates when the hypothetical rat is presented with a CS that predicts a shock during the Conditioned Emotional Response procedure in Experiment 3. Laboratory report requirements. We recommend instructors use the following laboratory requirements: 1. Have the students complete all 9 parts of the experiment. 2. To demonstrate the students understanding of cumulative records, we require them to create a cumulative record for an imaginary experiment in which rat s lever pressing occurs with and without a CS that had previously been paired with a shock. The students construct their cumulative record with Microsoft Excel. However, there is no reason another program, such as QuattroPro, could not be used. We ask them to annotate parts of the cumulative record by using the letters A, B, and C, and D for fast, slow, reinforcement, and CS presentation, respectively. The annotations should also be described in the figure caption. For an example, please refer to Figure 2. It presents cumulative lever presses emitted by a hypothetical rat. 3. Finally, the students should submit a lab report that includes a) an APA (2010) style cover page, b) figure and figure caption page, and c) the Experiment 3 Worksheet. Experiment 4: The Blocking Effect Learning objectives. Once students have completed this experiment, they should be able to demonstrate the following skills: a) define the blocking effect, compound conditioned stimulus, CS1, and CS2, and b) design and carry out a classical conditioning experiment to examine how the US intensity and the number of CS-US pairings in Phases 1 and 2 influence the blocking effect. Principle. The blocking effect was reported by Kamin (1969) and was immediately recognized as a finding that provided an important clue about the mechanisms underlying classical conditioning. In the blocking effect, certain kinds of experiences are shown to prevent future classical conditioning. The most significant contribution of the blocking effect has been the challenge it offers to theories of classical conditioning to explain why it occurs. In Experiment 4, the students will simulate a variant of Kamin s classic study using Virtual Pavlov 2000 (Benedict, 2000). The blocking effect involves a compound conditioned stimulus which is a pair of CSs presented together simultaneously. For example, a tone (CS1) and a light (CS2) might be presented at exactly the same time prior to presenting the US. The combined tone and light is called a compound CS. It has been known since Pavlov (1927) that pairing a compound CS with a US produces learning to both individual CSs. In the example above, if the light and tone are presented separately after the compound stimulus has been paired with the US, a CR will be observed for each stimulus element. However, Kamin (1969) showed in some situations conditioning can be blocked for one of the stimulus elements so that the CR is established only for one of the stimulus elements and not the other. Kamin discovered that the key operation required to produce the blocking effect was giving the subject prior experience with one of the stimuli in the compound CS. After such an experience, Kamin found that pairing the compound stimulus repeatedly with the US would produce only a very small CR from the stimulus element not included in the prior trials. There were several steps in his classic study: Kamin (1969) used the CER procedure where rats were initially taught to press a lever for food reinforcers on a VI 1 minute schedule. This operant behavior served as the baseline for evaluating the strength of the CRs conditioned later in the experiment. In Phase 1, some of the subjects were given experience with one of the stimulus elements. For example, the subjects might be exposed to 10 trials in which the tone (CS1), but not the light (CS2), was paired with shock (US). In Phase 2, the compound stimulus (tone + light) was repeatedly paired with the same US. During the testing, the subjects were put back on the VI 1 minute schedule for the final phase of the study. The light and tone were separately presented without shock in these test trials. After responding was observed, a suppression ratio was computed for each stimulus element to 127

9 measure the CRs. Kamin (1969) found that the stimulus presented in Phase 1 (CS1, the tone) showed a strong CR while the other stimulus element (CS2, the light) showed a weak one. Kamin concluded that the experience with the tone in Phase 1 blocked acquisition of a CR to the light seen in Phase 2. In this experiment, each lab team will investigate one of the following questions about these variables: 1. Does the number of pairings in Phase 1 affect the magnitude of the blocking effect? 2. Does the number of pairings in Phase 2 affect the magnitude of the blocking effect? 3. Does the intensity of the shock in Phase 1 affect the magnitude of the blocking effect? 4. Does the intensity of the shock in Phase 2 affect the magnitude of the blocking effect? We randomly assign one of the above questions to each lab team. The team is required to design an experiment and carry out appropriate simulations to provide the answer. Materials. The materials for conducting the experiment are Virtual Pavlov 2000 (Benedict, 2000), Experiment 4 Design Sheet, and Experiment 4 Worksheet. The Experiment 4 Design Sheet contains spaces to specify the independent and dependent variables, experimental procedure, and the hypothesis. The worksheet requires the students to a) explain the blocking effect and give the conditions necessary to produce it, b) give the mean and standard deviation of the suppression ratios for each group, and c) interpret the suppression ratios for each group by discussing what the suppression ratios imply about their rats' behavior during the tests and what they suggest about the strength of the rats' fear responses. Procedure. The computer should be turned on with the program menu appearing on the desktop. The students should select the Fear Conditioning subprogram from the menu and follow the instructions on the screen to begin the experiment. Initially, we recommend the students conduct a preliminary experimental design of their question. We require the experimental design to be a simple two group between-subjects design. Each group should have 10 subjects each to investigate the assigned empirical question. If your students are more advanced in statistics and research methods, we advise you to modify these requirements to include more groups. The computer program will allow the students to vary the number of pairings and the shock intensity in each phase. Students should select reasonable (and theoretically important) values that address the empirical question. After the students completed the preliminary design, we recommend they conduct a partial replication of Kamin s (1969) experiment. Based on our experience, we believe replicating Kamin provides a solid demonstration of simple blocking while also increasing the students familiarity with Virtual Pavlov 2000 s (Benedict, 2000) blocking simulation. Here are the parameter values we recommend for replicating Kamin with Virtual Pavlov The tone in Phase 1 should be presented for 16 trials with the US intensity set to 1.0 ma. In Phase 2, both the tone and light should be presented for 8 trials and the US intensity set at 1.0 ma. Using these values, students should observe relatively little fear in response to the light but significant fear to the tone. Once the students have completed the partial replication of the blocking effect, we recommend they modify the parameters in the second group to answer their empirical question. Next, each lab team s experimental design should be approved by the laboratory instructor or teaching assistant. The students should then conduct the experiment by using the program menus to select the appropriate independent variables for the simulation and collect the data. We ask them to assume that each rat in the experiment has been taught to lever press on a VI 1 minute schedule before the classical conditioning phases begin. Another point to emphasize is Virtual Pavlov 2000 (Benedict, 2000) will automatically report the suppression ratios from Phase 3 (i.e., the test phase). Students often find this to be confusing because Virtual Pavlov 2000 does not provide a visual simulation to accompany the data as in the previous experiments. Therefore, it is important to remind the students that the hypothetical rats are again placed in a Skinner box where each CS is tested separately. As in Experiment 3 (The CER Procedure), disruptions in their lever pressing are used to calculate the suppression ratios. The students should record a suppression ratio for each subject in response to both the tone and the light. Also remind the students to run the simulation twice: once for the control group and once for the experimental group. Laboratory report requirements. We recommend instructors use the following laboratory requirements: 1. Students should complete the Experiment 4 Design Sheet and Experiment 4 Worksheet. The design should include a signature by the 128

10 Table 3. Suppression Ratios as a Function of the Number of Tone/Shock Pairings in Phase 1 Trials in Phase 1 Tone Light M SD M SD Note. These data were collected to answer Question 1 in Experiment 4 (Blocking Effect). instructor or laboratory assistant. The data for each group should be summarized with the mean and standard deviation of the suppression ratios for the tone and the light. The answers to the questions on the Experiment 4 Worksheet should also be included. Table 3 presents data collected to answer Question 1. During Phase 1, rats were exposed to either 1 or 16 tone (CS) - shock (US) pairings. During Phase 2, both groups were exposed to 8 compound CS (tone & light) shock (US) pairings. The shock intensity in both phases was 1.0 ma. The suppression ratios were derived from Phase 3 of the Conditioned Emotional Response procedure. Each group consisted of 10 rats simulated by Virtual Pavlov 2000 (Benedict, 2000). 2. The students should submit a lab report that includes a) an APA (2010) style cover page, b) Experiment 4 Design Sheet, and c) the Experiment 4 Worksheet. Additional requirements. The above requirements may seem unchallenging. However, the data from Experiment 4 are also the basis for first major paper in the course. In addition to the above requirements, the students are required to submit a paper that includes a cover page, Methods, Results, a table summarizing the data, and an appendix detailing the statistical calculations. We ask the students to write the paper as if they conducted the experiment with real rats rather than a computer simulator. In addition to the descriptive statistics, the students are also required to conduct to independentsamples t-tests to determine if a significant difference in the blocking effect occurred. These should be summarized according to APA (2010) format (i.e., correct statistical copy) in the Results section. We also ask them to provide an appendix with the t-test calculations to ensure that they were done properly. If any errors are detected, we give them feedback on the mistakes in the hope that the errors will not repeat themselves later in the semester. Operant Conditioning Laboratory Activities General Overview of Operant Laboratory Procedures The next four experiments are designed to illustrate important operant conditioning principles discovered by B. F. Skinner (1938). During this part of the semester, lab partners work with real behavior in real time. The basic criterion for successfully completing each of the experiments is the observed behavior of the rat. When the rat performs at criterion level, the experiment is finished except for some brief paperwork. Lab partners apply established operant conditioning principles and see firsthand what happens. The results are interesting and have implications for the understanding of human behavior (e.g., Skinner, 1953). The experiments emphasize the relationship between behavior and its consequences. After learning the conditioned reinforcement procedures in Experiment 5, the students shape the most classic of all operant behaviors, the lever press (Experiment 6). Lever pressing is brought under stimulus control in Experiment 7 so that it only occurs in the presence of a light. The final operant experiment expands the rat s repertoire by teaching a complex sequence of behaviors. Taken together, the following four experiments are a good introduction to practical techniques of operant conditioning and have applications to real world situations ranging from animal training to behavior therapy (Skinner, 1953). If the reader wished to expand upon these operant experiments, we recommend consulting the classic operant laboratory manuals by Michael (1963) and Reese (1964). Rats Our teaching laboratory has an animal holding room for 50 adult, singly-housed Norway rats (Rattus norvegicus). Our facility is inspected on a semi-annual basis and is compliant with the National Institutes of Health (National Research Council, 1996) guidelines. The husbandry, educational, and research protocols are approved by the University of Alaska Anchorage s Institutional Animal Care and Use Committee (IACUC). We order 45-day old Wistar or Long-Evans rats from Simonsen Laboratories (Gilroy, CA) before each 129

11 Animal Welfare Act, ethical treatment of research animals, and the IACUC review process. The approved IACUC protocol covering the rat experiments is also available for each student to review. Operant Conditioning Chambers Figure 3. A schematic of the custom made Skinner box used in our operant laboratory. All of the operant contingencies are manually controlled by the students. The apparatus is used for training real laboratory rats in Experiments 5-8. semester. Following their quarantine period, the rats are handled for docility on a regular basis by our teaching assistants. When the rats are approximately 90 days old, they are placed on a feeding schedule. The rats are given free access to water, but access to food is controlled between Saturday and Thursday of each week so that food can be used as a primary reinforcer in the experiments. During the controlled feeding days, rats are given a daily ration of food (8-10 g) each afternoon. The laboratory sessions are held before this daily feeding. The rats have free access to food from Thursday evenings to Sunday afternoon. This system maintains the health of the animals and motivates them to respond for food reinforcers. Based on our experience, the rats are maintained at approximately 85% of their ad libitum weight by using this procedure. At the beginning of each laboratory session, the rats are weighed. The students record their rat s weight in their lab notebook and on the cage card. Before working with the rats, each student is required to complete two online training courses offered by the Collaborative Institutional Training Initiative ( The required training module for everyone working with vertebrates at our institution is the Basic Course for Investigators, Staff, and Students. The course covers basic information on the Animal Welfare Act (United States Department of Agriculture, 2005) and ethical considerations when using animals in research and teaching. The second course is called Working with Rats in Research and covers the unique husbandry and handling requirements of the species. Additionally, we dedicate a portion of the first operant conditioning laboratory to discussing the Our teaching laboratory has custom-made operant conditioning chambers (i.e., Skinner box). Each Skinner box has a lever, discriminative stimulus light, clicker, and food dish (see Figure 3). Forty-five mg food pellets (BioServ, Frenchtown, NJ) are delivered manually through a funnel. Each lab team uses the box exclusively during the operant segment of the semester. During the first operant conditioning session, we ask the students to familiarize themselves with it before beginning the first experiment. The best method for doing this is to experiment with the buttons and switches. Accurate and prompt reinforcement is crucial for successful conditioning, so we ask the students to practice delivering a few reinforcers before the first experiment. Each Skinner box is stocked with a bottle of 45 mg food pellets, a timer, a marble, and a bolt. Experiment 5: Creating a Conditioned Reinforcer Learning objectives. Once students have completed this experiment, they should be able to demonstrate the following skills: a) define positive reinforcer, primary reinforcer, conditioned/secondary reinforcer, and magazine training; b) handle a laboratory rat; c) systematically observe a rat s behavior in a Skinner Box; and c) establish a clicker as a conditioned reinforcer. Principle. In order to strengthen a specific behavior, it is necessary to follow that behavior with an immediate, positive reinforcer. A positive reinforcer is a stimulus (e.g., food) that follows a behavior and increases the frequency of that behavior. The noise that the Skinner box s clicker makes when it is actuated serves this function because the clicker predicts the delivery of a food pellet. The purpose of Experiment 5 is to establish that relationship between the clicker and food. The clicker is called a conditioned reinforcer (sometimes called a secondary reinforcer) because of its ability to immediately reinforce the rat s behavior after being paired with the primary reinforcer, food. The food is called a primary reinforcer because no training is necessary to teach a rat to respond for food. The establishment of a conditioned reinforcer is an important part of operant conditioning because it 130

12 allows a reinforcer to be more precise than it would be using the primary reinforcer alone. Materials. The materials for conducting the experiment are one healthy laboratory rat, a functioning Skinner box, a good supply of 45 mg food pellets, and Experiment 5 Data Sheet, and the Experiment 5 Worksheet. The Experiment 5 Data Sheet contains lines for recording the number of food pellets delivered and the date of the lab session. The Experiment 5 Worksheet requires the student to a) explain the purpose of establishing the clicker as a reinforcer, b) explain why it necessary to give the final reinforcers while the rat was away from the food dish, c) why the reinforcers were given at irregular intervals, d) report how many reinforcers were delivered, and e) report if there was any signs of satiation toward the end of the experiment. Procedure. Magazine training is a procedure that pairs a click with the delivery of food. As a result, the click becomes a positive reinforcer through classical conditioning. During this experiment, we require the students to keep track of the number of pellets the rat is given. The purpose of this record is to determine how many pellets it will take to satiate the rat. Based on our experience, rats typically satiate after pellets. The students should place some food pellets into the food tray and allow the rat to habituate to the new environment and find the food pellets. After the animal has eaten the food, the students should introduce the clicker. Occasionally, some rats will startle to the initial sound of the clicker, but this should dissipate after several trials. We recommend each pellet be given by performing the same sequence: sounding the clicker once followed by dropping the food pellet into the food tray. After the rat has eaten the pellet, another click and pellet should follow. The students should precede each pellet with a click and always wait until the rat has eaten one before giving another. They should continue to sound the clicker and deliver pellets while the rat is near the food dish for about 20 reinforcers. The interval between pellets should be varied so that the rat associates food delivery with the sound of the click rather than the passage of time. When the rat has begun to eat as soon as he hears the click, the students should withhold reinforcement (clicker) until he moves away from the food dish. Thereafter, reinforcers should be given at irregular intervals until the animal s response to the click is a direct movement to the food dish regardless of where the animal is in the chamber when the click is sounded. The students will have successfully established the clicker as a conditioned reinforcer when the rat moves directly to the food dish when the clicker is sounded from any position in the Skinner box. Laboratory report requirements. We recommend instructors use the following laboratory requirements: 1. The rat approaches the food dish whenever a click is sounded. 2. The students can verbalize the conditioned reinforcement procedure. It is very important the verbal description include a statement about classical conditioning contingencies rather than the rat expecting that the food pellet is in the dish because it hears a click. Students will often infer mental explanations for their rat s behavior even when more parsimonious reasons exist (Leigland, 1995). 3. The students should submit a lab report that includes an APA style cover page and the Experiment 5 Worksheet. Experiment 6: Lever Pressing Learning objectives. Once students have completed this experiment, they should be able to demonstrate the following skills: a) define extinction, shaping, successive approximations, target behavior, and ABAB reversal design; b) shape a target behavior by systematically reinforcing successive approximations; c) conduct an experiment using an ABAB reversal design; d) present data using a cumulative record. Principle. The object of this experiment is for the students to investigate the effects of reinforcement and extinction on the lever pressing behavior their rat. Extinction is a procedure that stops the delivery of a reinforcer for the behavior that it follows. As a result, rate of responding decreases. First, the students will observe their rat s spontaneous lever pressing in a 10 minute period without reinforcement and then the rat will be trained to press the lever using shaping techniques. Shaping is the process of systematically reinforcing successive approximations to a target behavior. The target behavior in Experiment 6 is the lever press. Successive approximations are steps toward the target behavior. Once lever pressing is firmly established using continuous reinforcement (i.e., every response is reinforced), the students will count the number of presses in a second 10 minute period. Next, the students will examine the control exerted 131