Associative learning occurs when an organism links two or more items of information. The simplest forms of associative learning are classical conditioning and instrumental conditioning. Classical conditioning is also known as Pavlovian conditioning in honor of Ivan Pavlov (1849–1936) who was the first person to conduct extensive research of this nature. In a typical experiment with dogs, Pavlov would present a neutral auditory stimulus such as a metronome immediately before applying sand or food powder to the dog's tongue, which produced salivation. After a number of these pairings, Pavlov presented the metronome alone, and the dog now salivated. Pavlov developed terminology for these components of classical conditioning: The dog experienced a relatively neutral stimulus or conditioned stimulus (CS, the metronome) in conjunction with a biologically significant stimulus or unconditioned stimulus (US, the food powder), which always produces an unconditioned response (UCR, salivation). After multiple CS-US pairings (i.e., acquisition), presentation of the CS alone elicited a response, the conditioned response (CR, also salivation), which is appropriate for its corresponding US. Following acquisition of the CR to the metronome CS, Pavlov also reported that presenting the CS alone a number of times would eventually eliminate the salivation CR, a procedure termed extinction.

Although most classical conditioning experiments have used nonhumans, classical conditioning readily occurs in humans (e.g., Hermans, Craske, Mineka, & Lovibond, 2006). With nonhumans, many model systems have been developed to explore classical conditioning, including conditioned eyeblink, conditioned taste aversion, and conditioned approach/avoidance (Domjan, 2003). Clearly, these have little classroom application, but the most common classical conditioning paradigm, conditioned emotional response, is applicable. Conditioned emotional reactions can be either positive or negative. A positive conditioned emotional response is produced by pairing a relatively neutral stimulus with a US that elicits a positive emotion such as happiness. For example, a parent may use the preferred taste of cheese to cover the flavor of broccoli. After a few meals of cheesy broccoli, a child will be more willing to eat broccoli by itself (for a review of food preference learning, see Capaldi, 1996). Although it is possible to produce positive emotional reactions, broad application of this methodology has not been implemented (other than in advertising when an attractive model is paired with a product). In practicality, it may be difficult for a teacher to spend substantial time during the earliest portions of class to pair their presence with a positively affective US. One can imagine only the youngest of children would not see through an instructor plying them with candy or treats on the first day of class.

Instead, the more commonly studied phenomenon, and the more likely classroom occurrence, is the negative conditioned emotional response. A classic example of conditioned fear in humans is the Little Albert Study conducted by Watson and Rayner (1920). Watson and Rayner examined if a phobia could be induced in a human, so they borrowed nine-month-old Albert from the nursery at Johns Hopkins University. After recording Albert's baseline responses to a range of stimuli such as animals and neutral objects, conditioning began two months later. During acquisition, a white rat was paired with a loud noise US (Watson clanged a steel bar with a hammer) seven times. Five days later, Albert was tested with a range of stimuli, including the white rat. Albert cringed and cried in response to the rat, behaviors that were quite different from his curiosity about the rat during the baseline phase. They tested long-term retention of fear 30 days later, and Albert was still scared of the white rat and other white objects such as a rabbit, white fur coat, and Santa Claus mask. In addition to demonstrating conditioned fear in humans, Watson and Rayner planned to examine the conditions necessary to extinguish Albert's fear. Unfortunately, on the day prior to the implementation of the extinction phase, Albert was released from the hospital with his fear intact.


Just as Albert learned to fear the white rat, the potential for students to learn a phobia to a neutral classroom or instructor is always present. Few instructors aim to produce a threatening or fearful situation, but a wrong answer or embarrassing situation may induce a negative emotion in the student and confer learning to the cues present at this time. As a result, the student may choose to miss class or decrease participation during class. A specific example comes from an advanced course in tests and measurements. Here, students are often given a short, timed, math ability test. Students frequently report experiencing anxiety during this exam. Some report their heart beating faster and harder, shortness of breath, and inability to concentrate because they cannot ignore the stopwatch used for timing. They often state “I hate math,” or “I can't do math.” Occasionally the anxiety is so strong, they stop in the middle of a problem and say, “I can't go on.” When asked about their anxiety many trace its beginning to One Minute Arithmetic Tests in elementary school. In these tests, they had a sheet of arithmetic problems to complete correctly in one minute, or repeat the test until they did so. From a classical conditioning perspective, the CS is the arithmetic problems and the US is the time pressured testing situation that produces pressure and anxiety (the UCR) to both finish in a rapid time (one minute) and calculate problems correctly. After the CS-US pairing the CS (the math problems) alone produces anxiety (CR).


Because the student overcomes the anxiety enough to pass these tests, an instructor who has learned about conditioned fear might expect an extinction treatment has effectively eliminated the fear. A common misconception of extinction is that it is the equivalent of unlearning: Once a CS-US association has been learned, this association could be unlearned if the CS is frequently presented alone, like erasing a word from a blackboard so no trace remains. Yet, numerous studies have shown that extinction is not unlearning, but is actually new learning. Two extinction-related phenomena, spontaneous recovery and renewal, illustrate this interpretation. In spontaneous recovery, an organism experiences acquisition of the CS-US association, followed by CS-alone presentations (i.e., extinction). If the student is tested within a few days of the extinction trials, a weaker CR is observed. However, if CS testing is delayed for a few weeks (e.g., 21 days), a significantly stronger CR (anxiety reaction) is recorded. The fact that the CR returns without any additional training suggests that the original CS-US association is still intact (Rescorla, 2004). Indeed, some neuroscience studies suggest that the locus of the acquisition memory (CS-US association) is a different anatomical region from the locus of the extinction memory (CS-no US association) (e.g., Sotres-Bayon, Cain, & LeDoux, 2006). Thus it would not be unusual for a strong math anxiety reaction to recur spontaneously in students who had not experienced panic in quite some time.


Further support for this view comes from a second extinction-related phenomenon, the renewal effect (e.g., Bouton 2002; Bouton & King, 1983). The renewal effect is produced by alterations in the contexts of learning and extinction. For example, a control group will learn the CS-US association in an aqua room (A), experience CS alone experiences (i.e., extinction) in this aqua room (A), and then be tested with the CS in the same aqua room (A). Not surprisingly, following this order of experiences, members of Group AAA will show weak responding to the CS in the aqua room during testing. In contrast, the experimental group will receive learning in the aqua room (A), extinction in a blue room (B), and testing of the CS in the aqua room (A). Group ABA will show a significantly stronger response to the CS during testing than Group AAA. This outcome provides convergent evidence that the original CS-US association learned in the first (A) phase is still intact and can be retrieved if the contextual cues during testing are the same as during learning.

One might ask why some students develop test anxiety while others did not. In 2001, Bouton, Mineka, and Barlow proposed a modern learning theory approach to panic disorder that can be extrapolated to this anxiety. According to their model, during an experienced panic episode, various external or internal cues (CSs) can become associated with the negative emotion US. As individuals re-encounter these CSs, they experience conditioned anxiety, and this may lead to a panic attack. Bouton and colleagues also argue that a major contributing factor to panic and conditioned anxiety is catastrophic misinterpretation of somatic symptoms. In the math test anxiety situation, the students experience panic during a testing situation, perhaps as an increasing heart rate or impaired recall or attention. These cues can lead to catastrophic thoughts, such as “I can't do math.” There are a number of other cues that can serve as the CS in this situation. The contextual cues of the classroom are encountered in academic settings, the quiet shuffling of feet and the sound of pens scribbled on paper, the visual features of other individuals quietly hunched over their desks could be another cue. Importantly, a host of internal cues may be most salient, suchasthe heart rate and racing thoughts. If the students engage in catastrophic thinking at this stage, then they may increase their anxiety worrying that their performance on this test is the first step to career success or failure. Overall, this combination of events can then be generalized to a host of other testing situations because most tests will have common visual, auditory, or contextual conditions, these will be coupled with internal somatic sensations that seem unique to the situation, and even though the students have passed the previous exam, the current exam holds the same potential for life-long failure.


Bouton, M. E. (2002). Context, ambiguity, and unlearning: Sources of relapse after behavioral extinction. Biological Psychiatry, 52, 976–986.

Bouton, M. E., & King, D. A. (1983). Contextual control of the extinction of conditioned fear: Tests for the associative value of the context. Journal of Experimental Psychology: Animal Behavior Processes, 9, 248–265.

Bouton, M. E., Mineka, S., & Barlow, D. H. (2001). A modern learning theory perspective on the etiology of panic disorder. Psychological Review, 108, 4–32.

Capaldi, E. D. (Ed.). (1996). Why we eat what we eat: The psychology of eating. American Psychological Association, Washington, D.C.

Domjan, M. (2003). The principles of learning and behavior (5th ed.). Belmont, CA: Thomson-Wadsworth.

Hermans, D., Craske, M. G., Mineka, S., & Lovibond, P. F. (2006). Extinction in human fear conditioning. Biological Psychiatry, 60, 361–368.

Rescorla, R. A. (2004). Spontaneous recovery. Learning and Memory, 11, 501–509.

Sotres-Bayon, F., Cain, C. K., & LeDoux, J. E. (2006). Brain mechanisms of fear extinction: Historical perspectives on the contribution of prefrontal cortex. Biological Psychiatry, 60, 329–336.

Watson, J. B., & Rayner, R. (1920). Conditioned emotional reactions. Journal of Experimental Psychology, 3, 1–14.