ON THE HANDLING OF LABORATORY RATS
EXPERIMENT 3: SHAPING THE LEVER PRESS
EXPERIMENT 4: EXTINCTION AND SPONTANEOUS RECOVERY
EXPERIMENT 5: FIXED INTERVAL REINFORCEMENT
EXPERIMENT 6: STIMULUS CONTROL
EXPERIMENT 7: INDEPENDENT EXPERIMENT
home | MSUS,M home page | CSBS home page
ON THE HANDLING OF LABORATORY RATS
An important part of this course will involve the use of animal subjects. In many types of research, animals are preferred to humans because of the greater degree of experimental control and genetic uniformity possible. Because of this genetic uniformity it is often said that "there are no dumb rats, just dumb experimenters." That is, there are no external influences on your rat's behavior other than what you as an experimenter do to her, and laboratory rats are as alike as any mammals can be in regard to their intellectual capabilities. Therefore, if your rat is not performing the way that you expect her to, it is probably because of something that you are doing -- or not doing.The rats that we are using this semester will be on loan from a commercial breeder and are usually about 50 days old on arrival. Their actual birth date and genetic information is printed on the card on the cages. You'll need this information for your subject section.
Since the normal life expectancy of (domestic) laboratory rats is about two years, they are still minors when we get them. They will be just approaching adulthood while we are working with them. We use female rats since they are smaller than the males and (supposedly) more docile. Unlike humans, laboratory rats will bite only under extreme provocation, or when very hungry. Since we do not deprive the rats of food, it is up to you to avoid extreme provocation. It would be a good idea (even if you were not in this course!) to have had a tetanus booster shot within the last ten years.It's a good idea to spend as much time as possible handling your rat before you start to work with her. As she gets tamer and more accustomed to you and the experimental situation, she will concentrate more on the business at hand.The rats are housed in Room 43; experiments are run in Room 40B. We (myself and the work study student assigned to animal duty) will feed the rats. They (the rats) will have unlimited access to food, and will be maintained on a 24hr water deprivation schedule with water available to them only for a half hour in the late afternoon, and on weekends. Therefore, they will be approximately 18-20 hours water deprived when you work with them.
The reinforcer that we will use is saccharin sweetened water (about 0.2%w/v) which is a very effective reinforcer for all rats (reread the first paragraph). These rats will normally work for between 75 and 150 reinforcers per session.
Each lab team will have a plastic cage with a stainless steel wire lid which holds laboratory chow and a 500ml water bottle. The floor of the cage is filled with bedding which absorbs moisture and odors. My assistant and I will take care of cleaning the cages, changing the bedding (at least weekly), and maintaining the food and water supply.
Your main responsibility in the area of animal care will be to monitor your rat's weight. This is to ensure their health and to verify that they are indeed on the water deprivation cycle as specified. The rats should gain weight steadily, with the exception of a 10-20gram drop when they are first placed on water deprivation. They may also tend to be a bit heavier on Mondays, since I like to give them at least a day's free water over the weekends to make sure that they do not run up too much of a water deficit. Rats can tolerate up to 72 hours without water before their lives are endangered; your rats should never go more than 30 hours! The rats should be weighed at least three (3) times each week during the semester &endash; always before running them. A log should be kept of their weight, and the date and time of each weighing. Record your rat's weight to the nearest 5 grams; that is the limit of the accuracy of the scales.
The rats will be marked with one or more rings around their tails, to tell who is who in a cage. Be careful!!! One three-ring rat looks very much like another, and most will not answer to their names. Make sure that you put your rat back into the cage with her (and your) name on it! It's a good idea to re-mark the tail rings every month. We use regular felt markers; some should be in a drawer in the animal room.The rats should not be carried loose. Take the whole cage back into the experimental room. The custodial staff will appreciate this!The most common ailment of rats is a respiratory infection, which can be treated with terramycin or sulfa. If your rat develops an obvious 'bubbling' sound as she breathes, let me know. It is very seldom fatal -- to either rats or students -- if treated.
§ Some final notes:
Weigh the rats before the start of each day's running! Never put the rat into the conditioning chamber until the equipment is turned on, checked out, and functioning properly!
Do not let the rats run around amongst the equipment whilst their sisters are at work! They tend to eat unauthorized items such as recording pens, and wires. Fried rats are of little use, behaviorally!
The last thing in the procedure is to put the rat into the chamber; the first thing to do when you are finished running is to take the rat out of the chamber and put her back into her home cage.
Only then should you turn off the equipment and record your results!
One question that always comes up is why we study animal behavior when our primary interest is people ....
First of all, there are ethical and practical restrictions on what we can do with human subjects.
Second, we can design better experiments with rats since they are simpler, bred for uniformity, and raised under controlled conditions.
Of course, this would mean little if we could not apply the results (with confirmation, of course) to human behavior. The following quotation addresses this topic:ON THE BRAINS OF MICE AND HUMANS
"No category of cell, no particular type of circuit is specific to the human cerebral cortex. The components of our cerebral machinery derive from a stock very similar, if not identical, to that of the mouse. The major event in the evolution of the mammalian brain is the expansion of the neocortex. This growth is accompanied by an increase in the total number of neurons, and thus in the number and complexity of the operations which the cortex can perform. The number of cellular elements per unit of surface area has not changed. The cortical thickness varies, but much less than its surface area. On average, the cortex of man is only three times thicker than that of the mouse, although the increase is not uniform in all layers... The more the surface area of the cortex expands, the more the number of neurons capable of establishing association connections increases... This translates, finally, into an increase in the mean number of connections per neuron,
with a consequent burgeoning of the dendritic and axonal trees, reaching a maximum in man. Nevertheless, the increase in the mean number of synapses per neuron is not directly proportional to the increase in cortical area. Far from it. The density of synapses per cubic millimeter of cortex is of the same order in the rat as in man... At the levels of both the macroscopic anatomy of the cortex and its microscopic architecture, no sudden qualitative reorganization marks the passage from the "animal" brain to the human brain. There is, on the contrary, a continuous _quantitative_ evolution in the total number of neurons, the diversity of areas, the number of possible connections between neurons, and, therefore, the complexity of the neuronal networks that make up the cerebral machine."Jean-Pierre Changeux: Neuronal Man: The Biology of Mind;
Oxford University Press, Oxford 1985, p.66
WEIGHT LOG
(this is for your records; it should not be handed in)
DATE
TIME
WEIGHT
NOTES
OPERANT CONDITIONING OF THE LEVER PRESSING RESPONSE OF A RATYou will of course begin with an Abstract that states the general topic studied, summaries the method and results (defining any technical terms used such as shaping and magazine training, and draws some general conclusions about the effects of these consequences on behavior (not just lab rats).This experiment involves conditioning a female laboratory rat to operate the lever in an operant conditioning chamber.
For your Introduction, look up a reference (an article published in a professional journal) describing experiments on the basic principles of operant conditioning. You should have gotten some background from the text by now, but this is not an acceptable reference. One possibility might be to look at some of the applied journals, such as The Journal Of Applied Behavior Analysis.
This experiment will demonstrate the principle of reinforcement: any behavior which is followed by a reinforcing event will tend to occur more frequently in the future. In this study the reinforcer will be saccharin-sweetened water, delivered by means of a solenoid controlled dipper.
The current experiment will consist of three parts.The first part of the experiment will be the adaption of the rat to the experimental situation; habituating her to the various sights, sounds and smells of the laboratory setting. This will also satiate her for the novelty of the situation and reduce later exploratory behavior.
In the second part -- called magazine training, the rat will be conditioned (taught) to drink from the dipper cup when she hears it operate. This process will also establish the sound of the dipper as a conditioned reinforcer, making possible the immediate reinforcement of other behaviors by the sound of the dipper operation.
In the third part of this experiment we will investigate the shaping procedure.This is one of two basic procedures for conditioning a behavior which is not currently in the organism's repertoire. In the shaping procedure, we change the topography of the subject's response by looking for variations of her current behavior that are closer to the target behavior (pressing the lever) in the sense that they make it more likely to occur. We describe this procedure as reinforcing successive approximations to the desired performance of the lever press. Each time that we reinforce an approximation to the lever press we make that behavior more likely. This in turn makes it more likely that she will emit another variation which is even closer to the target behavior. Thus, we gradually shape the new behavior.
This can be contrasted to the chaining procedure; another way to produce a new behavior (this distinction is not as clear in the Chance text as it could be). Unlike shaping, in chaining all of the elements of the target behavior are already in the subject's repertoire (if not, they would have to be individually shaped). We build a complex chain (sequence) of behaviors out of these elements by prompting them in sequence. The completion of each element in a chain produces a discriminative stimulus (prompt) for the next element in the chain. Primary reinforcement occurs when the chain is completed.
v Before you start magazine training your rat .... v
I've got two computer exercises that I'd like you to do first as practice for this experiment. You'll be asked to comment on them in the Discussion section.
First, I'd like you to go through a simple demonstration where the computer will shape your behavior by teaching you to move the computer's mouse towards a particular location on the screen. This demonstration is on all of the computers in Rm 29.
- This Shaping Demo program can be a bit tricky (it's an old program).
To run the program, doulbe click on the <ShapingRunner> button in the Psychology or IntroBA folder.
When this program runs, click on the arrows to read some instructions for the program.
When you've done this, begin by typing in the five parameter values (they're explained in the documentation provided in the main program). I'd suggest values of five for each -- you can play with it if you want.
- The main problem with this program is that it may end as soon as you start it, and show you some graphs with nothing in them. To avoid this, try locating the cursor in the upper right-hand corner of the box before hitting <a> to run the program.
I'm not particularly interested in the graphs; it's the experience of having the computer shape your behavior that is valuable.
One point: Chance talks abouit shaping being both an art and a science because we cannot state exactly what form the shaping process will take. This demonstration is completely determined by a computer program following set rules. In this sense it achieves the scientific characteristics that Chance discusses.
The CyberRat program is not currently running reliably on the new iMacs in AH 029. I'll have it available on a few other machines; it only takes 5 to 15 minutes to run.
As I demonstrated, run the program, follow the prompts (entering as visitor), and shape your rat by clicking in the 'reinforcer' box when she emits an approximation to the target behavior (pressing the lever).
Please record how many reinforcers it takes to shape the rat, and how much time was required. You'll report this in the Discussion section when you compare shaping CyberRat with the real rat
In a sense we are shaping your behavior by successively approximating the target behavior of shaping a living organism!When you are down with this program, quit it by using the program's pulldown menu. Anything else is likely to hang the computer.
METHOD Subject
The subject for this experiment will be an experimentally naive female laboratory rat. They are usually (check the cage card) albinos. Report the vital statistics for the rat that you are working with only!!
The rat's birth date is printed on the label on the home cage which has your name on it -- you may add the rat's name if you wish. Use the birth date to calculate the rat's age (in days) at the start of the experiment. Your subject's age is more meaningful to a reader than her birth date.
You must not only weigh your subject at the start of each experimental session; you must state this fact in your Subject section. Otherwise, the reader is not sure when the weighing occurred, and she (your subject, not the reader!) will be drinking about ten grams of water a session (see "saturation").
Our rats are allowed free access to food, and will be about 18hrs water-deprived at the start of each day's session. Report what your rat weighs at the start of each day's session as well as the deprivation information.
Since the rat's weight is part of your description of your subject rather than a behavioral outcome, her weight at the start of all sessions should be stated here, to the nearest five grams (5g); the limit of accuracy of our scales.
Apparatus
We will be using a Gerbrands Model G2155 Rat Student Research Chamber with a liquid dipper, a control unit, and a Gerbrands Model G3200 Flat-bed Student Cumulative Recorder. Saccharin-sweetened water at about 0.2% (w/v) concentration will be used as a reinforcer; the dipper cup holds about 0.10ml.
This has been a description of the apparatus that you have used.
With this apparatus you should have used metric-ruled (millimeters or centimeters) graph paper for the cumulative recorder. Since we have been having problems with the commercial graph paper available from the bookstores I've included a sheet in the manual (see the Appendix). Make as many copies of it as you need.
As far as it is possible, use the same sheet of graph paper for all of the sessions in a given experiment. This makes it easier to compare the results for different conditions.
You will also need a pen for the cumuilative recorder. Unfortunately I can't tell you exactly what to buy; it partly depends on the paper that your graph sheet is made of. You need a pen that will write without your adding any pressure other than its own weight -- good ballpoint or fiber point pens will do this. Cheap fiber pens may leave large yucky blotches on the paper, while cheap ball-point pens will not leave a visible line.
At the start of the second session, set the pen about four centimeters above the start of the graph from the previous session. It is desirable to get all of the records on the same sheet of graph paper if it is practical and readable.
Õ Calibration: the recorder pen travels across the paper at the rat of 5mm/minute and is stepped up the paper by lever-pressing responses at the rate of .5mm/response. This means that it takes two lever presses by the rat to move the pen up one millimeter, or about 50 responses per inch.
The above "calibration" information is stated for your benefit in reading and preparing your cumulative records. It should not be stated in your Apparatus section, as this information should be available from inspection of those records. The same applies to the pen and graph paper used.
Final note: on some of the setups external response and reinforcer counters have been added, because the ones on the control box are dead (the equipment is probably older than you are, which is really old in rat years!).
2.Dipper duration control. Clockwise increases the duration to a maximum of about 20 seconds.
3. Response indicator light. It lights when the lever has been completely pushed.
4. Response pushbuttons. They duplicate the rat's lever and rod.
5.Program control: EXP Ð only you can operate the dipper; PARALLEL Ð you or the rat can operate the dipper; SERIES Ð both you and the rat must respond to operate the dipper.
6. This switch should always be set on MANUAL.
7.These switches control whether the stimulus will be a LIGHT or a TONE, and whether it is ON or OFF.
8.Connections for your hand switch. The two snaps at the end of the cord switch should connect the first two (stimulus and gnd) snaps. The second and third snaps should be connected by a jumper. The fourth stud should NOT be connected.
9.Counters to record responses (complete lever presses) and reinforcers (dipper operations). Many of the setups will have external counter units relacing dead conters on the control unit. Reset the counters before each session.
10.Connections to the cumulative recorder. Black to +24vdc, white to response, green to reinforcer.
Procedure
You will be working individually with the rats;
there should be plenty of time during the day for all of you to get at one of the eight conditioning setups. The main advantage of working during the scheduled lab period is that I will generally be available during that time. The lab will be open from 10am to 4pm, except for the times that other classes such as Advanced Behavior Analysis are scheduled.
Normally, training a live rat should take two sessions; one each for magazine training and shaping. We will add a short preliminary session to allow the rats to adapt to the experimental chambers.To start: READ THE INSTRUCTIONS!!Check the dippers carefully. They often slip out of adjustment and need attention or the rat will not be able to drink. Another problem concerns continuous operation of the dippers, such as during continuous reinforcement. This may cause the dippers to freeze in the up position. If the dipper does not drop back down when it should, turn off the power switch on the control box for about two seconds. This should take care of it.
Set up the equipment so that the stimulus light is on (the auto/manual switch should be on manual) and reinforcement is continuously available for lever pressing responses. To do this, set the dipper for an operating time of 15 seconds, with the PROGRAM control on PARALLEL, so that either you or the rat can operate the dipper. Make sure that a snap lead connects the second and third snaps so that the stimulus light is on.All of this should be done before you put your rat into the @#*?$! chamber!!Now -- weigh your rat!
ADAPTATION
Place all of the team's rats into the chamber and allow them to adapt to this new environment, which may be mildly aversive to them at first. If they are afraid of the dipper you can make reinforcement continually available by propping the dipper on the side of the liquid container. Depending on their startle response to the sound of the dipper, you might occasionally operate the dipper with the hand switch.
Leave them in the chamber for about 30 minutes, then remove them for the day.
MAGAZINE TRAINING
You are now ready to start magazine training. Place your rat in the chamber by herself. Make reinforcement continuously available (by pressing the button on the hand switch as soon as the dipper has dropped back into the liquid).
You will know that she is ready to start magazine training when she continues to drink from the dipper even while it is operating.
Set the dipper operating time to about ten seconds. The behavior that you want to condition is for the rat to run to the dipper when and only when you operate the dipper. Therefore, do not deliver reinforcers when the rat's head is already inside of the dipper enclosure. What you reinforce is what you get, and if you reinforce your rat for sitting with her head in the dipper enclosure she will do just that (and nothing else)!!
Deliver reinforcers only when she is actually moving towards the dipper.
Note -- if your rat does not move at all, there are a few possible explanations. The first is that the dipper is not working properly and she is not getting reinforced (check the liquid level, and the screw that attaches the dipper arm to the solenoid). The second is that she has not adapted to the chamber, and particularly to the sound of the dipper operation. If this is the case, I will show you how to give her relaxation training. Finally, you may have inadvertently been reinforcing inactivity by delivering reinforcers while she is standing still, or that you have in fact extinguished activity by not delivering reinforcers frequently enough. Again, I can help you to modify your and her behavior.
Deliver several reinforcers as she is running to the dipper, just before she reaches it -- as her head us just about to enter the dipper enclosure. Remember the results of the reaction time experiments! By the time you can push the button, she may have moved several inches! Therefore, you will have to 'lead' her -- start to operate the dipper before she has reached the point where you want her to be. In other words, you have to anticipate -- predict -- her behavior!
The next step is to gradually reinforce her as she is farther and farther from the dipper, still being careful to reinforce her for moving towards the dipper while the dipper is operating. One technique is to watch the location of her nose in relation to the floor rods as a measure of distance. As you are doing this, gradually reduce the operating time of the dipper as she reaches it more quickly; it should take her only a few seconds to clean out the cup! If she leaves the cup before the dipper goes down, shorten the operating time. Don't trust the numbers on the dial; 'calibrate' the control by timing the actual dipper operation.
When you have completed this procedure, your rat should be thoroughly magazine trained. As an operational definition: she is magazine trained when -- as she is peacefully meditating in the far corner of her chamber -- you operate the dipper, a metaphoric trumpet sounds, she goes three feet up into the air, spins, does a triple somersault and charges immediately to the dipper.
In other words, it should be obvious!
SHAPING
You and your rodent colleague are now ready for the acquisition of the lever response ("barpressing"). This will probably be on the second day of the experiment. From this point on, the cumulative recorder is an absolute necessity, although it is also helpful as a measure of your reinforcer-delivering behavior during magazine training.
Before you start the session, make sure that the recorder is properly set up with an operating pen in the holder, a sheet of metric-ruled graph paper in place, and the pen at the starting position near the motor.At first, reinforce any movement towards the lever (remember -- if you do not reinforce something fairly frequently you will lose control of her behavior). Then, make reinforcement contingent upon her coming closer to the lever, and then rising above it. Nosing the lever from underneath may appear to be a step in the right direction, but it is incompatible with the final target behavior of pushing the lever down, and thus a bad habit for the rat to get into. Require her to first touch the lever with one paw, then two paws, then partially depress it, and finally operate it all by her lonesome. When the lever is depressed far enough to operate the dipper, the white light on the control box will go on and the RESPONSE counter on the control box will increase by one.
You may find that you have to 'shape through' the lever press. These rats have good balance, and are quite capable of resting their paws on the lever without actually operating it. This is sometimes used as a definition of frustration! The trick here is to watch the part of the lever that is outside of the box, and require successively greater movements of the lever, while maintaining the current level of performance through intermittent reinforcement.In shaping, there is a fine line between reinforcing the same behavior too many times, resulting in 'fixation' at that level, and not reinforcing enough, so that you lose control of her behavior entirely! A good compromise is to use a low variable-ratio schedule, and reinforce about every third, fourth or fifth time that her behavior meets the level that you have already achieved, while reinforcing every response which meets your current criterion for reinforcement; the next step towards the target behavior of lever-pressing.
Once the rat has operated the lever by herself (as indicated by the light lighting, the counter counting, and the cumulative recorder accumulating) about a half dozen times or so, shaping has been accomplished and she is ready to go it alone. Reaching this stage is also known as 'reinforcement for the experimenter'! Allow the rat to receive at least one hundred reinforcers for responses. If a lot of reinforcers have been given during shaping, another short session may be necessary.Again, note that much of this section has been "cookbook" descriptions of what you, the experimenter, should be doing to the equipment. Your Procedure section, on the other hand, should describe only what happened to your subject.
Also note: Satiation is defined as the reduction in effectiveness of a reinforcer as a result of the consumption of a large quantity of that reinforcer. Saturation, on the other hand, is what happens to the floor of her home cage after she has been satiated with water!
RESULTS The anticipated result of the first part of the experiment is that the rat will learn to respond to the sound of the dipper by running to the cup and drinking. For each session your report should state:
1) The experimental condition.
2) The approximate time of day and
3) duration (in minutes) of each session,
4) The number of reinforcers given during each session of magazine training as well as
5) The number of responses which operated the lever (if there were any).Present this data in the form of a table. Type this Table into the body of the Results section. One table should describe the results of the entire experiment. Do not make separate tables for the magazine training and shaping procedures.
Your results for the shaping part of this experiment will include the cumulative records of the rat's lever-pressing behavior. Label the axes and add units, as illustrated in Figure 1. Draw the axes on the graph paper so that the zero point (the point where the axes cross) is within a centimeter of the starting point for your first session. The vertical axis is CUMULATIVE RESPONSES and the horizontal axis is TIME (in minutes).
Estimate the response rate for a representative portion of the cumulative record (one without any long pauses and with a consistent rate of response) by dividing the number of responses in that portion of the record by the amount of time that it took the rat to make those responses. Get this information from the cumulative record by making the sample portion of the cumulative record the hypotenuse of a right triangle. Divide the number of responses on the vertical side by the time on the horizontal side (you get these quantities by measuring the horizontal and vertical sides of the triangle and using the calibration data. Use lower case letters to indicate the start and finish of the portion of the record being used to estimate response rate, and use these letters to refer to that portion in the body of your Results sectione.g., "The rate of response was estimated from the portion of the cumulative record between points a and b on the cumulative record in Figure 1").In the example illustrated below, the rate of response would be 80 responses divided by 10 minutes, equaling a rate of 8.0 responses per minute; a respectable rate of response!
Note that there is only one vertical (cumulative response) scale on the cumulative record illustrated below, even though there are two cumulative records on the graph. You might even have three!In general it's a good idea to get as many records as possible on one sheet of graph paper. This facilitates comparisons among the data from different experimental conditions.
Round to one decimal place; I don't think that you could justify any greater precision in your measurement!Again, report the time, number of responses, and number of reinforcers for each shaping session, and the totals for the whole shaping process. Round your results!! As I've already pointed out, stating your results to four decimal places tells the reader that your measurements are this accurate! In other words, if you say that your subject's response rate is 9.4567 responses per minute, you are implying that you can back up the statement that the response rate is NOT 9.4566 responses per minute. Calculators will often give you more decimal places (sometimes referred to as "significant digits") than have any meaning in the real world; this is the difference between science and mathematics!
DISCUSSION
1. Anything worth commenting on? Anything go wrong? Anything go right? If you had difficulties in shaping the lever-pressing response, what might have been the cause?
- If there were equipment problems, in which booth did they occur?
- What specific responses might you have reinforced which might have interfered with the ultimate acquisition of lever pressing?
- In discussing you rat's behavior, avoid the trap of "cognitizing" (talking about what the rat expects, is thinking, etc). If you're not sure why I'm saying this, reread the text to see what "Dr. Cee" has to say! Talk about what you observe. There's plenty happening!
- One problem with talking about what your rat is thinking or feeling is circularity. Since the only evidence that you have for the existence of these private events is your rat's behavior, you cannot use them to explain that behavior. You'd need independent evidence of the existence of thought processes to use them as explanations of behavior.
- Explanations should not be metaphors; they should refer to real observable events. For example, if you observe that the rat shows certain reflexive responses to specific events (such as jumping when the dipper is operated), this is an example of a reflex interfering with an operant behavior, and might explain why the rat did not become trained. Saying that she was nervous or felt afraid would not add anything to this explanation.
2. How did the human shaping exercise compare with CyberRat?
With shaping the live rat?
- Compare the behavior of the CyberRat simulation to that of the live rat.
Is it convincing?
Did it help you in the shaping process?
4. We sometimes talk about superstitious
behavior: behavior that is accidently reinforced while we
are trying to shape some other behavior. Do you see your rat doing
anything which might fit this description?
5. Describe some realworld human situations where behavior
might be maintained by a reinforcement contingency that could be
described as superstitious. Be specific about the
reinforcement contingencies responsible. I'm most interested in
situations other than those commonly described as
superstitions.
6. What has this experiment got to say about the way in which
people acquire new behaviors? Inolve the Law of Effect in
your discussion.
7. Comment on the topic of being sensitive to the relationship between your behavior and someone else's behavior -- the extent to which your behavior is controlled by its effect on someone else's behavior.
Whatever you cited in your Introduction.
Just the photocopy of your reference! The cumulative record as drawn by your rat would be considered raw data, but by the time you have added axes, labels, units, captions, etc., it must be titled and presented in the Results section as a Figure.
A final reminder: Your
procedure section must report what you actually
did!!!
If your procedure
differed from that described in the Lab Manual describe those
differences in the Procedure section. If you think that
they affected your subject's behavior, comment on this in your
Discussion section.
topEXPERIMENT IV: EXTINCTION AND SPONTANEOUS RECOVERY
In the preceding experiment, operant conditioning was demonstrated by shaping a rat to press a lever in an operant conditioning chamber through the procedure of differentially reinforcing successive approximations to the lever pressing performance. In this experiment we will investigate the permanence of these changes in behavior by demonstrating the processes of Extinction and Spontaneous Recovery.
Extinction is defined as the nonreinforcement of a previously reinforced behavior, resulting in a decrease in the likelihood of occurrence of that behavior. This can be seen as a decline (or 'flattening') of the slope of the cumulative record. Spontaneous Recovery is defined as the observed effect that an extinguished behavior will, after a period of time has elapsed, reappear without any additional reinforcement! Hence, the term "spontaneous". Once you have delivered a reinforcer after an extinction trial the return of the previously reinforced behavior is called reconditioning, not spontaneous recovery.
The term extinction is applied here to behavior, not to the organism. When you extinguish the rat's behavior, you stop reinforcing it. When you extinguish the rat, you pick her up by the tail and bat her against the side of the conditioning chamber. I do not expect this to happen !!*@&%!
These phenomena will be demonstrated by extinguishing the rat's lever pressing behavior and then, at least one day later, repeating the extinction procedure without delivering any more reinforcers. If the rat presses the lever at all during this second extinction period, these responses are evidence of spontaneous recovery of the extinguished behavior.
If you must deliver reinforcers to reinstate the lever pressing behavior, then we have demonstrated reconditioning, not spontaneous recovery.
METHOD
Subject
The same subject will be used as in the previous experiment. Give her age, weights, deprivation data, and summarize her experimental history (a reference to your previous report would be appropriate here).
Apparatus
We will use the same apparatus as in the previous experiment . Describe it again here. If your apparatus were a custom built or otherwise unique piece of equipment, you might refer to a previous publication in which it was described completely. Our equipment is generally standard.Procedure
This experiment should require only two sessions.
In the previous experiment your rat was conditioned to lever press and was given one hundred or more reinforcers for lever responses. To test for extinction, place your rat in the chamber with the stimulus light ON and the program control on SERIES, and allow her to receive ten reinforcers for lever presses by holding down the push button on the cord switch. This will be on the side of some switches and the end of others.
If you are doubtful about the thoroughness of her conditioning in the previous experiment, you might want to allow her up to thirty reinforcers for lever presses. At any rate, her response rate when you start the extinction phase of this experiment should approximate that of the previous experiment.
To start the actual extinction phase of this experiment, simply release the button on the cord switch. If the program control was in fact set to SERIES, her lever presses should no longer be reinforced. Leave her in the chamber until a total of forty five minutes has elapsed -- this will be one excursion of the pen across the paper. This will generally be enough time to observe the extinction process. If for some unlikely reason she is still responding strongly at the end of 45 minutes, reset the pen to a point beneath the first extinction cumulative record and continue. I doubt, however, that this will be necessary. You may observe a tendency for the rat to respond during extinction when someone comes into the room, or when some other sudden change in her environment occurs.
When the extinction session has been completed, remove her from the chamber without giving her any more reinforcers and record:
- 1) the number of reinforcers given at the start of the session,
- 2) the number of responses made by the subject during extinction
(once reinforcement has been stopped), and- 3) the time from the last reinforced response before the start of extinction until the
end of the session when the rat was removed from the chamber.
This last may easily be measured from the cumulative record.At least one day later, test for spontaneous recovery by placing the subject in the chamber under the same conditions that were in effect at the end of the extinction session -- the stimulus light ON, and no reinforcement available. To repeat -- do not reinforce any responses!!
Leave her in the chamber under these conditions for another 45 minutes.
Use the same sheet of graph paper as for the previous extinction session.
Start this session with the pen about 5 centimeters above the start of the last extinction session (the start of nonreinforcement) as illustrated below. Again, record the number of responses and the time for extinction. There should not be any reinforcers!
Finally, still using the same sheet of graph paper, set the program control to PARALLEL and recondition the lever pressing behavior by delivering one or two free reinforcers and then waiting for the rat to press the lever. If she does not resume responding within a few minutes, you MAY have to very briefly repeat the shaping process. If you do this, you should not have to use more than half a dozen reinforcers! Allow her to respond for reinforcement until she shows evidence of satiation -- a decrease in the overall rate of response and pauses which become longer as the bursts of responses become shorter. She may also press the lever but not drink from the dipper, indicating that the liquid is no longer an effective reinforcer.
RESULTS
Your Results section should include your cumulative record prepared as in the previous experiment, plus a statement in the text of the Results section of:
1) the number of reinforced responses before the start of the first extinction session,
2) responses and time to the first extinction, and
3) responses and time for the second extinction session (spontaneous recovery).
4) the number of responses and the representative response rate during reconditioning.
Your evidence of the existence of spontaneous recovery will be any responses made at the start of the second extinction session. Since these responses were made after extinction had been completed in the previous session and before any more reinforcers were delivered to recondition the performance, this recovery of the extinguished responses may be termed "spontaneous".
DISCUSSION
1. Looking at the form of the first extinction curve, what was the immediate effect of instituting the extinction condition? Compare this with what Chance describes as the typical effect of extinction on behavior.
Compare the first and second extinctions, taking into account both the overall response patterns and local details.
- Again, can you find any places on the cumulative record where the rate of responding increased or decreased due to some environmental event?
2. What might account for the spontaneous recovery of the extinguished lever pressing behavior? Or the lack of it? Remember, I'm looking for a behavioral analysis!
3. Compare your subject's response rate after reconditioning with that from the CRF experiment. Again, your Discussion should include the Law of Effect.
4. How does the process of extinction differ from forgetting (check the text on this).Can you say that extinction 'removes' behavior from your subject in any sense?5. How might you set up a rat experiment to test for forgetting (it's really quite simple)?
6. What implications might these observations have for the application of extinction procedures to change human behavior? Talk about spontaneous recovery and the ease of reconditioning.
- What practical problems might this present in attempting to use extinction to eliminate a human problem behavior?
- What might be some other ways in which you could decrease the frequency of a behavior?
7. Were there any discrepancies between your procedure and the one in the manual? If so, what was the effect on your results?
REFERENCES Your previous report if you referred to it in this report (refer to it by author, title, and "unpublished manuscript").
APPENDIX Data sheets, etc.
Last week's experiment was concerned with continuous reinforcement; either every response was reinforced, or none were.
This week's experiment will use intermittent reinforcement to study the ways in which patterns of behavior will change to fit the current reinforcement contingencies (behaviors required for reinforcement under a given set of conditions).
To investigate this phenomenon we will use the fixed interval schedule of reinforcement. On a fixed interval schedule, two requirements must be met before a reinforcer is delivered:
First, some specified duration of time (interval) must have passed since the last reinforcer was delivered, and
Second, a response must have been made by the subject after the interval has ended.
There is no time limit set for how long it takes the subject to respond once the interval has ended and reinforcement is available.
The length of the interval determines the maximum rate of reinforcement which the subject can obtain. Because of this, interval schedules typically produce a low rate of response -- the longer the interval the lower the rate. In addition, since responses immediately following the reinforcer are never reinforced, the probability of the subject emitting a response increases as the time scheduled for reinforcer availability approaches (as reinforcement becomes more likely, responses become more frequent). This produces the fixed interval scallop; a cumulative record pattern characterized by a pause after reinforcement which lasts more than half of the interval, followed by a gradual increase in response rate as the time for reinforcement of a response approaches.
Interval schedules are said to have strong 'recuperative' powers, since if the subject stops performing, the end of the interval will allow the next response to be reinforced.
The fixed interval scallop may be quantified by using a measure termed the INDEX OF CURVATURE. This is defined as the proportion of the total number of responses which are made in the last quarter of the intervals -- including the reinforced response! Again, the microcomputer will help us with the calculations.
As an independent variable, we will examine two fixed interval schedules: FI-60sec and FI-30sec, and compare the patterns of behavior maintained by these two schedules, as well as the transition from FI-60sec to FI-30sec.
METHOD Subject
We will use the same subject as before, with additional experience. Provide the necessary information.
Apparatus
As before, plus some form of timing device. We have a number of stopclocks of various types which are particularly well suited to this type of procedure since they may be reset while they are running. Check the clock that you use for the manufacturer, as well as for the model number (if any).
In particular, it is useful to know whether the timing device you used was analog or digital.
ProcedureThis experiment may well take as many as four or five one-hour sessions That is why it is scheduled for two weeks. Most of this time will be spent establishing the stable Fixed Interval 60sec performance that we will use as a baseline. We will be observing a gradual change in behavior over time!Assuming that typical CRF (continuous reinforcement) performance has been recovered after the extinction sessions, we may proceed with the FI-60sec schedule.
As in the previous experiment, use the SERIES control setting to control the availability of reinforcement, and have the stimulus light on. Again, you should get as many cumulative records as possible on one sheet of graph paper. Depending on the rat's response rate, either telescope the records vertically (start the second record above the first one) or horizontally (start the second record underneath and to the right of the first record).
Start by allowing the rat to lever press a couple of times for reinforcement on CRF, and then move quickly up to FI 60sec through a sequence of 10, 20, 30, 45 and 60secs;.
A total of a dozen reinforcers should get you there!
With any luck (or, without any mistakes) you might begin to see typical FI behavior by the end of the first day's session. If not, continue with additional sessions on FI 60sec, until you have a period of at least 15 intervals that look something like the mature FI behavior illustrated in Figures 1 and 2 below.
Stable performance on fixed interval schedules is characterized by an Index of Curvature (to be explained later) greater than 0.50. This level of performance is the minimum REQUIREMENT for the satisfactory completion of this experiment. Do not go further until you and your rat have achieved this! It would be better to get a value greater than .66.
Once you have gotten a period of stable performance on FI 60sec (see the recording procedures below), change the schedule directly to FI 30sec. The transition to a typical scalloped behavior pattern under this new schedule may take a while, since at first the subjects may not respond at all until after the end of the intervals. This is particularly likely if her pauses on the longer interval have been longer than the total length of the shorter interval.
The best way to observe this transition is to stabilize behavior on FI 60sec, and then start another session on the same schedule. As soon as you have 15 intervals of good FI behavior, change the schedule to FI 30sec without removing the subject from the chamber. If you're lucky, you may get all three conditions in the same session and on the same sheet of graph paper!There will be two basic measures of behavior for this experiment.
The first will be the same overall response rate that we used in the previous experiments, which can be calculated directly from the cumulative records. Use your own judgement as to which portions of the record ought to be selected.
The second measure will be the Index Of Curvature, as defined in the Introduction. In your report, do not assume that the reader is familiar with this measure; define it in a sentence the first time that you report it, both in the Abstract and in the Results.Instructions for calculating this index can be found in the Appendix. Since there is now a program for the Macintosh computer which will do the actual calculations for you, the main thing is to collect the appropriate data. This means that once behavior appears to be stabilizing on FI-60sec, you should start using a data sheet which has columns for:
- 1) the first three quarters of each interval,
- 2) the last quarter of each interval,
- 3) the interval duration, and
- 4) the number of the interval, as illustrated below
You must record data for each interval separately!!
INTERVAL
SCHEDULE
FIRST 3/4
LAST 1/4
1 FI 60sec ///// // (7) /// (3) 2
"
/// (3)
///// (5)
3
"
//// (4)
// (2)
----
----
----
----
15
"
/ (1)
///// /// (8)
You should have this data recorded for at least:
1. The last 15 intervals on FI-60sec,
2. The first 15 intervals on FI-30sec, and
3. The last 15 intervals on FI-30sec.Thus, you can compare mature behavior on FI-30sec both to mature (stable) behavior on FI-60sec and to the transition from FI-60sec to FI-30sec.
This procedure is similar to data recording procedures that are used with human subjects in applied settings.Note: I've put this description of data recording and analysis in the Procedure section to make sure that you record the necessary data as you run the experiment. You should report the data and indices used in your Results!
RESULTS
The heart of your Results section will again be your cumulative records, with appropriate embellishments.
Note: Figure 1 is an illustrated example. Your Figure should contain the entire session, and be prepared and presented as in the previous experiments.In addition, calculate representative response rates for the three experimental conditions:
- 1. Stable performance under FI-60sec,
- 2. The beginning of the FI-30sec condition (at least 15 intervals),
- 3. Stable performance under FI-30sec (at least 15 intervals).
The response rates which you calculate for the three experimental conditions should be based on the same part of the cumulative record as the Indices of Curvature. These sections should be indicated on the cumulative record in the same way as on your previous reports.
If you have sheets of records from the initial FI training which are not referred to in your Results section, they should be in the Appendix rather than in the Results.Do not include satiation effects in your calculations! If the rat appears uninterested in the reinforcer and response rate drops at the end of a session, do not use these data for your final analysis.
Finally, calculate the Index Of Curvature for the same three conditions. For each sample period of at least 15 intervals, enter the number of responses in the first three quarters of each interval and the number of responses in the last quarter of each interval -- including the reinforced response. The computer will total the number of responses in the intervals in the sample, and divide the number of responses in the last quarter by the total number of responses in the intervals sampled to get the Index of Curvature. See the Appendix for a more detailed description of this measure of Fixed Interval reinforcement schedule performance.
Present your data in the form of a Table, as illustrated below:TABLE 1
session and response index of
schedule duration responses reinforcers rate curvature
1
60min
650
61
10.8r/m
na
2
120min
1080
110
9.0r/m
na
3
93min
610
90
6.6r/m
.58
4a (mature FI 60)
15min
80
15
5.3r/m
.68
4b (early FI 30)
7.5min
30
15
4.0r/m
.73
4c (mature FI 30)
7.5min
40
15
5.9r/m
.65
Again, the cumulative records for all of the sessions listed in the Table above should be included in your Results section; completed as shown in Figure 2.
Preliminary sessions not included in the Table should be in the Appendix.
DISCUSSION 1. Again, comment on your Results. You have three parameters of change: visual inspection of the individual interval patterns on the cumulative records, the overall response rates calculated from those cumulative records, and the indices of curvature. Do all three show the same effects of your three experimental conditions?
2. Compare the behavioral effects of the fixed interval schedule of reinforcement with the effects on behavior of continuous reinforcement. What might account for the differences? Discuss both the overall response rates and changes in the patterns of responding.
3. Do you think that most human behavior is in fact maintained by continuous reinforcement, or by intermittent reinforcement? Support your answer.
4. Now, compare your results with the patterns of responding that you would expect from a fixed ratio schedule of reinforcement.
- HINT: on a ratio schedule, each response makes reinforcement more likely to occur -- a higher rate ofa response results in a higher rate of reinforcement. However, like a fixed interval schedule, the first response after reinforcement is never reinforced on a fixed ratio schedule.
5. What might be an example from your personal experience of a human behavior that typically involves a fixed ratio schedule of reinforcement? (I've seen the Subway example too many times -- please find another one ;-).
6. Discuss the implications of these differences in schedule behavior for changing the frequency of human behavior. Include a discussion of the effects of a history of intermittent reinforcement on the resistance to extinction of a particular behavior.
REFERENCES
APPENDIX
Your raw data for calculating the indices of curvature, as well as cumulative records of preliminary sessions whose data is not referred to in the text of the Results section.
topSUPPLEMENTARY INSTRUCTIONS FOR CALCULATING THE INDEX OF CURVATURE
One way to describe an "average" fixed interval would be to calculate the average number of responses occurring in each portion of that interval. Thus, we would figure out, on the average, how many responses occurred in the first quarter, the second quarter, the third quarter and the fourth quarter of the intervals by recording the number of responses made in each quarter of each interval, and then averaging them for each quarter. We could then plot a graph which would be an idealized picture of a 'representative' fixed interval. If we wanted more detail, we could break the intervals down by tenths instead of quarters! However, the recording involved would be rather time-consuming, and it would still be difficult to compare a number of different conditions.
We can simplify the procedure by concentrating on the last quarter of the intervals; the period when most responses should occur if behavior has in fact stabilized! The more clearly defined the shape of the scallop, the greater the proportion of responses which should be occurring in the last quarter -- the end of -- the intervals!
Therefore, we can use the Index Of Curvature, which states the proportion of the total number of responses which occur in the last quarter of the intervals.To calculate this index:
1) For each interval, record the number of responses in the first three-quarters of that interval (e.g., the first 45 seconds of a 60 second interval).
2) When you have completed the number of intervals for which you are calculating the index of curvature, figure out the total number of responses for all of these intervals combined (fifteen is usually a good sample size). This may be done either by recording the number on the response counter at the beginning and end of the sample period, or by measuring the vertical distance on the cumulative record for that period, or by adding up the totals on your data sheet. You will now have the total number of responses made in the period for which you are calculating the index of curvature.
3) Find the total number of responses made in the last quarter of the sampled intervals, either by summing the last-quarter column on your data sheet, or by summing the first three-quarters column on your data sheet and subtracting that sum from the total number calculated as above.4) Calculate the Index of Curvature by dividing the number of responses in the last quarter of the intervals in the sample period by the total number of responses made by the subject during the sample period. Round to two decimal places.
RESULTS OF STEP #3INDEX OF CURVATURE = (divided by the)
RESULTS OF STEP #2As the number of responses in the first three quarters of the intervals decreases, the difference between the total number of responses and the number of responses at the end of the intervals gets smaller, so that their ratio approaches 1.00. Therefore, the better defined the interval, the higher the value of the Index of Curvature, with 1.00 as an upper limit.
On the other hand, if the subject distributed her responses equally among the four quarters of the interval (as illustrated in the developing FI part of the sample cumulative records in Figure 1), the value of the index would be 0.25. A value of the index of curvature less than 0.25 would indicate that most of the responses were occurring at the beginning of the intervals, as in the early FI record.Note: If you are using the computer and you have already summed the responses for the sample period, treat the number of intervals as one and enter the summed data as requested by the computer.
topEXPERIMENT VI: BEHAVIOR UNDER STIMULUS CONTROL
For this last assigned experiment we will examine the way in which behavior can be brought under stimulus control. Stimulus control refers to the observed effect that behavior occurs more frequently in some situations than others. Typically, this is the result of reinforcement contingencies; the behavior is more likely to be reinforced in some situations than in others. Stimulus control is another example of 'what you reinforce is what you get'. A behavior is more likely to occur in situations in which it has paid off in the past.In our experimental situation, we will use the lever-pressing behavior that by now should be thoroughly conditioned. Up to this point we have been concerned only with the frequency with which the behavior occurs in a standard situation (the operant conditioning chamber with the light on) under various contingencies of reinforcement. We did find out that the rats can tell when a behavior is likely to be reinforced, and that therefore that behavior is more likely to occur at those times.
We will now extend this principle to the observation that when we reinforce a behavior in the presence of a stimulus, and systematically withhold reinforcement when that stimulus is not present, the behavior will be more likely to occur in the presence of that stimulus than in its absence. We say that the rat discriminates light from dark.
By now it should be obvious that the stimulus that we will use will be the light in the operant conditioning chamber.
One procedure that we will be using is called DRO (Differential Reinforcement of Other Behavior). That is, we will deliver reinforcers when a specified behavior has not occurred for some period of time.
Originally, this term was interpreted as DR0 (D R zero). This was an extension from the term DRL (differential reinforcement of low rate). The idea was that if one could reinforce a low rate, then one could reinforce a rate of zero. After a while, behavior analysts realized that there was a major proplem with this usage -- the transition from 'low' to 'zero' was more basic than had been realized.
Note: Paul Chance and I have disagreed on this history. We've discussed it, and the next edition of the text may contain a modified account.When you reinforce a low rate of response, reinforcement is contingent upon the occurrence of the specified response.
However, in a DRO procedure, reinforcement is not contingent upon the specified response; it is contingent upon anything but the specified response. You always reinforce something. Any living being will be doing something just before a reinforcer is delivered!
METHOD Subject
The same subject will be used as in the previous experiment. Provide vital statistics, including history.Apparatus
Likewise.Procedure
This experiment should not take more than one or two sessions, since your subject should already be under a degree of situational control from the previous (fixed interval) experiment. We will build on this control by starting with an FI-60sec schedule, with the stimulus light ON at all times. Run the rat on this schedule until you see evidence of scalloping (an Index of Curvature of at least .50 if you want to get quantitative). This should not require a whole session!Now, we will turn the stimulus light OFF and turn it ON only when a lever press will be reinforced, at the end of a 30sec interval (note the change in interval duration).The easiest way to do this is to set the program control on Series, remove the snap lead from the stimulus and gnd studs, and connect the Stimulus stud to the Reinforce studs on the control box with a snap lead. Once you have done this, holding in the switch on the side of the hand control will both turn the stimulus light on and make reinforcement available. Wait until the subject responds and is reinforced, and then start the next interval.
It is likely that under these conditions the subject may not show any more stimulus control than during the simple fixed interval schedule. This is because lever presses will be occurring right at the end of the interval (with the light off) and will be immediately (a second or so) followed by the light coming on and a response being reinforced.
To separate the reinforcement of lever presses from the light on condition, you should add a DRO contingency (Chance, p 237). This is a requirement that there be an interval of at least five seconds without a lever press before the stimulus light is turned on. In other words, if your subject makes a response in the last five seconds of the interval, wait until the subject has gone at least five seconds without responding before turning the light on. This will guarantee that lever presses in the dark will be separated by at least five seconds from reinforcement. If the rat continues to respond at the end of the interval despite the DRO you might try lengthening it to 7 or 10 seconds.In most cases, this will result in a very low rate of lever pressing with the light out and a high rate of lever pressing with the light on. The cumulative record should show a flat line with reinforcers at approximately thirty second intervals.
If you don't get good stimulus control after a couple of hours, you might try turning the overhead lights off and just using the lamp in the room for illumination.The DRO contingency it is an important part of your Procedure and must be mentioned (and defined) in your Abstract.
RESULTS
The results should be obvious; you will appear to have an 'automatic' rat! When you turn the light on, the rat presses the lever. Thus, the lever pressing behavior is under Stimulus Control, and the subject can discriminate light from dark.
As evidence, you should have cumulative records showing a pattern characteristic of a behavior under stimulus control: a nearly horizontal line with a response step and a reinforcement mark at 30sec intervals.
Beyond this, the best evidence of stimulus control is the relative rate of response in the presence and absence of the stimulus. To calculate this Discrimination Ratio, record the time actually elapsed in each interval with the light off, the number of responses with the light off, the time elapsed with the light on, and the number of responses with the light on. Do this for a representative sequence of a group ofat least ten consecutive intervals after you (and your subject) have achieved stimulus control.
- From this raw data, calculate the response rate with the light off. Divide the number of responses with the light off by the total time with the light off. Then, do the same for light on.
- Finally, calculate the Discrimination Ratio by dividing the light-on response rate by the light-off response rate. If the subject responds more rapidly with the light on, this ratio should be greater than 1.00. Good stimulus control should produce a discrimination ratio of at least 10.00; this is a requirement for the satisfactory completion of this project.
If your rat appears to be under good stimulus control but the discrimination ratio doesn't seem to show it; check your calculations! If the rat has not in fact achieved a discrimination ratio of at least 10.0, then you must run another session.
If your subject emits no responses when the light is off you can't calculate a Discrimination Ratio, since its denominator would be zero. Simply report that stimulus control was perfect, and give the Latency (the average time between the presentation of the stimulus and the response) which in this case is equal to your light-on response rate.
DISCUSSION 1. As usual, discuss your results in terms of directly observed behavior, your cumulative records, and the numerical indices that you have calculated.
- What was the most obvious difference from the pattern of behavior maintained by the FI schedule of reinforcement?
2. How might you modify this procedure to include errorless discrimination training?
Review what the text says about this procedure.3. How might you modify this experiment's procedure to test the rat's reaction time?
4. Relate your results to the real world of human behavior.
In what ways has your behavior been brought under stimulus control by this course?By your rat's behavior?
- Make sure that you are talking about stimulus control, not simply control by reinforcement contingencies.
5.Reread the introduction regarding the meaning of DRO.
If one did try to reinforce "nothing" what would probably happen?What would the actual contingency be?6. Discuss whether a low rate of response necessarily means that a reinforcer is weak or ineffective. Again, the key concept is contingency!
REFERENCE
APPENDIX As usual.
topCHECKLIST FOR INDEPENDENT RESEARCH PROJECTS
Note: This project must be an experiment; that is, you must actively change something (the experimental variable) in the environment and observe and record the corresponding change (or lack of it) in some specified behavior under both conditions, where you are reasonably certain that nothing else has changed!Note also that for a a grade of 'A' for the course you must complete either the last text unit or this project.
To insure that subjects are treated appropriately, and that you have chosen a project that has a reasonable chance of being completed in the allotted time, it is required that you hand in a written proposal for your independent project before you start collecting data! Your proposal should be based on the outline below .
I'll return the proposal to you with comments and suggestions. A corollary of Murphy's Law states that there are more wrong ways to do something than there are right ways. I might save you some time and effort!
The text discusses the ethics of changing human behavior. Ethical considerations also apply to research done with nonhuman subjects.A thorough discussion of animal research ethics is beyond the scope of this course. Suffice it to say that when the purpose of a project is teaching rather than the advancement of scientific knowledge, no procedure that might cause harm or discomfort to a subject is acceptable. Even if we were doing publishable research, it would be incumbent upon us as researchers to demonstrate that any potential risk to subjects was justified by the knowledge to be gained. There are some who would go so far as to say that no risk to a research subject can be justified under any conditions. This presents a great challenge to the skill of the researcher!
NOT ALL psychological research is experimental in nature.
Introduction
Summarize the basic theoretical issues involved. Why are you doing this experiment (other than the fact that it is a course requirement)? Has it some relationship to a basic process in which you are interested? Some practical application? Simple curiosity of the "What would happen if....?" variety is quite scientifically respectable!
Since I don't know what you are up to, I will need a full lab report. However, if you have already met the '3 A' requirement I will grade it on an OK/Rewrite basis.
METHOD Subjects
Who or what? Where will you get them? Describe them appropriately!
Apparatus
What will you need, and where will you get it? In your report, you must provide enough information so that the reader could get equipment that would perform the same functions.
Procedure
Summarize the basic operations to be performed. What will happen to each subject? You must provide enough information in your report so that the reader could replicate (repeat the same way) your experiment. Please make sure that your project is an experiment; that you have actively done something to affect your subject's behavior!
If you are using human subjects you might want to consider a crossover experimental design, since it would be difficult to get enought subjects to have valid experimental and control groups.
Basically, this is the procedure that we used in the first two experiments, where each subject is tested under both experimental and control conditions. Split your subjects into two groups. Give one group the control condition first, then the experimental. Do the opposite for the second group.
RESULTS What types of data should the procedure as described above provide? What will you measures be? How will data be presented and analyzed? Do you have any hypotheses about the outcome of this experiment?
DISCUSSION Do you foresee any problems in running this experiment? What possible alternative explanations for your results might exist that would not be controlled by your procedures? What might be some of the implications of the results of your experiment?
REFERENCES Your sources of background information on the topic of your experiment. This is not necessary if you have already met your 3-A requirement.
APPENDICES Your raw data, work sheets, etc.
May be freely circulated and copied, but may not be sold for profit.
webmaster: pkbrandon@mnsu.edu
