SB – 10

Control of Variables
1/25/18

Related Humor If you drink a lot of scotch and soda, bourbon and soda, gin and soda, and you get drunk in all situations, it must be the soda that causes it! How many other conclusions about human behavior are based on the same error?

The topic of control in experiments will be presented in two categories — group experiments and individual experiments. Some general comments on control of variables can save repetition under both of these labels. To see what effect a certain variable has on another, control of all relevant variables is necessary. In doing a study of the effects of light intensity on productivity, if you also vary the amounts of attention given to the subject, it is not possible to decide if the intensity of light or the attention had an effect on productivity. Since both variables changed, either or both or some feature of each could be the causative factor. Control of variables permits keeping irrelevant variables from influencing the data being studied. Many times variables which are not effective alone may affect the results when interacting with other variables. Control techniques can be used to clear an experiment of an interaction-type error.

The essence of experimentation is control. Experiments can be viewed as controlled observations. Research in natural settings many times obscures the effects of experimental or independent variables. Greater precision of the experimental observation requires controls.

Physicists do not study gravity by studying falling leaves, nor do physiologists or biologists study the speed of nerve transmission by studying reaction time. Behavioral scientists also require the use of some controls.

10-1. The essence of the experimental method is the control of conditions of observation.
A. True
B. False

Answer. (A)

10-2. The experimental method attempts to control extraneous conditions in observation.
A. True
B. False

Answer. (A)

10-3. The role of the controls in an experiment is to keep irrelevant variables from influencing the data.
A. True
B. False

Answer. (A)

10-4. Good experiments involve the attempt to hold some variables constant while allowing the others to vary. This is known as the use of _______ .
A. control
B. variability
C. flexibility
D. constancy

Answer. (A)

10-5. Good psychological experimentation involves _______.
A. measurement
B. variables which are controlled
C. the study of variables
D. careful design that includes controls
E. All of the above

Answer. (E)

10-6. The experimental method is preferable to other methods in psychology because _______.
A. it cannot be used to study problems
B. it affords greater precision and control
C. the effects of many different independent and dependent variables can be studied simultaneously and forever
D. it has fewer advantages

Answer. (B)

10-7. Important characteristics of the experimental method include _______.
A. observations
B. control and systematic manipulation of the conditions
C. popularity (everyone likes it)
D. its beauty

Answer. (A,B)

10-8. One of the major difficulties of research in natural settings is that other variables may obscure the effect of the independent variables.
A. True
B. False

Answer. (A)

10-9. In psychology the distinction between the experimental method and other methods lies in the _______.
A. number of naturalistic observations
B. use of expensive apparatus
C. control of variables
D. theoretical issues involved

Answer. (C)

10-10. Physicists do not study gravity by studying falling leaves. Physiologists do not study speed of nerve transmission by studying reaction time.
A. True
B. False

Answer. (A)

“…the physicist who knows what events determine how a body will fall through space is in a position to predict the course of the fall of a specific body only if he is given access to information about the determining events in this case. It is for this reason that the trained physicist may be as helpless in predicting the flight of a stray feather through the air as is the layman. The layman knows neither the determining events nor the laws relating these events to the flight of the feather. The physicist knows laws relating the events, but without information concerning the occurrences of the actual determining events he is just as powerless as the layman to make predictions.” (Staats, Arthur W. and Staats, Carolyn K., Complex Human Behavior. Holt, Rinehart. & Winston, New York, 1963, p. 18)

If you tear up a tissue and let it fall, it will not fall directly down. Does this defy the law of gravity? Or does the account of this apparent defiance simply reflect the relevance of other variables such as air pressure?

One may not be able to blow out a candle from ten feet away. It would not appear wise to say the candle does not follow lawful scientific principles. When one has poor control of the variables that control a person’s behavior, s/he is as ineffective in controlling the person as is the individual cited above in controlling the candle’s behavior.

Research Employing Groups

10-11. In a group experiment, the scientist typically has two groups of subjects, an experimental group and a control group. The experimental group is presented the independent variable while the control group is subjected to the same condition but not presented the independent variable. As a result, the two groups may be compared in order to determine _______.
A. the effects of the independent variable
B. the effects of the dependent variable
C. the effects of the experimental group on the control group
D. the effects of all of the variables

Answer. (A)

A control group is necessary in order to have a base from which to judge. Unless one has a basis for judgement, he/she cannot determine the effects of the independent variable.

10-12. A judge selected delinquents and, in 12 of the 14 cases, found parental rejection, then concluded that, in over 80 percent of the cases, rejection is the cause of delinquency. Conclusion: _______ .
A. is warranted in the case where no careful investigation was made
B. is correct because in 9 cases no rejection was present
C. is correct because it confirms his hunches he just knew were correct
D. cannot be accepted until one establishes that less parental rejection occurred in a control group of non delinquents

Answer. (D)

10-13. Group experimentation in psychology and other sciences requires consideration of _______.
A. experimental and control groups
B. concomitant groups
C. graphical presentation
D. operant operationalism

Answer. (A)

10-14. Which one of the following does not belong with the others?
A. clinical method
B. control group
C. experimental method
D. dependent variable

Answer. (A)

10-15. In an experiment, a control group of subjects is the group to whom only the independent variable is applied.
A. True
B. False

Answer. (B)

10-16. If the hypothesis is that brand Y toothpaste can reduce the incidence of cavities and if the other variables are methods of brushing and diets, what variables should you keep constant in your group?

Answer. (Brushing & diet)

10-17. In psychology the distinction between the experimental method and other methods (e.g., naturalistic observation, clinical judgment) lies chiefly in the degree to which variables are _______.

Answer. (controlled)

Individual Organism Research

Sidman breaks down control into statistical and experimental control. Experimental control is that which is used in individual organism research. In this approach, when a variable is manipulated and the results occur in the predicted direction, you experimentally study what happened.

The method of individual organism research has been characteristic of an approach to behavioral study referred to by various names, such as a functional analysis of behavior, an experimental analysis of behavior, operant conditioning, or a Skinnerian methodology. This individual approach does not use large groups of subjects in statistical analysis. This approach contends that statistics with large numbers of subjects have too often been used to cover up lack of experimental controls. Sidman and others maintain that at times statistics have been used to tease effects from data. With adequate controls a single subject should be sufficient.

When large amounts of data from a single animal are collected under stringent controls, the results will be reliable, and no statistical techniques are necessary. (Sidman, Murray; Tactics of Scientific Research, Basic Books, 1960.)

10-18. Sidman contends that statistical techniques at times tease the desired effect from data.
A. True
B. False

Answer. (A)

If you are going to show that the procedure you introduced has an effect, you must show that the variable you manipulated influences the rate of responding. You need a control condition referred to as a baseline, or operant level. A rat put in a Skinner Box will have a low rate of lever pressing as the rat wanders around the cage. After the rat is given food pellets following pressing the lever, level of lever pressing will increase. If you want to increase the amount of time a child spends in contact with other children on a playground, you first count the number of contacts the child has before you introduce your attempt to increase the social behavior.

A study by a group of individuals at the University of Washington illustrates reinforcement with a child from the nursery school. After obtaining the operant level, they increased the time spent in contact with other children by paying attention to the child only when he was in the proximity of the other children. Before they could demonstrate that this procedure worked, they had to show what the baseline was. Recording of data in most individual organism research is by cumulative record. This is a graph in which events (usually responses) are cumulated and plotted as a function of time, e.g. a total of 15 responses emitted in 1 minute. The amount of something per unit of time defines rate. Time is listed on the chart along the bottom of the graph. Technically speaking, this horizontal axis is called the X axis, or abscissa. Usually the independent variable is plotted on it. The most prominent advantages of recording by cumulative record are contained in the memory device CAD (Continuous Automatic Direct).

Mnemonically speaking, one might use T B A HIX (to be a hick)

Time
Bottom of sheet
Abscissa
Horizontal
Independent variable
X axis

The other side, the one on the side of the sheet, is designated as follows: OY VD BS (think of it as the laments of the sexually promiscuous)

Ordinate is its technical name
Y axis is its letter designation
Vertical is its direction
Dependent variable
Behavior is plotted in the usual operant paradigm
Side of paper

On the chart used in this unit for illustration, percent of time interacting with a child is placed on the ordinate (axis on side of paper). In phase 1 you see a baseline (or, technically, an operant level). You see that very little interaction is taking place. This is the existing condition at the start of the program.

In phase 2 the teacher begins to attend to the child when the child interacts with another child. The percentage of time spent in interacting increases. Phase 3 is a repetition of phase 1 with the conditions of phase 2 reversed. This is intended to show that the variable manipulated is the one that is actually having the effect. This change in the condition shows that time itself could not have caused what is taking place. If the condition is not reversed, one has no evidence that the passage of time was not responsible for the change noted.

In phase 4 conditions are again reversed, this time in terms of reinforcing interaction. Again one notes an increase in interaction behavior.

 

10 – 19. Complete the mnemonic device OY VD BS
__________________________________________
__________________________________________
__________________________________________
__________________________________________
__________________________________________
__________________________________________

(Ordinate Y axis Vertical Dependent Behavior Side)

Individual and Group Procedures in Reference to the Clinical- Science Issue

Some individuals state that science and clinical practice are in disagreement, but that is not the case; there are simply different ways of conducting scientific studies. Group methods frequently are not useful for clinical practice because in clinical work you are treating a patient, not all patients. The two major approaches to scientific investigation are statistical techniques and control techniques. These techniques differ procedurally. In a statistical, predictive approach the scientist attempts to determine the interaction of variables by using statistical techniques, such as analysis of variance. In such a study of fertilizer, for example, the two groups of plants get varying degrees of different variables, such as sunlight, water, heat, ground condition, cultivation, wind, shade, etc. The experimental design attempts to balance the effects of these variables. Control techniques of the same problem would use two seeds and control all variables except for the fertilizer. (Summary from a speech by Goldiamond at St. Cloud State)

Related Humor A patient in bad shape in the hospital’s burn ward hears the instructor of the TV-taught chemistry course say: “Yesterday, students, time ran out just as we were about to learn what happens when magnesium is heated to its kindling point…” “I never have any fun in school… I’m always in the control group…”