In the four group six study design, it is possible to determine the effect of the experimental variable from only two groups can be seen experimental group 2 and control group 2. Referring to the summary table of the factors affecting the results the results in that design, one can see that the difference between the before and after measurements of experimental group 2 is made up of the effects of the experimental variable and controlled variables; the difference between the before and after measurements of control group 2 is the result of uncontrolled variables. Since the before measurement in both these cases was inferred, it would be the same in both instances. Therefore the effect of the experimental variable can be determined simply by computing the different between the after measurements for the two groups (x3 – y3)
This raises the questions: Why include the other two groups in the experimental design? The answer for the scholar is that the four groups and six studies enable the experimental variable to be studied under different conditions, permits the study of individual respondents who change, and encourage better methodology to be developed. To the average business executive, these may not be compelling enough reasons to sustain the added expenses. Therefore the after only with control group design becomes a logical modification. Would appear as:
Experimental Group Control group
Before measurement No No
Experimental variable Yes No
After measurement Yes (x1) Yes (y1)
Effect of experimental variables = x1 – y1
The experimental and control groups are selected in such a way as to be equivalent. No before measurement is made in either group. The effect of the experimental variable is determined by computing the difference between the two after measurement (x1–y1). Notice that this design escapes the problems of pretest effect and interaction. Compared to the four group-six study design this two group – two study design is much simpler to administer and much less expensive. It is not surprising that it is by far the most widely design in marketing.
In a classic after only with control group study determine the image of the homemaker who uses instant coffee, the following design was used. Two comparable groups of consumers were shown similar shopping lists and asked to describe the home maker who prepared the list. On the list shown the control group, one of the items was Maxwell House Coffee (drip grind). On the list shown the experimental group, this item was replaced by Nescafe Instant Coffee. The results measured were the percentages of the respondents who described the shopping list author as having various characteristics. The effect of the experimental variable (Nescafe Instant Coffee user) was the difference in the percentage ascribing each characteristic to the instant coffee user from the percentage ascribing the same characteristics to the drip grind user.
No problems of before measurement effect were encountered because no pre-measurement were made. Uncontrolled variables such as history and maturation influenced both the experimental and control groups to the same degree.
On one basis the after only design is at a disadvantage relative to the before after design. The before after design permits an analysis of the process of change whereas the after only design does not. Thus, individual respondents can be identified and their reactions noted in a before after study. For example in an attitude and opinion study one can measure the effect of the experimental variable on those people who have favorable attitudes as contrasted with those who had unfavorable attitudes in the before measurement.
The after only with control group design fits many marketing problems and is easy to use. Many promotional devices can be tested this way. A dry milk company believed its biggest problem was to get consumers acquainted with its product. Therefore, it put most of its promotional money into sampling campaigns, but it had no real knowledge of their effect. An after only experiment was devised whereby the experimental group was given samples of dry milk. Then the experimental group and the control group were both sent coupons for purchase of the dry milk at a discount at grocery stores. The coupons were coded to indicate whether they were sent to the experimental group or control group, and the number of coupons redeemed by each group was counted.