Dr Green advertised in the local paper, asking parents of hyperactive children to volunteer their children for an experiment. He used the first thirty volunteers for the experiment, assigning ten to each group at random. The three groups were as follows:
Group A - had to drink nothing but sunshine squash for 1 week. Group B - had to drink half sunshine squash and half orange juice for 1 week. Group C - had to drink orange juice for 1 week.
At the end of the week, each child was videotaped in a classroom situation for 30 minutes. Dr Green used the videotapes to count the number of times each child fidgeted. The results are given in Table 1.
Table 1. Mean number of fidgets and standard deviation for the three groups in Dr. Greens study.
|Mean no. fidgets||S.D.|
|Squash + juice||49||6.2|
A one-way between subjects analysis of variance was carried out to evaluate these results. This demonstrated that the groups differed significantly at the .05 level (F(2,26) = 5.23). Dr Green concluded that a moderate amount of squash was not harmful, but that children should not drink large amounts of sunshine squash.
Sunshine plc was not very happy with Dr Green's findings. They questioned the small sample size used and argued that since the parents knew what type of drink they were giving the children, perhaps those giving the children the squash were most anxious about the outcome, and had communicated this to their children.
The marketing department was instructed to contact the next 45 mothers who had volunteered, and to repeat the experiment, except that this time all the children were given the squash only. As well as measuring the frequency of fidgeting the mothers were asked to complete a questionnaire which measured their level of anxiety about their child.
The mean number of fidgets for this group was 61. The mean level of anxiety (measured on a seven point scale) was 5.4. There was also a significant correlation, r = .29, between the two measures.
Sunshine Squash plc concluded from their results:
(i) Dr Green's results were probably not reliable because the mean fidgeting score obtained by their marketing department was significantly lower than that obtained by Dr Green (t = 2.21, p < .05);
(ii) the correlation between fidgeting and parental anxiety showed that the effect of additives on the children's behaviour was the product of a self-fulfilling prophecy on the part of the parents.