DECISION ON STRATEGY
Research Findings on Approaches to Solution
Statistical analysis in this practicum was done by
accepted statistical methodology. The general approach is discussed by
Popham and Sirotnik (1973), and is recognized as standard statistical practice.
The computer program which was used in the analyses is the Statistical
Package for the Social Sciences, one of the most widely used computer programs
for educational research, especially at major universities.
Analysis of the Problem getting and Situation
Because this part of the practicum is looking at statistical relationships, there is little impact on its development or results by system politics, colleague characteristics, local "traditions", or intervention timing. Because the data for both variables were gathered free from interference of the above and other factors, this particular section is not relevant to this practicum with one exception - the population which is being studied.
Because the SELCs accept only students with violent/acting out behaviors, it is by definition a select student population. Had the results of the analysis showed a significant relationship between certain factors and behavior, there are some who would say that the relationship exists only with our select population. This may be true, as Landsberg found:
Some do not react to weather at all, while others respond like a barometer to weather changes. Perhaps the former have an ability to compensate for the changes, while the latter are not similarly endowed. This difference makes all studies on weather influences among groups of people very difficult. (1977, P.8)While this writer understands that point, and agrees that these selective students may exhibit more severe behavioral effects, the writer does not feel that this selectivity makes any resultant relationship invalid. What it says is that it does show a difference, which may be reflected to lesser or greater degree by the nature of the student population. Students prone to violent behavior may become more violent than those who are not so prone. But even the non-prone should exhibit behavioral change, though perhaps in a milder form, were the relationship shown. Another influence on the outcome of the study is the unknown effect the variables have on the staff in their interactions with the students. Unfortunately, the effects of the variable changes on staff and their interactions with students are impossible to determine. For the purposes of argument, let us suppose that the variable changes affect only staff behavior, as demonstrated by changes in interpretation of student behavior. Regardless, the end effect is still the same. Barometric pressure, moon phase, and welfare check delivery have a relation to misbehavior of the SELC students. This practicum is investigating the relationship between sets of variables. It is not attempting to show causality. Consequently, how that relationship demonstrates itself is unimportant to this study. Further, since the atmospheric variable factor s are not subject to human intervention, we could not change barometric pressure even were it shown to be related to behavior.
Each variable was compared with the number and average
lengths of time-outs and with the incidents of restraint, to test for correlations
- a quantifiable relationship between the two variables. The computer program
did the actual analysis, utilizing the Pearson product-moment correlation
coefficient. When this coefficient was determined for each set of variables,
it was tested by null hypothesis for its level of significance. Coefficients
having a .05 level of significance or better were accepted as significant
for purposes of this practicum. (Popham, 64-82)