This study attempted to identify factors which impact on decision making relating to physical restraints. It was hypothesized that nurses who scored higher on instruments used to measure levels of fear, anger and need for control would restrain patients earlier in the progression of aggression. The collective effect of these three emotions on the decision to physically restrain was examined using multiple regression analysis. The individual effect of each of these emotions was examined using Pearson product-moment correlations. Decision to restrain was measured by an audio visual tool created by the researcher (Moylan Assessment of Progressive Aggression Tool - MAPAT).

One hundred and six female nurses were given the Fear Survey Schedule III (FSSIII), the Multidimensional Anger Inventory (MAI), and the Desirability of Control Scale (DCS) in random order. They then were given the MAPAT measure. None of the hypotheses were confirmed. The collective effect of these emotions were not statistically significant (p =.135). When measured individually, the level of fear was found to be statistically significant (r =.179 at p =.030) in a positive direction. When scoring of the MAPAT was revised to reflect the progression of aggression as ordinal level of measure rather than an interval level of measure, anger also had a significant positive correlation with decision to restrain scores (r =.167 at p =.045), in the opposite direction of the proposed hypothesis.

Supplemental analyses using demographic data were done. Nurses who had a history of injury from patient assault significantly later than uninjured nurses (p =.032). The mean score for the injured group (253.346) represented a time in progression of aggression when the patient was dangerously violent. These findings indicate a pattern of higher fear levels, history of injury, and later decision to restrain, which could lead to cyclical assault.

The failure to find validity for the proposed hypothesis may have resulted from problems with the MAPAT tool or with the application of fight or flight theory. Further testing for construct and criterion validity is indicated. The finding that fear and anger were significantly related to later restraining may result from an application of fight or flight theory different from that originally proposed. The dynamic of "freezing" (failure to take action) rather than neutralization may be occurring.

Findings which identified groups whose scores reflected unsafe decision making is a cause for concern. This study confirmed findings relating to the frequent level of assault and variability in restraint use discussed in the literature. Sixty nine percent of the subjects in this study had been assaulted. MAPAT scores varied greatly both between and within each of the demographic groups. These findings indicate a need for continued assessment of safety issues in acute psychiatry.