Abstract
Multiple regression lag analyses demonstrated that four to five major components of the weather matrix accommodated approximately 30% to 60% of the variance in daily mood scores for 5 male and 5 female university students over either a 2- or 3-mo. period. Between 10% and 30% of this variance was also explained by day-of-the-week for 4 subjects. Optimal values from the weather matrix occurred about three days before mood measures. Two of the most frequent variables were geomagnetic variation and mean barometric pressure. Temperature, which was contaminated by serial correlations, could be substituted by other variables. Standard errors of the estimate for the equations ranged between 8% to 15% of the mean mood measures. However, equations for three of the subjects' predicted scores were significantly correlated with the observed scores for the 10 days and to a lesser extent the 20 days after the period used to generate the equations.
Get full access to this article
View all access options for this article.