log[ (x^2*y)/(sqrt(z)) ] = log[ x^2*y ] - log[ sqrt(z) ] .... Use Rule 2 log[ (x^2*y)/(sqrt(z)) ] = log[ x^2 ] + log[ y ] - log[ sqrt(z) ] .... Use Rule 1 log[ (x^2*y)/(sqrt(z)) ] = log[ x^2 ] + log[ y ] - log[ z^(1/2) ] .... Use Rule 4 log[ (x^2*y)/(sqrt(z)) ] = 2*log[ x ] + log[ y ] - (1/2)*log[ z ] .... Use Rule 3 log[ (x^2*y)/(sqrt(z)) ] = 2*a + b - (1/2)*c .... Use Substitution
So log[ (x^2*y)/(sqrt(z)) ] will break down into 2*log[ x ] + log[ y ] - (1/2)*log[ z ] which is equivalent to 2a + b - (1/2)c when a = log(x), b = log(y), c = log(z)
When two variables have a positive linear correlation, the dependent variable increases as the independent variable increases. So when the independent variable decreases, the dependent variable decreases as well. Both variables decrease and increase simultaneously.