What is the general linear model?
y = m*x + b
where y is the dependent variable, x is the independent variable, m is the slope of the line, and b is the x-axis intercept.
Any model's intent is to allow some degree of understanding correlation between different things, and sometimes trying to predict future outcomes.
A “linear” model is simple to understand, but real-world data do not always come in a linear fashion. “Linearization” techniques are applied to data to attempt a closer “fit” (correlation) with a linear model. The correlation factor indicates the relative correctness of the linear model to the actual data. In all cases the level of “accepted” error must be understood.
No statistical “answer” is absolute or definitive. They can only give a relative probability of an occurrence with an associated level of error in the estimation as well. Interpolations of data sets can be done within the known error constraints. Extrapolation (prediction) increases error dramatically and non-linearly, and should be done only with extreme caution.