Even though the
reported McFadden R-squared (a pseudo R-squared) is only 0.12, it
is inside the range generally obtained by these models with a limited
dependent variable using cross-sectional data (Greene, 2003). In the
case of the PTM general logistic model, the McFadden R-squared
increased from the complete sample logistic model, reaching 0.17.
The model correctly predicted 87% of the respondents in the PTM
who said they would buy branded beef (probability over 0.5), higher
than the prediction power for the model using the complete sample.
It predicted correctly only 45% of the respondents in the PTM who
would not buy branded beef products or were uncertain (probability
below 0.5).