I find that it is very hard to interpret quadratic parameters.
When I can see an inflection point in the scatter plot, does it make any sense at all to create a dummy variable such that higher than inflection x's are 1, and lower are 0 and then to interact that with the continuous?
e.g. say i have income in $ and age in days
maybe i can see in scatter plot that the income age chart inflects at around 50 years of age (18250 days).
Does it make any sense, and is it even possible to run a model:
When I can see an inflection point in the scatter plot, does it make any sense at all to create a dummy variable such that higher than inflection x's are 1, and lower are 0 and then to interact that with the continuous?
e.g. say i have income in $ and age in days
maybe i can see in scatter plot that the income age chart inflects at around 50 years of age (18250 days).
Code:
gen age_high = 0 replace age_high = 1 if age > 18250
Code:
reg income age##age_high
Comment