explaining marginal effects of a conotinous variable

[Ashish Chouhan]

Dear statalist members,
while estimating the average marginal effects of an IV probit regression, using
. margins, dydx(sdi age educ famsize) predict(pr)

The AME for sdi (diversification index) are -0.987. The variable "sdi" is an index of diversification and ranges from 0 to 1. How should I interpret it?
Any help is appreciated.

[Maarten Buis]

For a unit increase in sdi (i.e. going from the minimum possible value to the maximum possible value) the probability of "y" (you did not mention what the dependent variable is) decreases by .99 or 99 percentagepoints.

[Ashish Chouhan]

Thank you for response, Maarten Buis. Sorry for the confusion. I think I did not explain my question properly. My dependent variable is adoption (Y=1 and N=0). The SDI is the Simpson's diversification index, which is a continuous variable with a range between 0 and 1, where 1 indicates complete diversification. My question is how to interpret the results of the index, and how is it different from the other continuous variables such as age or years of education?


Here are the summary statistics for the variables

 sum sdi age    educ .  famsize

Variable        Obs    Mean    Std. dev.    Min    Max
                        
sdi        373    .6041497    .0893173    .46    .82
age        373    52.23592    12.57785    22    80
educ        373    4.97319    4.566631    0    15
famsize        373    5.718499    2.691805    1    18

Thank you again.

[Maarten Buis]

That does not change my answer.

[Ashish Chouhan]

Thank you. That clarifies my doubt.