Help with confirming what I can and cannot conclude from this regression output

This study tried to determine the factors influencing the probability that voters would be asked to show ID in order to vote in a certain state during a certain election. (Model 1 identified Hispanics by self-identification and Model 2 identified Hispanics by whether they had an Hispanic surname). My understanding is that I can conclude, based on the output in Model 2, the following:

(1) I could calculate the probability that someone with an Hispanic surname was asked to show ID, given the other factors controlled for in the model, by adding 1.011 to the constant, which is 1.365, and then converting that figure, which is 2.376, from the logit to a probability and getting 91.5%.

(2) I could do the same type of calculation to get the probability that each of the other demographic/voter characteristics were asked to show ID and in each case the correct way to interpret the probability would be to say, "the probability that a [demographic/voter characteristic] was asked to show ID to vote in this election was X, controlling for all of the other characteristics in the model"

(3) Gender, having an Hispanic surname, and being a first-time voter are the only characteristics whose relationship to being asked for ID to vote is significant in this model.

However, I believe that I could NOT conclude:

(1) the probability that someone who is BOTH Hispanic and male would be asked to show ID, because this would require an interaction term and I would need the original data to calculate this.

Is there anything about my understanding above that is incorrect?

I'm trying to learn how to interpret social science studies precisely and I really appreciate your help!


Super Moderator
Sorry for the delay in releasing this post - it was placed in moderation by our anti-spam software (probably just since you're a first time poster and included an attachment). All good now!