Logistic Regression Problem - HELP!

#1
A local health clinic sent fliers to its clients to encourage everyone, but especially older persons at high risk of complications, to get a flue shot in time for protection against an expected flu epidemic. In a pilot follow-up study, 160 clients were randomly selected and asked whether they actually received a flu shot. In addition, data was collected on their health awareness, which was combined into a health awareness index, for which higher values indicate greater awareness. The estimated logistic regression model is
log(/(1-)) = -2.447 + 0.0568*(health awareness index)
where is the estimated probability that the client received a flu shot.

Estimate the health awareness index at which 48% of all clients receive a flu shot.
I said 42.469 and was incorrect.

Estimate the health awareness index at which the odds of receiving a flu shot are 2 to 3.

THANK YOU!
 
#3
Would 48% be probability or odds?

I set up the problem like this: log(.48/1-.48)=-2.447+0.0568x and solved for x. so that x=42.469

Alternatively if I set up the problem like this: log(.48)=-2.447+0.0568x, I find that x=37.469
 
#4
For the second question, estimate the index for which the odds are 2 to 3 I did this:

log(2/3)=-2.447+0.0568x and found that x=39.981.

Does this seem correct? Do I find the log(odds) and solve for x for this type of problem?
 

Dason

Ambassador to the humans
#5
Would 48% be probability or odds?
Estimate the health awareness index at which 48% of all clients receive a flu shot.
Sounds like probability to me...

I set up the problem like this: log(.48/1-.48)=-2.447+0.0568x and solved for x. so that x=42.469
I mentioned this before but... you gotta stop rounding in your intermediate steps. That's the right method but you're getting an answer thats off by almost an entire unit and I'm guessing it's because you're rounding everything to 1 or 2 decimal places in your intermediate steps.