# Thread: nominal (yes/no) and distance. What test to use to proof correlation?

1. ## nominal (yes/no) and distance. What test to use to proof correlation?

Hi all,

My boss has a list that shows the interest of people for a new venue (interest? Answer: yes or no) and the distance to that possible venue. He asked me to show if there is any correlation between the two of them..
Although I searched and searched, I find it hard to decide which test to use for this (it's been quite a while since I followed statistics classes). Although I think it should be really easy and straightforward. Any guru who has an easy answer on how to proceed?
Am I correct this is a nominal data set (yes/no) and a continuous data set (distance)?
The list looks this:

Yes / 20km
Yes / 15 km
No / 8 km
Yes / 2 km
No / 5 km
etc. / etc

2. ## Re: nominal (yes/no) and distance. What test to use to proof correlation?

Binary logistic regression would be a good approach for this.

Seconded!!

4. ## Re: nominal (yes/no) and distance. What test to use to proof correlation?

I am wondering if there would be a problem with spatial autocorrelation here?

5. ## Re: nominal (yes/no) and distance. What test to use to proof correlation?

Originally Posted by gianmarco
I am wondering if there would be a problem with spatial autocorrelation here?
Very interesting...

I can see where distance is not the only factor involved in this scenario. One might be willing to go further if there are good roads to drive vs. a short distance and heavy traffic. Or if a venue is right off a train stop vs. needing a cab also. Or amenities at new venue vs. old.

6. ## Re: nominal (yes/no) and distance. What test to use to proof correlation?

I was thinking "spatially", in other words there could be spatial auto-correlation going on there, with people being closer to the venue being more keen to go.
While Logistic Regression could be used, one should use some form of auto-logistic regression. There is a sizable amount of literature on the topic.

 Tweet