Hope this helps you: http://rdc.uwo.ca/slides/Fernando,LogitTobitProbitMarch2011.pdf
In short, when your independent variable X increases by 1 s, probability that data came from THIS category (instead of all lower categories - if Y was treated as ordinal / or instead of all the rest categories - if Y was treated as nominal) increases by 10.5%. If it's odds ratio (OR) of 1.105 you've reported.