0

I have a data set with the amount of bicycle accidents and the year in which the bicycle accident occurred. I would like to see if there is a significant increase in the amount of bicycle accident per years and by which amount.

I am doubting between Pearson's correlation and Spearman's Correlation. (Or should I use a linear regression analysis?)

T.E.G.
  • 1,676
  • 6
  • 16
  • 29
Michiel Herteleer
  • 1
  • 1
  • 3
    There are many related questions asked before on this site, for instance [here](http://stats.stackexchange.com/questions/29489/how-do-i-study-the-correlation-between-a-continuous-variable-and-a-categorical), [here](http://stats.stackexchange.com/questions/102778/correlations-between-continuous-and-categorical-nominal-variables), and [here](http://stats.stackexchange.com/questions/119835/correlation-between-a-nominal-iv-and-a-continuous-dv-variable). However, I am not sure why you think one of these variables is nominal. – T.E.G. Mar 02 '17 at 10:20

2 Answers2

0

as far as I can see your variables are both metric, therefore you could use a simple regression analysis (you can do a scatterplot, to see what kind of regression is suitable, f.e. a linear one).

The advantage of a (linear) regression is, that it provides a better interpretation of "the amount of bicycle-accident increase", because you can report the slope and say: "the expected amount of accidents increases about (the slope-coefficient) per year".

bucky
  • 205
  • 2
  • 11
0

Regression analysis and correlation solve two very different problem, so what to do depends on what you want to get from the data.

If you're interested in only a number that describes the relationship, you may want to use Pearson (for linear) or Spearman (for monotonic).

If you want a dependent variable and interested in how the variables interact, you should use regression. There's no mention the relationship is linear, so pick the most appropriate regression analysis.

SmallChess
  • 6,764
  • 4
  • 27
  • 48