Interpretation with a share on the left, and a delta share on the right side

[Xavier Pedros]

Hi,

I have regression of the following form:

Y_i = a + bΔX_i + e_i

Where Y is the share of votes for a political party, in municipality i.

and ΔX is the percentage point increase on a variable X (which is a share, too), in municipality i.

Both X and Y were introduced with values ranging from 0 to 100. Hence, I understand that the units of my variables are percentage points.

For b = 0.30, is it accurate to say that a 1 pp increase in X, leads to a 0.30 pp increase in Y?

Or perhaps more rigorously:

A 1 pp increase in ΔX, leads to a 0.30 pp increase in Y?

Many thanks!

[Clyde Schechter]

Almost. Fully correct, since this is observational, not experimental data, would be "A 1 pp increase in ΔX is associated with a 0.30 pp increase in Y." Avoid using causal language in presenting results from non-causal analyses.

[Xavier Pedros]

Thanks Clyde!