That depends on the particular situation and how the association was calculated. Sometimes, the more alleles you had the more risk you have, but sometimes one risk allele is same as two. Usually, written like this it would suggest an aditive effect of a SNP.
I would take a caution of taking any "association studies" too serious. I mostly consider them something as a "place to look" for some actually functioning thing that affects the risk. Not always those are possible to found and in clinical settings many SNPs can be used as a predictive or so on. But mostly it is what is it called "association" only. And possibly not even with the single base of that SNP itself, but likely with something linked to it, that actually has the functional effect.
Most SNPs that are well established, have some functional background anyway. Your usual SNPs, you need to take into consideration: 1) what was tested (metodics, whole genome SNP chip or WGS) 2) who was tested (population/family linked effects only) 3) how was the "effect" measured (qualitatively/quantitatively, was the diagnostic good or vague) 4) how it was statistically calculated (no dirty games with "we need at least one significant SNP for this project" things, were cofounders assesed including enviromental, genetic etc., sample size and this whole thing. This also means if they calculated the genotype effect or just allele effect and how to interpret this.
There are so many factors, why SNPs can only show association in one study and not in other (which does happen more often than not). Also genes do not work alone, on the contrary, so single SNP having some serious effect without regard for all the rest of the genes? Sounds like a way too shady.
But since consumer genetics/genomics became a thing, people are just crazy for SNPs having folds of risk on this and that. Some of the lay people even add or multiply the risk assessed in different studies (never do that). Most of those would never be proven as anything more than a "marker" and only in certain circumstances.