I would add in the same vein: rather than engage in the simplistic battle of labels ("socialist, extremist") , which generally leads only to one side declaring its labels acceptable/applicable and the others' not, one is usually better served to argue the consequences of competing points on the ideological continuum. Here empiricism can be employed, and the efficacy of competing approaches at reaching desired outcomes measured. Whether or not Label X applies to Politician Y is generally a less interesting discussion than the effects of policies A,B,C on our welfare. It's quite a bit more work, but as our viewpoint is informed by the harsh chill of the marginal propensity to produce as articulated by von Mises, Hayek, and Friedman we must embrace the arduous path as the one to our ideological victory as well as to greater human happiness generally.

In any case the label "socialist" should not really be considered purely perjorative-- virtually everyone agrees a certain level of coercive redistribution is desirable (as few would like to see Social Security end as would see it extended to those aged 35), the question needs to be understood as how much socialism we can afford now consistent with maximizing the long-term productivity growth that is the source of the improvement in overall living standards.

Of course, the audience must be amenable to reason in the first place, and such determination should be made before investing the effort, as I've learned (too late, to my detriment) from interacting with a couple popular fiction authors whose authorial talents did not translate to an amenability to reason, or even an honest interest in it. One laughs at the clown, one applauds the clown, one may even discuss the intricacies of comic timing with the clown, but one is not necessarily advised to attempt to teach the clown vector calculus.