grocer and associate

How do you pronounce the words "grocer", "grocery", and "associate" (both noun and verb)? I myself have:

grocer ["gRoSR=:]
grocery ["gRoSRi:] (or quite carefully ["gRoSR=:i:])
associate (noun) [@"soSi:I?]
associate (verb) [@"soSi:e?]

In General American and Received Pronunciation all of these words seem to have [s] instead of the [S] that I have. These likely are due to differences in the way that palatalization goes on.

Unstressed /i/ followed by a vowel seems to more readily palatalize preceding consonants than in other dialects (even on the South Side of Milwaukee thay say "associate" with [s] rather than [S]).

Also, /@r/ seems to palatalize preceding consonants in my dialect rather irregularly under conditions that I have not ascertained (and which may vary greatly from individual to individual) and which may be lexicalized. However, "grocer" and "grocery" are the only cases of it palatalizing /s/ alone that I can think of, as all the other cases involve it palatalizing /t/ (which may in turn trigger a preceding /s/ to become [S]).

I have:

grocer ["gr\7Us@`]
grocery ["gr\7Us@`i]
associate (noun) [@"s7USiIt]
associate (verb) [@"s7USi%eIt]

<<In General American and Received Pronunciation all of these words seem to have [s] instead of the [S] that I have.>>

I don't think the situation is as uniform as that. [s] in "grocer" seems to be predominant in GA and probably universal in RP, but I think that [S] in "associate" is predominant in both GA and RP. m-w.com, for example, gives [S] as the primary pronunciation for "associate" but [s] as the primary pronunciation of "grocer, grocery". And the Longman Online Dictionary ( http://www.ldoceonline.com/ ) gives [S] as the primary pronunciation for "associate", for both RP and GA, but they don't even mention the [S] pronunciation for "grocer, grocery".

In fact, the Dialect Survey ( http://cfprod01.imt.uwm.edu/Dept/FLL/linguistics/dialect/maps.html )included a question about "grocery". For the whole US, 52.3% used [s] and 45.5% used [S]. Our respective pronunciations make sense: in Wisconsin, 60.5% used [S] and 37.1% used [s], whereas in Massachusetts, 76.2% used [s] and 22.3% used [S].

Then we would probably have to say that these are common variations throughout the whole of North American English rather than local innovations. Of course, I would expect [S] in "associate" from dialects that more readily palatalize /s/ (from following /j/ or unstressed /i/ before another vowel) than from dialects which do not readily do such.

As for "grocer", if the use of [S] in it is not related to things like the use of [tS] (or just [S]) in "sister", "yesterday", "history", "restaurant", "faster", and so on, then I am not sure what is from. (Of course this varies a lot, as most people here have [st] in "sister" even if they have [stS], [StS], or [S:] in "yesterday" (even though some people here have [stS], [StS] or [S:] for "sister", and I myself alternate between [st], [StS], annd [S:] for it), and usually I hear [st] in "faster" even though I have heard individuals who have [stS] or [StS] instead).

I have:

"grocer" [gr\7USr\=]
"grocery" [gr\7USr\i]
"associate" (noun) [@s7USi@t]
"associate" (verb) [@S7USieIt]