In the title of a post that I came across several days ago, linguist Arika Okrent asked the question: "Why does ‘a quarter of’ mean the same thing as ‘a quarter to’?" As the "quarter form" for telling time was elicited in interviews for the Linguistic Atlas of the Middle Rockies (LAMR), this post spurred me to finally get around to analyzing my data for it. However, rather than asking why, as Arika had, I was curious to see what the range of responses for the prompt was in LAMR and if there was any evidence that these different responses were influenced by the social characteristics of the informants who gave them. Here's what I learned.

First, 53 of the 70 LAMR informants provided quarter forms. This doesn't mean that 17 informants didn't use these form in their natural speech (only one informant explicitly reported not using a quarter form); rather that these informants, for some reason, were either not asked or failed to answer the question (and quarter forms never occurred in discourse of the interview other than as the product of direct elicitation). Per usual in Linguistic Atlas interviews, each informant could provide multiple responses to any prompt they so desired, although only one did so in response to this prompt (and is mentioned in the next paragraph). Tallies of the quarter forms that were provided by LAMR informants are presented in the figure below:

As shown, four quarter forms are found in the collection: quarter to, quarter of, quarter til, and quarter before. The sole informant who provided quarter before also reported using quarter to, the most common quarter form, by far, in the collection.

The areal distributions of the three most common forms are shown in the maps below:

As shown by the maps, the distribution of the three primary quarter forms shows little areal patterning. Utah is the most homogeneous of the three states with respects its choice of quarter form, being nearly unanimous in its preference for quarter to, but aside from that and a rather small cluster of quarter tillers in the northeast section of Colorado (where one of these purple markers actually represents two informants, both offering quarter til, in Limon, the Hub City of Eastern Colorado!

With respect to the social distribution of quarter forms in LAMR, I looked at sex of the informant and education. Of the 53 quarter-form users, 28 were female and 25 male; their choices of quarter forms are presented in the table below:

As shown in Figure 5, while the use of quarter to and quarter til seems to be relatively unaffected by sex of the informant, the use of quarter of has a much stronger association with females than males in LAMR.

With respect to education, the informants were broken down into four categories: unknown; did not finish high school, finished high school, and attended college. The breakdown of quarter forms with education variable are presented in the figure below:

The most apparent trend in this figure is the low figure provided by high school graduates for quarter to and their higher numbers elsewhere. This is not only interesting in and of itself, but because it leaves the group with the least education and that of the greatest amount of education to pattern very similarly together, which is usually not the case in studies of linguistic behavior. 

To summarize, in LAMR the most common quarter form is quarter to, but there is ample evidence of two other forms, quarter of and quarter til, as well as a oncer: quarter before. Although there is little evidence in LAMR that region plays much of a factor in the forms that speakers use, there is some evidence that social factors, namely sex and education, might play some role in the variants that speakers choose to use, even in something as commonplace as telling time.