Determining where farms fall within the five categories is difficult. The data I used came from in-depth interviews with farmers of 54 CSA farms in California's Central Valley and surrounding foothills, 48 of whom also completed an online survey (Galt et al. 2011; Galt et al. 2012). To use these data in my framework, I calculated “per farmer annual earnings” for 2009—which, for ease of discussion, I refer to hereafter as “earnings”—and used it as a proxy for farm profit. I calculated earnings from survey responses and answers to interview questions, including inquiries about general profitability, gross sales, and net profits in 2009; farmers’ own salaries (whether and how much they formally pay themselves, or whether they use “what's left over” accounting); how they qualitatively value their own labor power; and the characteristics of the farms, farmer's philosophies and motivations, and pricing of their shares. More specifically, when farms paid farmers a formal salary, I used this amount for earnings. When farms used “what's left over” accounting, which is far more common, I divided the annual “what's left over” total by the number of farm partners (those with ownership obligations and who were involved in management decisions). For 36 of the 48 farms, the surveys and interviews provided enough detailed information to calculate per farmer annual earnings, while in 12 cases, the farmers did not know their profits or what was left over as personal compensation or did not want to share that information.