... the important part is to pick the metric that you like:
First, we have our possible definitions of 'family farm' :
1. Farms operated by indvidual families
2. Farms owned by individual families
3. Farms owned or operated by individual families that produce agricultural products for sale
4. Farms owned or operated by individual families that aren't incorporated. (might be a death tax dodge, might be a huge corporatation that's tightly held)
5. Farms owned an operated by individual families that qualify as a 'small business'.
6. Farms under a given acerage.
And we can further modify what we're analyzing:
a. ...only those farms that produce agricultural products for sale.
b. ...only those farms that produce food.
c. ...only those farms that produce food intended for human consumption. (no sod or flower farms, feedstock for biodiesel)
d. ...only those farms that produce food that contributes to the human food chain. (so allow hay, alfalfa and animal feed if grown for cows, but if the cows are to be dog food).
e. ...only those farms that 'contribute meaningfully to the market'.
Then, we have our metric, selecting the definiton of 'family farm' that's most advantageous of what we're trying to show, comparing "family farms" to either "corporate farms" or to "all farms":
1. Percentage of the count "family farms"
2. Percentage of the acerage of "family farms" 3. Percentage of the acerage used for farming in a given year.
4. Percentage of the products produced by "family farms" (in tons)
5. Percentage of the products produced by "family farms" (in dollars)
6&7. Percentage of the food produced by "family farms" (tons / dollars)
8&9. Percentage of the food sold by "family farms" (tons/dollars)
Some of these, I'm not even sure which way the selection bias will be. (family farms might sell at farmer's markets and get a better price per pound ... or they might focus on herbs and things typically sold at higher margins that don't tend to be grown on a massive scale).
But like anything, you run all of the different combinations, and pick the one that gives you the answer to support whatever argument you're trying to make.