PrimeTime_Fall_2018

Selection Indexes

BY JOHN GENHO, NEOGEN AGRIGENOMICS

S everal years ago, a group of researchers at Colum- bia University performed a study on choices. They placed a large display of jams in a gourmet grocery store. Customers who stopped for a few moments to sample the jam were given a $1 coupon toward the purchase of one of the many jams on display. With 24 different types of jam to choose from, only 3 percent of the customers who sampled the jam and looked through the choices made a purchase. The next day, the research- ers set the display up again. However, this time they only displayed six different types of jam. Fewer customers stopped to sample the jam (40 percent of those passing stopped to see the smaller display, while 60 percent stopped at the larger display). But of those who stopped for the smaller display, 30 percent purchased a jar of jam. The lesson learned in this study is clear: people have an easier time making decisions when there are fewer choices. We’ve all been in the situa- tion of hurriedly looking over a menu at a restaurant when everyone else at the table has ordered and we haven’t decid- ed yet. When I’m in that situ- ation, I immediately skip past the sections labeled “veggie wraps” and “chicken.” If there is a ribeye steak on the menu, that’s what I order. I’m sure I’m passing over some good dishes by doing this (although prob- ably not in the “veggie wraps” section), but falling back to what I know will be acceptable is much easier than trying to sort through a large amount of data quickly.

I’ve taken

sale prices from several bull sales and correlated them back to the expected progeny dif-

ferences (EPD) that were displayed in the catalogs.

Selection indexes simply merge EPDs together by us- ing certain weights of impor- tance. Usually, these weights are driven by the economics of a certain business model. A bull buyer retaining heifers and selling calves at weaning is running a different business than a buyer purchasing heif- ers and marketing fed cattle on the grid. These different business models result in dif- ferent weights being placed on the different EPDs, and the resulting indexes being dif- ferent. These two buyers may be interested in very different bulls at the sale (as opposed to the situation where they are both bidding on the lowest- birth-weight bull.) Selection indexes makes it much easier to find the bull that fits a par- ticular situation. There are some buyers who already have a decision-mak- ing model that they apply to sale catalogs. They may say that they never buy a bull be- low a certain yearling weight EPD, or above a certain birth weight EPD. Animal breeders call these independent cull- ing criteria. Many bull buyers remove potential bulls from those they are interested in based on these cutoffs until they are down to a certain number of potential purchases. While this method is much better than randomly selecting

The only EPD I’ve ever found that highly correlated to sale price is birth weight. It’s as if bull buyers are looking at the menu of EPDs in front of them as their waitress taps her pen- cil impatiently on her pad of paper (in this case it’s an auc- tioneer, not a waitress). The buyer finally decides that there are likely some important things that he’s missing, but the one thing he remembers is pulling a calf on an icy, cold night, and he knows he doesn’t want to repeat that experi- ence. So, he goes with what he knows – low-birth-weight bulls – just as I go with my trusty ribeye steak. In this situation, we could learn something from the Columbia researchers trying to peddle jam – plac- ing too much information in front of decision makers often causes paralysis in the deci- sion-making process. The problem with limiting information is deciding what needs to be removed. We have a host of EPDs that are incredibly relevant and important to seri- ous breeders. We certainly don’t want to hide information or limit a buyer’s ability to sort through what they feel will help them make a better decision. The answer is summarizing informa- tion to make what is displayed more concise. This is where selection indexes come in.

Continued page 38

Akaushi Prime Time • Fall 2018

36