The math behind “straight-run” (RESULTS)

[sksmass]

A couple weeks ago I posted and asked fellow BYC'ers to tell me the male/female ratios they experienced in their “straight-run” hatchery orders. My intent was to gather data to better understand what “straight-run” really means. I received a lot of data and present below the findings of my analysis. Many thanks to all who commented on their “straight-run” ratios.

First of all, the data consisted of 16 people who ordered a total of 417 straight-run CX from various hatcheries. Of these birds, 152 were male. So, the proportion of males in these orders averaged 36%. So, on average, when you order straight-run CX from a hatchery, you are likely to get approximately one male for every two females.

Interestingly, the 6 people who ordered from McMurray averaged 31% male while the 10 people who ordered from other hatcheries averaged 43% male.
We can definitively prove, using a binomial test, that these male ratios significantly vary from a 50:50 ratio. So, if you thought “straight-run” meant having an equal likelihood of getting boys and girls, this little mathematical exercise proves you wrong.

However, thanks to the insightful comments of JohnL11935 (@comment#6), BackyardAR (@comment#22), and WinsorWoods (@comment#25), we have an alternative hypothesis for what “straight-run” means and I think they are right.

They posit that “straight-run” just means “unsexed leftovers” That is, a hatchery will fulfill its orders for sexed chicks first, then fill orders for “straight-run” randomly out of what remains. Imagine two buckets into which chicks go at the hatchery: sexed females and sexed males. Once the sexed orders are filled from the two buckets, all the remaining chicks get combined into a bucket from which all straight-run orders are filled.

So, if you order straight-run meat birds, for which many people prefer to buy sexed males, the hatchery will first fill those male-only orders, and then fill your straight-run order from what remains, which will be disproportionately female.

And the opposite is true for ordering LAYER breeds. If you order straight-run layers, for which almost ALL people prefer to buy sexed females, the hatchery will first fill those female-only orders, and then fill your straight-run order from what remains, which will be disproportionately male.

The moral of the story here is if you really care about the male/female ratio of your order, buy them sexed. Otherwise you can expect a roughly 1:2 sex ratio skewed in an “undesirable” direction.

As for me, I will probably continue to buy straight-run CX because I prefer to have them finish out over a period of a few weeks rather than all at once. I like to process in a few batches rather than one marathon session.

 

[barred-rocks-rock]

Thanks! I found this really helpful!!!

 

[Thomas423]

Interesting. Thanks for taking the time to figure it out. I agree with you about the staggered processing. We processed the males (9) and the 3 largest females on week 9 and then processed the remaining 12 females on week 10. (we had cornish roasters, not cornish x).

 

[DragonEggs]

Makes perfect sense. Thanks!

 

[dancingbear]

I always wondered, I knew that the orders for straight run did NOT come out anywhere near even, I ordered straight run of heavy breeds once, and got all males. Since then I've ordered gender specific.

 

[haTHOR]

this is really great! now, i figure you already know this...it would be so much better if you had a larger number. around 30 would be a large enough sample to extrapolate data from, whereas 16 is not considered to be a sizable enough sample to "ensure" statistical accuracy.

that said (and my stats teacher hat is now off) i love that you did this little experiment. nice job! very very interesting.

 

[sksmass]

It depends on what you are considering the sample. 16 is number of orders, but 417 is the number of male/female observations. Of course a larger sample size is always better but I think there is enough power in n=417 to run the binomial test.

 

[twentynine]

I disagree with your opinion, not math.

The math on the responses you recieved is correct, however I believe your science is in error.

For the following reasons.

1. Not a true random sample. You could only do the math pertaining to the replies you recieved. This is not random sampling. Nor could it be considered a random cross section sample.

The only reliable way to find the "ratio" of M to F would be random gathering of samples, from randomly selected hatches through out the hatching season.

2. The tone of your initial post possibly/probably skewed the responses towards the results you were desiring to see.

I ended up with 7F and 3M CX from McMurray**

Statements such as the above copied from the original post are leading, people with your point of view are the ones that then tend to reply.



What your mathimatical/scientific testing has revealed is the results for only the (possibly biased) people who responded. If and only if they were being accurate and truthful. Not a true ratio of M/F anywhere.


I have a suggestion, call the hatchery/ies, and ask them the method (if any) used to seperate straight run chicks.

 

[sksmass]

Quote:
Researchers often are stuck with non-random samples in the social sciences. Every survey you have ever taken is subject to response bias making it, in effect, non-random. But that doesn't stop researchers from drawing inferences from the data. You go to war with the data you have, not the data you'd like to have. I have no way of obtaining the data you suggest nor could a normal backyard flock raiser be expected to have access to such. I am not publishing this in Nature. I am posting it to a online community message board! If researchers really let the "perfect be the enemy of the good" in the way you suggest then science could never move forward.

Quote:
Well, I actually had no such preconceived desired outcome. In fact, in my original post I suggested that we'd actually find the hatcheries would come out close to 50:50.

Quote:
Of course! That is the case for any survey research on any topic. All surveys are subject to bias. That is why they are used in the social sciences. However, I think the conclusions are logical, the sample is believable, and the method is reasonable. You disagree. And that is fine.

Quote:
Apply a little economic common sense and Occam's razor here. Do you not think that hatcheries are aware of their bottom line? Do you think it would be in their economic best-interest to take straight-run directly "from the egg"? I don't. This is not a conspiracy, it is just business. I don't fault them for it. I totally get it.

 

[Brunty_Farms]

Quote:
I can tell you from experience that is just not the case. I've worked at a hatchery here in Ohio and one of the larger competitors in the national market for retail chicks and can honestly say that 90% of the time you are going to get a 50.... 50 mix. The only time that this may not be true is early in the spring months. However if you ordered now, you will get an even mix. There is some truth to your statements... one is the fact that most DP breeds and rare breeds are sexed prior to packing. However the list of orders is usually alphabetically in order and they are filled that way too. Each packer is given a list of orders to be filled and we were told that pullets get pullets, males get males, and st. run was exactly 50 / 50. They do not jumble them all together they are kept separate all the way to the end.

CX chicks are quite the opposite. Because it cost money to sex chicks and it is an extra step in the process... if they do not have to sex they do not. For instances, most CX orders are ST. Run anyhow so basically they are pulled from the hatchers and the chicks are separated from the shells some go into plastic chick crates but the majority go right into 100 count boxes. The ones that are in the plastic crates get feather sexed the ones in the boxes are shipped as is.

You have to think that when a hatchery is separating, sexing, and packing 500,000 + chicks at a time they don't have time to do what your implying it's just not economical. However the reason why they are sold st.run is it covers their but if they over sell pullets. If they know they are going to have a shortage prior to hatch they tell the packers to pack 70 / 30 instead of 50 / 50. Especially if someone comes in and orders 5,000 females.... you see what I'm saying? 90% of the time though it is fair and it is just plain and simple luck of the draw.... especially with broiler chicks.

Every hatchery is different I'm sure but they all have to deal with the economic parts of it and sexing every single chick is not necessary unless it's needed.