A Better Government, One Tweak at a Time | Hacker News Reader

A Better Government, One Tweak at a Time

nytimes.com

36 pointsdean10 comments

The way I read this is that the key is the difference between running government at the macro level versus the micro level.

Traditionally, government policies and lawmakers create laws and regulations with broad brushes expecting the people closer to the ground to execute it. This type of centralized authority rarely works.

What they deem A/B testing is executing at the micro level down to almost individual choices. This gets you better and more immediate feedback.

The same thing happens in large companies. When there are so many layers between management and the consumer, it is hard to make the right decisions. But if management is talking directly with the consumer, they get better feedback and better decision making.

Technology has disrupted and continues to disrupt many industries, but those focusing around human relationships have been the hardest to automate. As technology gets better, you will see more disruptions to how government goes about its business. The main obstacle to this is that governments don't have natural competitors, and need to be dragged kicking and screaming into the future.

This is such an obvious idea you have to wonder why it is even news. Why shouldn't government be constantly iterating and ever so slowly getting better?

What isn't so obvious is that government developers should design from the viewpoint of their users.

http://www.citylab.com/tech/2014/01/what-really-happens-when...

It's amazing how much of this boils down to "If you make sure people know about things, people know about things!".

I mean it happens over and over in the article "Remember, you can default to double sided", "Did you know about this savings program?", etc.

Still, love the idea, and I hope they keep it up. It sounds like it's had some really positive effects so far and what's the point of the scientific method of testing and evaluation if you never use it?

> eight text messages at a total cost of about $7 a student

lol dumb government wasting money!

> The text intervention was particularly cost effective: the messaging campaign cost $7 per student, inclusive of the expense of hiring school counselors to support students who needed additional assistance.

This is from the study[0] linked in the article, and makes it clear that labor is the expense, not sending text messages.

This isn't just an error or a bad story, it's the deliberate removal of defining information. Moreover, I'd argue that it's not only intentional but systemic at the New York Times, and has been for decades.

[0] http://curry.virginia.edu/uploads/resourceLibrary/9_Castlema...

"Certainly eight text messages are a more cost-effective way of promoting college than offering thousands of dollars in grant and scholarship aid."

I do not understand how in the world the person writing this article came to this conclusion unless they just straight up have zero clue how people that did not grow up with a college fund pay for college. Doubly so when paired with "The effect was particularly large for low-income ... students." How do they think that low-income students pay for college?

The point is that adding more grants is less effective than making students aware of grants that already exists. For low-income students whose information through local social networks isn't getting the message passed about potential grants, trying to shake the pot around harder isn't going to attract the attention they want.

Ah, that makes a lot more sense. I did not think that was very clear in the article, however.

Can't be any worse than the way they do it now.

I admire your optimism.

Reply on news.ycombinator.com