Hacker News Reader

Comment by DanI-S | Hacker News Reader

Interesting idea, but the slider may not be a great interface. It stops the experiment being purely about numbers - I found I picked '5' because it was close to being nice and symmetrical.

View on HN

I think that is a good datapoint, myself. If there is a predisposition for people to choose "5" under those circumstances, we will have learned something.

This would be a fantastic app to run A/B -like testing on to determine whether say a circular "picker" gives different results than a horizontal "picker". You could even tighten it down to see if a random distribution of numbers between 1 and 10 gives different results than a ordered distribution.

It fits the model perfectly and, at least to my perspective, very clearly.

Edit: Doh. Looks like you're doing just that.

I also took 5 because it was close. Also, since I know people tend to pick 7, I somewhat tend to don't pick 7 when I'm asked something :p You could maybe "randomize" the order of the numbers, or hack them up so, for instance, 7 is far left and 3 is far right.

If there are enough data points (which it seems like there is), and each UI is chosen a random 1/4 of the time, then it will be very easy to isolate and determine the effect of the slider UI. Problems would arise if the slider was chosen 60% of the time in Firefox, but only 30% of the time in Opera, and if Opera had a 20% higher population of hackers, who had a distribution centered around 6 as opposed to the aggregate population centered around 5; as is, it sounds like there's plenty of random sampling solving these complications.

Also, unlike other experiments, slider is preset (to 1.)

It may also conflict with cultural expectations. Check out the chapter "How We Think About Space" in Lera Boroditsky's talk "How Language Shapes Thought". http://fora.tv/2010/10/26/Lera_Boroditsky_How_Language_Shape...

Make another experiment with the same numbers but change the order?

i didn't even know that was a slider but it looked like something that needed clicking... got a 5.

EDIT: for clarity i used opera, then chrome then safar then ff. The UI is different for all!

No, the UI is different every time you refresh the page. There are 4 UIs, randomly presented.

I'll be fascinated to see how the different UIs skew the results. Of course, there are now so many variables in this "experiment" it'll be hard to make any concrete determinations.

But there will be data and it will be cool.

Reply on news.ycombinator.com