3 ms·

Slightly OT, but I tried taking the same test at https://politiscales.party https://politiscales.party. And the questions are as inarticulate / ambiguous as yo

by Radim 4y ago

Slightly OT, but I tried taking the same test at https://politiscales.party https://politiscales.party. And the questions are as inarticulate / ambiguous as you'd expect from a "political" or "social sciences" quiz: 17. "It is a small group that consciously and secretly controls the world." Like, controls all of it? They control the whole world? Obviously disagree, a no brainer. Or a relatively small group of humans has an outsized impact on human affairs? Obviously agree, again, a no brainer but in the opposite direction. The question could be made more meaningful by better framing: What's "small"? What kind of "control"? How "secret"? What are you really asking? I wish there was an option "None of the above – the question is ill posed, and I try to not partake in sloppy thinking". The "Neutral / Hesitant" option hardly covers it. 18. "The right to be anonymous on Internet should be guaranteed". Guaranteed by whom, how? What does this even mean? Taking the test I feel like I'm trying to read the test authors' mind, rather than probing my own. I'm not sure what the AI makes of it, but I doubt the test authors really took into account that it's mostly a coin toss between "Strongly agree" and "Strongly disagree" (or "Neutral hesitant"?), even on the same question for the same person. Or maybe I'm just not the target audience, the sloppiness is an intentional self-selection ruse.

arc-in-space 4y ago
A lot of these questions are infuriating. "Offshoring and outsourcing are necessary evils to improve production." This one can be took to imply that offshoring is evil regardless of whether you agree that it's necessary or not. "A language is defined by its users, not by scholars." This shouldn't even be considered a political question. EDIT: Don't bother with that quiz, it's clueless.
- jefftk 4y ago
  Yes: should I mark "disagree" if I think they're not evil?
Aachen 4y ago
It's always this way. Personality quizzes, voting guides, all of it. I don't know why people can't either phrase them better or take this whole labeling/polarising business less seriously.
ordu 4y ago
> And the questions are as inarticulate / ambiguous as you'd expect from a "political" or "social sciences" quiz Nothing wrong with it. Your way to interpret a question can tell a lot about you. Though I agree it would be fun if the pedantic way of asking clarifying questions back was treated on its own. But I'm not sure it is possible in all cases. > Taking the test I feel like I'm trying to read the test authors' mind, rather than probing my own. Just relax and answer questions. > I doubt the test authors really took into account that it's mostly a coin toss between "Strongly agree" and "Strongly disagree" (or "Neutral hesitant"?), even on the same question for the same person. I don't know about this particular test, but good tests are robust, i.e. they give the same result on the same person. There are some tricks to measure robustness (for example, just give the test twice is not a good measure) but it is measured and reported in articles proposing new tests. How these tests works is a magic, i.e. sufficiently advanced technology. There is a lot of work and research goes into really good ones. The trouble is you need a certification to reliable interpret results of such tests.
- guntherhermann 4y ago
  > Just relax and answer questions. It is good to be inquisitive, this is HN, being curious is in our nature :) > How these tests works is a magic Is it not a multiple-choice with a vector given to each answer, add them up and you get an x,y? If not, why not?
  - ordu 4y ago
    > It is good to be inquisitive, this is HN, being curious is in our nature :) Yeah, but as I see it, if you wonder does this test work or not, you'd better try to play by its rules, which are: just relax and answer questions. Then you need to find a description of those dimensions, how authors of test define them, what do they mean. And the last step is to interpret results and to compare with the reality. Then you'll find something about this test. At the very least you'll know does it work for you. > Is it not a multiple-choice with a vector given to each answer, add them up and you get an x,y? It is easy to do math of a test, but it is hard to make a good test.

by Radim 4y ago

arc-in-space 4y ago

A lot of these questions are infuriating. "Offshoring and outsourcing are necessary evils to improve production." This one can be took to imply that offshoring is evil regardless of whether you agree that it's necessary or not. "A language is defined by its users, not by scholars." This shouldn't even be considered a political question. EDIT: Don't bother with that quiz, it's clueless.

jefftk 4y ago
Yes: should I mark "disagree" if I think they're not evil?

Aachen 4y ago

It's always this way. Personality quizzes, voting guides, all of it. I don't know why people can't either phrase them better or take this whole labeling/polarising business less seriously.

ordu 4y ago

> And the questions are as inarticulate / ambiguous as you'd expect from a "political" or "social sciences" quiz Nothing wrong with it. Your way to interpret a question can tell a lot about you. Though I agree it would be fun if the pedantic way of asking clarifying questions back was treated on its own. But I'm not sure it is possible in all cases. > Taking the test I feel like I'm trying to read the test authors' mind, rather than probing my own. Just relax and answer questions. > I doubt the test authors really took into account that it's mostly a coin toss between "Strongly agree" and "Strongly disagree" (or "Neutral hesitant"?), even on the same question for the same person. I don't know about this particular test, but good tests are robust, i.e. they give the same result on the same person. There are some tricks to measure robustness (for example, just give the test twice is not a good measure) but it is measured and reported in articles proposing new tests. How these tests works is a magic, i.e. sufficiently advanced technology. There is a lot of work and research goes into really good ones. The trouble is you need a certification to reliable interpret results of such tests.

guntherhermann 4y ago
> Just relax and answer questions. It is good to be inquisitive, this is HN, being curious is in our nature :) > How these tests works is a magic Is it not a multiple-choice with a vector given to each answer, add them up and you get an x,y? If not, why not?
- ordu 4y ago
  > It is good to be inquisitive, this is HN, being curious is in our nature :) Yeah, but as I see it, if you wonder does this test work or not, you'd better try to play by its rules, which are: just relax and answer questions. Then you need to find a description of those dimensions, how authors of test define them, what do they mean. And the last step is to interpret results and to compare with the reality. Then you'll find something about this test. At the very least you'll know does it work for you. > Is it not a multiple-choice with a vector given to each answer, add them up and you get an x,y? It is easy to do math of a test, but it is hard to make a good test.