The site doesn't work well on mobile. It keeps randomly jumping up to the beginning.

That's alright, it's hideously difficult to read on my 27" retina display iMac so it's a bad experience all 'round.

Funny how that works. I could feel my eyes relaxing when I clicked the link, and I thought to myself that this is what a readable, content-first, website should look like. Using FF on a regular 24" full-HD screen.

For me, it's too low contrast. Extremely thin, monospaced, light-grey font on a black background just doesn't read nearly as well as well-proportioned proportional dark font on a light background — or, if it must be a dark background, yellow text on dark blue is the most legible. Regular Courier over Courier New could help, too, but ultimately a proportional font directs the eyes more easily.

It's the monospaced font, with grey letters on black background that makes it hard to read.

I'm using FF62 on Android and I thought "what a refreshingly simple design", maybe the problem was js that my browser blocked.

thanks for the feedback ... the CSS must be much better now

The real trick is how the shiny book and class were sold to the company.

You can get good ideas from a book, or a box, or a class, but you don't get results from them.

Just a heads up, my corporate laptop is reporting this site as a malware distributor.

Strange. There is nothing there but Google Fonts and a script for the background lines and the one image. No other external resources. It is hosted on GitHub.

"Every organization strives to improve and empower its employees." lol

If you consider employees as taking money as input, and producing more money (in profit) as output, I guess that statement holds.

It does take a lot of cynicism, though.

Would 25 multiple choice questions be enough to detect random answering with certainty? Surely many combinations come up way more often than others when answered truthfully.

A perfect test would, first, try to prevent correlation between questions. It would be a waste to ask 4 questions when 3 suffice to predict the last answer with good accuracy.

After that, some questions supposed to be perfectly correlated with previous are added. These should, however, use different words or concepts altogether, to avoid detection. These questions are meant to establish consistency, and would pick up on randomness

25 strikes me as too few to get into that kind of verification. I seem to remember serious personality tests to usually contain in the range of 100 to 200 questions.

You're on to something. I recently took several, and they contained a lot of questions (100+). I noticed that many of the questions were testing suspiciously similar things, only worded differently.

>It would be a waste to ask 4 questions when 3 suffice to predict the last answer with good accuracy. //

How you ask a question can perturb the answer. If an axiom is that you're determining static elements of personality, then you'd need to ask about the same facets in different ways to avoid skewing the analysis.

"I am quick to react?"

vs

"I take time before reacting to make a measured response"

Both a priori sound positive, so people will tend to "somewhat agree" (as the answer format of these things goes), but they are seemingly antithetical.

Of course they're not really antithetical they're just shallow questions that aren't really going to discern much of your personality. People can react quickly/slowly according to the needs of a situation. Maybe the value is in measuring perceptions of one-another, despite the ease with which we are misguided.

Also, I’m sure many people tend to stray away from extremes.

And thus we see that cold-reading can be automated.

See, e.g.: astrology, enneagram, MBTI, Big Five...