When this came out I was skeptical and looked at transactions that I had metrics for to corroborate with some colleagues. Not at liberty to discuss specifics, but we all rejected the proportions that the study used.
End of the day, complex processes get completed when offered to the public, and computer based processes are generally the norm. People somehow get hunting licenses, navigate complex DMV processes, book transportation, and get passports, even when they are not above average IQ.
I think if you asked these questions based on end goal rather than method, you’d see smaller numbers of level 0/1 people.
It sounded to me like the assessment was framed around PC use — most people don’t use PCs! I’ve seen scenarios where complex workflows delivered in social services scenarios (where education, literacy and language competency are an issue) where getting tasks completed in mobile (which is ubiquitous and often more complex) is significantly better than a similar complexity task (travel vouchers) with a professional audience on PC.
If less than 5% of the population is capable of performing complex tasks, society would break down. Something is wrong with the assessment of complexity or questions asked.