Look at Lego and Playmobil. Lego started out as a construction game that, although very popular in my youth, almost completely disappeared whereas Playmobil, a storytelling/roleplaying game, was an instant hit. Then Lego adopted a Playmobil-like strategy with little men, stories and kits that, like Playmobil kits, are not intended to be reused for something else. They did this very well and Lego went through the roof to become one of the most successful brands in the world.
To me this shows that most people are willing to go through the effort of assembling something but only when the assembled object is useful to them somehow. Being challenged, getting creative with a kit’s parts, rearranging them to make something different, does not interest them.
And this is exactly what all these electronics kits that we see today, all these microcontroller-made-easy boxes, IoT for newbies packs, Raspberry Pi for girls and what not, overlook. All these campaigns at schools to teach children to build and program line-following robots; the BBC micro:bit. They just provide parts, add one or two basic examples, and leave the rest to the user. But users want fully worked-out applications (to assemble if you must) that are useful in their daily lives. Who needs a line-following robot? Only if many of such complete application examples are provided, then maybe a few users will become interested and dive a bit deeper and start tinkering themselves.
But... our world is already full of excellent engineering examples and brilliant applications that you can buy and use straight out of the box, without assembling anything. Great technology is everywhere around us, makes our lives more comfortable, and yet only few people pursue a career in engineering. If it has always been like that, then it must be something genetic. Maybe we must engineer the human to get the human to engineer?