Another issue with experiments is that they are not realistic scenarios. For example, in the Economist experiment no one actually paid any money. Which makes it basically worthless. We’ve often found in economics research that what people say they will do and what they actually do is very different. Moreover, studies only find that some % act irrationally but then it’s reported as if everyone did.
Now obviously humans are not totally rational, but it does us a disservice to act as if we are so terrible at decisions all of the time.
In the first experiment I can choose if I'd rather have online (59$) or print version(125$). Either way I'm risking the disappointment of taking the wrong choice, or I can choose online for 59$ plus a 66$ hedge against choosing wrong.
In the second experiment you've eliminated the risk of choosing print and realizing later that I wanted online, so hedging makes less sense (i.e. I'm willing to pay less for hedging my choice).
If there’s an easy to pick offer bundling the two, I’d rather pay the extra to keep the choice open, that is, to choose later.
Most of his experiments I read about have other explanations that fit the data just as well. For instance, the "The Economist" subscription has another explanation to simple comparison, choice deferral.
People don't really know if they'll want the online or print version, so are happy to pay extra for not having to make that choice now. If you only give them the two choices, you force them to choose. I'm not saying he is wrong, just that I miss him considering other alternatives (that to me sound way more plausible, but are less publication-sexy).