Did Epidemiologists Grossly Underestimate The Public's Willingness And Ability To Reduce Daily Social Contacts?

From Gallup, Gallup Blog, April 21, 2020, "Americans' Social Contacts During the COVID-19 Pandemic" by Jonathan Rothwell:

Adults practicing social distancing generate at least 90% fewer contacts per day than those who are making little effort to social distance, according to new Gallup data. Those who completely or mostly isolate themselves generate about five contacts per day, compared with an average of 52 for those not attempting to isolate themselves.***Yet, closing workplaces may not be the only way to mitigate contact and suppress viral transmission. Workers who visit their work site but have made changes in how they do their job reduced contacts by 29% to 69%.

Why Measuring Contacts Matters
In epidemiological models, social distancing has large effects on expected infection growth. Yet, different modeling teams have used varying assumptions for the efficacy of social distancing when it comes to reducing contacts. Getting these estimates right would thus appear crucial to accurately understanding the risks and rewards of different mitigation policies.

For example, in research published by public health scholars at Columbia University (PDF download), these scholars have assumed that social distancing can reduce at most 40% of daily contacts. An influential model from Imperial College London defined "social distancing of the entire population" as reducing workplace contacts by 25%, household contacts by 25%, and social or nonwork contacts by 75%.***Only 3% of adults and 4% of workers have not made any attempt to isolate themselves.