A Contradiction Left Unresolved?

Here are a quotes from two different Subcommittees of Social Security's Occupational Information Development Advisory Panel (emphasis added):

The Physical Demands Subcommittee strongly believes that data collection for the physical demands of work cannot be done by self-report. [See below. Another Subcommittee wants to give self-reporting a try.] There are numerous studies that demonstrate that self-reported physical demands are neither reliable nor valid ... (Physical Demands Subcommittee report, pages 34-35 of pdf)

Recommendation: The SSA should train existing Experts [Are we talking about Vocational Experts who testify at ALJ hearings or state agency vocational specialists or both?] in the new OIS [Occupational Information System] and use them as a source to provide job level data for the pilot study. The SSA should also provide job incumbents with the opportunity to provide job level data in the pilot study and compare the quality of results from the two sources. (Work Taxonomy and Classification Subcommittee report, page 282 of PDF )

The references are to the PDF of the entire Panel report. It looks to me as if there is a serious contradiction about a key matter. I do not see that this contradiction was ever resolved by the entire Panel.

I do not understand why the idea of surveying workers about their job duties should even come up. If you spend any time interviewing people about their job duties you quickly discover that when you ask people detailed questions you get confused answers that run along the lines of "I never weighed what I was lifting" and "I never timed how long I was on my feet" and "I never measured how far I had to walk." When pressed, people just give wild guesses. Surveying workers about their job duties may sound plausible, but it would yield only confusing, inconsistent, meaningless data.