ction standard under section 315 is still in its
experimental stages, it may promote accuracy and help to bring about
scientific amendments of the present law to illustrate in this
investigation the possible danger of using the commission's figures
to fortify different and inconsistent conclusions. The data obtained
by the commission in the straw-hat investigation are unsatisfactory
in the following particulars:

_Representativeness of samples._--In selecting hats assumed to be
representative of American production, it was found impracticable
to determine the respective percentages of production of cheap,
medium-priced, and high-priced hats. In consequence there is some
reason to believe that the limited figures secured with respect to cheap
American hats has tended to exaggerate American costs beyond what an
exactly representative selection would have shown. Figures were secured
for only a few producers of cheap American hats, and while it is
impossible to say what weight should be given to such cheap American
production, expert opinion is not wanting in support of the view that
because of the method of sampling employed, American costs as a whole
have been unduly elevated for comparison with Italian costs. While it is
too late to make any exact mathematical adjustment on this account, it
is only fair to urge distinct caution in accepting at their face value
and following to their inexorable conclusions the comparisons based on
the domestic and foreign data.

Probably the most important principle of sampling employed by the
commission's agents when confronted with the problem of selecting for
cost comparison a few types of hats from the many manufactured was
the choice of those types of hats with respect to which the domestic
industry has been suffering the keenest competition. It must be clear
that the selection of such hats tended to show the widest cost
divergence for the two countries, since it was to be expected that the
severest competition would have been experienced when the relatively
higher-cost hats of the United States met the relatively lower-cost hats
of Italy. Nor could it be said that such hats as were chosen were the
only "similar competitive articles," since the foreign manufacturers
can and do produce all types and styles sold in the United States. The
fact that the American industry earned approximately 10 per cent on its
invested capital (even after the payment of large salaries) must be
chiefly