EXPERIMENTAL DESIGN AND DATA
The focus of the study is to determine whether accountability and sanction awareness promote
increased reporting of use tax and whether sanction awareness increases B&O tax reporting for the
third quarter of 2003 (3rd Q). The experimental treatments were implemented under the guise of
the DOR sending construction firms educational tax-related information. As illustrated in Table 1,
the experimental design is a 2 X 2 design. For the accountability manipulation, half of the taxpayers
were required to file signed affidavits stating that all use taxes had been paid (i.e., there
were no additional use taxes due); the other subjects did not receive such a request. Likewise, for
the penalty manipulation, half of the taxpayers received a letter that discussed penalties applicable
to nonpayment of Washington taxes, and the remaining subjects' letters did not contain the penalty
information. Thus, a taxpayer could receive a letter with no manipulations (control), only the
accountability manipulation, only the sanction manipulation, or both.^ The third non-manipulated
factor was the firms' revenue state. Firms were partitioned based on current revenue states (improving
versus declining over the preceding period).* This causes the experiment to have a
2 X 2 X 2 design.
The DOR randomly selected the experimental sample from construction firms located in
Washington that filed excise tax returns in 2002. Specifically, the DOR used a random number
generator to place taxpayers into each group.