Across six studies, we validated a new measure of helpful and hurtful behaviors, the Tangram Help/Hurt Task. Studies 1 to 3 provided cross-sectional correlational convergent and discriminant validity evidence for the Tangram Task using college-based and adult online samples. Study 4 revealed that previously validated empathy primes increase helpful behaviors on the Tangram Task. Studies 5 and 6 revealed that previously validated provocation manipulations increase hurtful behaviors on the Tangram Task. The effects of various experimental manipulations on the Tangram Task were similar to or larger than on other established indices of helpful and hurtful behaviors. In addition, motivation items in all studies indicate that tangram choices are indeed associated with the intent of helping and hurting. We discuss the advantages and limitations of the Tangram Help/Hurt Task relative to established measures of helpful and hurtful behaviors.