Although often credited with prompting a paradigm shift in social comparison theory, T. A. Wills's (1981) downward comparison (DC) theory has received some criticism recently. In particular, several recent studies have failed to find support for T. A. Wills's (1981) contention that threat and accompanying negative affect lead to a desire for DC. These apparent failures have led some investigators to question basic principles of the theory. To resolve this controversy, 5 studies were conducted examining preferred comparison level (PCL) after performance; 4 of the studies also assessed change in this preference. Results supported DC theory, but with modifications. Specifically, individuals who performed poorly lowered their PCLs. Under some circumstances, this "downward shift" included an increased interest in "true" DC--comparing with worse-off others. A reconciliation of these results with those of previous studies is offered.