There is a dichotomy in framing research that can be traced back to its multidisciplinary origins in psychology and sociology. Definitions of framing rooted in psychology are concerned with the differential presentation of the otherwise identical information and are often referred to as equivalence framing. Definitions rooted in more sociological traditions investigate how a message can be constructed with different sets of information to highlight contrasting perspectives on the same issue. The latter is typically referred to as emphasis framing. Although often subsumed under the same label, equivalence framing and emphasis framing are systematically different, both conceptually and operationally. Therefore, the two traditions need to be carefully distinguished in terms of their origins, conceptualization and operationalization of frames, underlying mechanisms, cognitive outcomes, and their relationships with other media effects theories. Categorizing existing studies revealed two major pitfalls in framing effects literatures. First, many political communication studies to date have adopted the emphasis framing approach. However, as substantial manipulation of information introduces confounding variables making it difficult for researchers to attribute the effect on the audience to the change of frames, this approach has relatively low internal validity in experiments and can hardly be distinguished from other cognitive media effects models, such as agenda setting and priming. Thus, the bias toward emphasis framing needs to be addressed by conducting research with equivalence frames so that a more concrete causal relationship between message framing and its effects can be established. In addition, little attention has been given to visuals in framing effects research so far. Considering that people consume information in a multimedia environment online, visual frames and verbal-visual interactions need to be further investigated.