"Why do people vote against their own best interest?" This question defines one of the more perplexing paradoxes of the effect of verbal behavior development on human performative behavior (aka "voting"). Framing Theory (the study of how to present information through public forums to bias the effect of various interlocking contingencies that influence individual behavior) offers a tool but not an explanation of how such behavioral paradoxes evolve. Framing Theory is all about how to effct bias and to influence the performative behavior of others. For example, see the techniques employed by Faux News to bias and influence the performative behavior of its regular viewers. Such daily contact with such sustained "Frames" maintains the strength of an established bias and partitions that bias as a separate VB repertoire (aka "narrative") isolated from and detached from the influence of other VB structures or narratives.
From a verbal behavior perspective this bias formation and VB partitioning occurs within the skin of the same individual and within that person's otherwise unified verbal repertoire. This leads to a question, "How does a unified verbal structure become partitioned into such disparate and contradictory narrative repertoires?" The complexity of what should be discussed here versus the simplicity of what will be discussed here must be forgiven by the reader, for that complexity requires a tome requiring a writer repertoire that I've not yet developed. So here's one take on what might be going on within the evolution of a verbal repertoire to create an individual who acknowledges concrete and extended relational frames (i.e., the names of observable object features and object collections -- that a toothbrush has a handle, bristles, a length, a shape, a function, and is typically found with a tube of toothpaste, a container of dental floss, and is typically found in a bathroom cabinet or on a bathroom counter). Even the most politically (narrative) biased individual is likely to comport his or her verbal behavior to these frames. It is when the narrative is reliant on predominately arbitrary relations is when we most often see egregious bias take hold. To understand this outcome from an RFT perspective a slight detour and presentation on the structure and nature of Mobius Strips is required.
At its core, a Mobius Strip is a 2D structure, a rectangle, that is shaped into a 2D circle with some extension of its outer bondary and dimunition of its inner boundary, allowing the opposite ends of the strip to come together at a common coordinate. It is what is done at this common coordinate that is what turns the strip into a manifold of unique properties. That is, at the location of the common coordinates, one of the ends is flipped in OPPOSITION to its counterpart, and what used to be A1...Ax and B1...Bx frames of COORDINATION are, at this location A1....Bx and B1...Ax coordinations, but ONLY at this intersection. One would expect that this flipping of frames would have a disruptive impact on the entirety of the relational network (aka "strip") in which this "end flipping" occurs. Most likely it does and one of the disruptors may be to split the entire network into parts that contain predominately non-arbitrary to purely arbitrary relations, with other splits container various proportions of both. This is where understanding what happens with the longitudinal partitioning of Mobius Strips suggests what is occurring. When you longitudinally partition a Mobius Strip in two its boundary remains both intact and CONTINUOUS -- i.e., there is only one boundary. However, when you cut the strip into three partitions, you create two separate bounded structures, with a smaller structure caught forming a ring that is traversed by the larger and also continuous structure. This allows the smaller structure to move the entire length of the larger structure, much like an orbiting moon, forever constrained in its path and movement by the shape of the larger strip. If we note that the larger strip represents the networking of both primary non-arbitrary and transitory arbitrary networks -- that generally maintain contact with prevailing social norms verbal behavior use -- and that the smaller, self-contained but constrained structure that is free to conform to its own internal "narrative," then we have a model for how verbal behavior becomes unshackled to reality.