Suboptimal choice: A review and quantification of the signal for good news (SiGN) model.

Abstract

As first reported several decades ago, pigeons (Columba livia) sometimes choose options that provide less food over options that provide more food. This behavior has been variously referred to as suboptimal, maladaptive, or paradoxical because it lowers overall food intake. A great deal of research has been directed at understanding the conditions under which animals and people make suboptimal choices and the mechanisms that drive this behavior. Here, we review the literature on suboptimal choice and the variables that play a role in this phenomenon. Suboptimal choice is most likely to occur when the outcomes following a choice are uncertain, when the outcomes are delayed after the choice, and when the outcomes are signaled only on the option that provides food less often. We propose a mathematical formalization of the signal for good news (SiGN) model which assumes that a signal for a reduction in delay to food reinforces choice. We generate predictions from the model about the effect of parameters that characterize suboptimal choice and we show that, even in the absence of free parameters, the SiGN model provides a very good fit to the choice proportions of birds from a large set of conditions across studies from numerous researchers. R code for SiGN predictions and the data set are available on the Open Science Framework (https://osf.io/39qtj). We discuss limitations of the model, propose directions for future research, and discuss the general applicability of this research to understanding how rewards and signals for reward may combine to reinforce behavior.

On the value of advanced information about delayed rewards

“…On average, pigeons took 16 sessions to complete each long/short ratio (range: [15][16][17][18][19][20][21][22][23][24][25]. Table 1 shows the number of sessions per ratio for each bird.…”

“…Even though the non-informative option usually yields more reward (2.5 times more in the gure), human gamblers (Molet et al, 2012), pigeons (e.g., Fortes et al, 2018;Gipson et al, 2009;Smith et al, 2016;Stagner & Zentall, 2010), starlings (Vasconcelos et al, 2015), rhesus macaques (Blanchard et al, 2015), and rats (e.g., Ajuwon et al, 2023;Chow et al, 2017; Cunningham & Shahan, 2019, 2020) prefer the informative option. This seemingly paradoxical preference generated an intense empirical and theoretical effort to understand its causes (e.g., Cunningham & Shahan, 2018;Daniels & Sanabria, 2018;Dunn et al, 2023;González et al, 2020a;McDevitt et al, 2016;Vasconcelos et al, 2018;Zentall, 2016). Whatever the mechanism involved, this preference for informative signals when they do not yield any tangible instrumental bene t is reminiscent of the information-seeking hypothesis rst suggested in the 'observing response' literature.…”

Differences in paradoxical choice between pigeons (Columba livia) and rats (Rattus norvegicus): The problem of cue trackability.

“…Because of delay aversion (e.g., Mazur & Biondi, 2009), this may have directly contributed to their preference for the suboptimal option after all TL durations were equalized. Beyond this fact, the consistency tracking hypothesis cannot account for the slow emergence of paradoxical choice with consistent associations in both options, and other interpretations have been provided (e.g., Dunn et al, 2024; Laude et al, 2014; Stagner & Zentall, 2010). Nevertheless, paradoxical preference might somehow relate to the informational content of the signaled 50% option relative to the 100% option.…”

“…In the opposite, a lower contrast (sometimes no contrast) exists with the IL noninfo alternative (in the example, 20% vs. 50%), which is therefore not preferred. Alternatively, the signal-for-good-news or SiGN theory (Dunn et al, 2024; McDevitt et al, 2016) attributes preference for the IL info alternative to the reinforcing value of the onset of the TL+ in this alternative (which improves local context, as a TL− might occur), compared to the TL± (or even a fully predictable TL+, see further) in the other alternative (which does not improve local context). Both theories may sound similar, the former insisting on the importance of probability contrast for paradoxical choice, while the latter puts forward the reinforcing value of delay reduction once the individual knows what is coming.…”

On the value of advanced information about delayed rewards

“…This seemingly paradoxical preference generated an intense empirical and theoretical effort to understand its causes (e.g., Cunningham and Shahan 2018 ; Daniels and Sanabria 2018 ; Dunn et al 2023 ; González et al 2020a ; McDevitt et al 2016 ; Vasconcelos et al 2018 ; Zentall 2016 ). Whatever the mechanism involved, this preference for informative signals when they do not yield any tangible instrumental benefit is reminiscent of the information-seeking hypothesis first suggested in the ‘observing response’ literature.…”