Walters Caroline E, Kendal Jeremy R
Department of Mathematics, University of Durham, DH1 3LE, UK; Department of Anthropology and the Centre for the Coevolution of Biology and Culture, University of Durham, DH1 3LE, UK.
Theor Popul Biol. 2013 Dec;90:56-63. doi: 10.1016/j.tpb.2013.09.010. Epub 2013 Sep 23.
Epidemiological models have been applied to human health-related behaviors that are affected by social interaction. Typically these models have not considered conformity bias, that is, the exaggerated propensity to adopt commonly observed behaviors or opinions, or content biases, where the content of the learned trait affects the probability of adoption. Here we consider an interaction of these two effects, presenting an SIS-type model for the spread and persistence of a behavior which is transmitted via social learning. Uptake is controlled by a nonlinear dependence on the proportion of individuals demonstrating the behavior in a population. Three equilibrium solutions are found, their linear stability is analyzed and the results are compared with a model for unbiased social learning. Our analysis focuses on the effects of the strength of conformity bias and the effects of content biases which alter a conformity threshold frequency of the behavior, above which there is an exaggerated propensity for adoption. The strength of the conformity bias is found to qualitatively alter the predictions regarding whether the trait becomes endemic within the population and the proportion of individuals who display the trait when it is endemic. As the conformity strength increases, the number of feasible equilibrium solutions increases from two to three, leading to a situation where the stable equilibrium attained is dependent upon the initial state. Varying the conformity threshold frequency directionally alters the behavior invasion threshold. Finally we discuss the possible application of this model to binge drinking behavior.
流行病学模型已应用于受社会互动影响的人类健康相关行为。通常，这些模型没有考虑从众偏差，即采用普遍观察到的行为或观点的过度倾向，也没有考虑内容偏差，即所学特征的内容会影响采用的概率。在此，我们考虑这两种效应的相互作用，提出一种SIS型模型，用于描述通过社会学习传播和持续存在的行为。行为的采用受对群体中表现出该行为的个体比例的非线性依赖控制。我们找到了三个平衡解，分析了它们的线性稳定性，并将结果与无偏差社会学习模型进行了比较。我们的分析重点在于从众偏差强度的影响以及改变行为从众阈值频率的内容偏差的影响，超过该阈值频率，采用行为的倾向会被夸大。研究发现，从众偏差的强度会定性地改变关于该特征在群体中是否会成为地方病以及当它成为地方病时表现出该特征的个体比例的预测。随着从众强度的增加，可行平衡解的数量从两个增加到三个，导致达到的稳定平衡取决于初始状态的情况。定向改变从众阈值频率会改变行为入侵阈值。最后，我们讨论了该模型在酗酒行为方面的可能应用。
