A moderated mediation model is a model in which the mediation process is influenced by moderating variables. If a model includes both mediating and moderating variables, it is considered a moderated mediator, with the focus on exploring whether there is a mediating effect and whether the moderating variables have a significant impact on the mediating effect, that is, whether there is a moderating effect.

    Data description:

    Background description:

    Mediation analysis with moderation is an analytical method used to explore the changes in the relationship between the dependent variable and the independent variable under the influence of moderating variables. Here, we investigate the mediating effect of A4 on A1 through A2 through path analysis, and further analyze it by adjusting for different levels of variable A2.

    The analysis results are as follows:

    The results of model I showed a significant positive effect of A3 on A1 with a beta value of 0.5912 and a p-value of 0.011 (p<0.05).

    The results of model two showed that none of the independent variables played a significant role in influencing the dependent variable A2 (p>0.05).

    The BootCI confidence interval for the path A4 => A2 => A1 is [-0.048, 0.031], which includes 0, indicating that the mediation effect of this path is not significant, i.e., A2 does not play a significant mediating role in A4 and A1, nor does it explain the relationship between the cause and outcome variables.

    The 95% BootCI confidence interval at low level (-1SD) for this path was [-0.143, 0.181] including 0, indicating that the mediation effect of this path was not significant, i.e., A2 did not significantly mediate in A4 and A1 and did not explain the relationship between the cause and outcome variables.

    The 95% BootCI confidence interval at the mean level for this path is [-0.248, -0.011], excluding 0, indicating a significant mediation effect for this path, i.e., A2 plays a significant mediating role in A4 and A1, with a total effect value of 0.181 and a mediation effect value of -0.1.

    The 95% BootCI confidence interval at high level (+1SD) for this path was [-0.462, 0.079], including 0, indicating that the mediation effect of this path was not significant, i.e., A2 did not play a significant mediating role in A4 and A1, and did not explain the relationship between the cause variable and the outcome variable.

    In summary, the 95% BootCI confidence interval at the low level (-1SD) of this path is [-0.143, 0.181], excluding 0, indicating that the mediation effect of this path is not significant, i.e., A2 does not play a significant mediating role in A4 and A1, and it does not explain the relationship between the cause variable and the outcome variable. The 95% BootCI confidence interval at the mean level of the path is [-0.248, -0.011], excluding zeros, indicating that the mediation effect of the path is significant, i.e., A2 plays a significant mediating role in A4 and A1, with a total effect value of 0.181 and a mediation effect value of -0.1. The 95% BootCI confidence interval at the high level (+1SD) of the path is [-0.462, 0.079], including 0, indicating that the mediation effect of this path is not significant, i.e., A2 does not play a significant mediating role in A4 and A1, and does not explain the relationship between the cause and outcome variables.

    Reference:

    [1]方杰,张敏强,顾红磊,梁东梅.(2014).基于不对称区间估计的有调节的中介模型检验.心理科学进展,22(10),1660-1668.

    [2]黄彦,谢晓琳,周晖.(2016).亲子依恋与儿童问题行为的相关:师生关系的调节作用.中国临床心理学杂志,24(6),1074-1078.

    [3]聂光辉,吴俊端,唐峥华,韦波,杨莉.(2014).工作场所欺负与护士心理健康的关系:自我效能感的调节作用.中国临床心理学杂志,22(5),901-903.

    [4]温忠麟,刘红云,侯杰泰.(2012).调节效应和中介效应分析.北京:教育科学出版社.

    [5]温忠麟,吴艳,侯杰泰.(2013).潜变量交互效应结构方程:分布分析方法.心理学探新,33(5),409-414.

    [6]温忠麟,叶宝娟.(2014).有调节的中介模型检验方法:竞争还是替补?心理学

    [7]Ledgerwood, A., & Shrout, P. E. (2011). The trade-off between accuracy and precision in latent variable models of mediation processes. Journal of Personality and Social Psychology, 101(6), 1174-1188.

关注微信公众号发送【示例数据】获取SPSSMAX练习示例数据。