A Mixed Effect Location-Scale Model with Mixture Distributed Scale Random Effects to Analyze (Near) Identical Entries in Ecological Momentary Assessments.

A Mixed Effect Location-Scale Model with Mixture Distributed Scale Random Effects to Analyze (Near) Identical Entries in Ecological Momentary Assessments.
复制标题

具有混合分布尺度随机效应的混合效应位置尺度模型,用于分析生态瞬时评估中的(接近)相同条目。

DOI:
10.1080/00273171.2020.1857217
复制
发表时间:
2021
影响因子:
3.8
通讯作者:
Hedeker,Donald
Hedeker,Donald
中科院分区:
心理学3区
文献类型:
--
作者:
Ma,Qianheng;Dunton,GenevieveF;Hedeker,Donald

文献摘要

相似文献

生态瞬时评估(EMA)研究提供了足够的信息,可以使用混合效应位置尺度模型(MELS,Hedeker 等,2008)对均值(位置)和受试者内方差(尺度)进行建模。但在 EMA 研究中,受试者可以随着时间的推移提供(接近)零方差的响应,这会给 MELS 估计带来问题。我们扩展了 MELS 来解决这个问题。在传统的两水平 MELS 中,特定于受试者的位置随机效应 i 影响平均值,尺度随机效应 xi 影响受试者在时间 j 时响应的变异性,即 yij $ NðXT ijb þ i, expðW T ijs þ xiÞÞ:但 xi 中可能存在潜在簇,以区分具有不同方差模式的受试者。而不是假设 xi $N 0, r2 x
Ecological Momentary Assessment (EMA) studies provide sufficient information to model both the mean (location) and the within-subject variance (scale) using the mixed effect location-scale model (MELS, Hedeker et al., 2008). But in EMA studies, subjects can provide responses with (near) zero variance over time which causes problems for MELS estimation. We extended MELS to address this issue. In conventional two-level MELS, the subject-specific location random effect i affects the mean, and a scale random effect xi affects the variability of the subjectes response at time j, ie, yij $ NðXT ijb þ i, expðW T ijs þ xiÞÞ: But latent clusters may exist in xi to distinguish subjects with different variance patterns. Instead of assuming xi $ N 0, r2 x