Tarlow Kevin R, Brossart Daniel F
School of Public Health.
Department of Educational Psychology.
Sch Psychol Q. 2018 Dec;33(4):590-603. doi: 10.1037/spq0000273. Epub 2018 Oct 4.
Single-case experimental methods are used across a range of educational and psychological research. Single-case data are analyzed with a variety of methods, but no statistic has demonstrated clear superiority over other methods. The time-series nature of single-case designs requires special consideration for baseline trend and autocorrelation when estimating intervention effect size. However, standard correction methods are limited because they assume precise statistical estimation of trend and autocorrelation. Unlike standard correction methods, Monte Carlo simulation methods can address the poor precision of single-case effect size indices. This paper presents the rationale for a new simulation method, Interrupted Time-Series Simulation (ITSSIM). A small field test was also conducted, and ITSSIM performed similarly to sophisticated multilevel methods for single-case research. ITSSIM is accessible as a free software application that requires no prior knowledge of statistical computing or syntax. ITSSIM may be used to estimate the effect size of a single interrupted time-series (AB design), and multiple ITSSIM effect size estimates may be combined via meta-analysis. (PsycINFO Database Record (c) 2018 APA, all rights reserved).
单案例实验方法被广泛应用于一系列教育和心理学研究中。单案例数据通过多种方法进行分析，但没有一种统计方法显示出比其他方法具有明显优势。单案例设计的时间序列性质在估计干预效应大小时需要特别考虑基线趋势和自相关。然而，标准校正方法存在局限性，因为它们假定趋势和自相关的统计估计是精确的。与标准校正方法不同，蒙特卡罗模拟方法可以解决单案例效应大小指标精度较差的问题。本文介绍了一种新的模拟方法——中断时间序列模拟（ITSSIM）的基本原理。还进行了一个小型现场测试，结果表明ITSSIM在单案例研究中与复杂的多水平方法表现相似。ITSSIM作为一个免费的软件应用程序可供使用，无需具备统计计算或语法的先验知识。ITSSIM可用于估计单个中断时间序列（AB设计）的效应大小，并且多个ITSSIM效应大小估计值可通过元分析进行合并。（《心理学文摘数据库记录》（c）2018美国心理学会，保留所有权利）
