Comparing N = 1 effect size indices in presence of autocorrelation
Generalization from single-case designs can be achieved by means of replicating individual studies across different experimental units and settings. When replications are available, their findings can be summarized using effect size measurements and integrated through meta-analyses. Several procedur...
| Autores: | , |
|---|---|
| Formato: | artículo |
| Estado: | Versión aceptada para publicación |
| Fecha de publicación: | 2008 |
| País: | España |
| Recursos: | Universidad de Barcelona |
| Repositorio: | Dipòsit Digital de la UB |
| OAI Identifier: | oai:diposit.ub.edu:2445/34753 |
| Acesso em linha: | https://hdl.handle.net/2445/34753 |
| Access Level: | acceso abierto |
| Palavra-chave: | Investigació de cas únic Correlació (Estadística) Single subject research Correlation (Statistics) |
| Resumo: | Generalization from single-case designs can be achieved by means of replicating individual studies across different experimental units and settings. When replications are available, their findings can be summarized using effect size measurements and integrated through meta-analyses. Several procedures are available for quantifying the magnitude of treatment"s effect in N = 1 designs and some of them are studied in the current paper. Monte Carlo simulations were employed to generate different data patterns (trend, level change, slope change). The experimental conditions simulated were defined by the degrees of serial dependence and phases" length. Out of all the effect size indices studied, the Percent of nonoverlapping data and standardized mean difference proved to be less affected by autocorrelation and perform better for shorter data series. The regression-based procedures proposed specifically for single-case designs did not differentiate between data patterns as well as simpler indices. |
|---|