View source: R/REML-ES-functions.R

g_REML | R Documentation |

Estimates a design-comparable standardized mean difference effect size based on data from a multiple baseline design, using adjusted REML method as described in Pustejovsky, Hedges, & Shadish (2014). Note that the data must contain one row per measurement occasion per case.

```
g_REML(
m_fit,
p_const,
r_const,
X_design = model.matrix(m_fit, data = m_fit$data),
Z_design = model.matrix(m_fit$modelStruct$reStruct, data = m_fit$data),
block = nlme::getGroups(m_fit),
times = attr(m_fit$modelStruct$corStruct, "covariate"),
returnModel = TRUE
)
```

`m_fit` |
Fitted model of class lme, with AR(1) correlation structure at level 1. |

`p_const` |
Vector of constants for calculating numerator of effect size.
Must be the same length as fixed effects in |

`r_const` |
Vector of constants for calculating denominator of effect size.
Must be the same length as the number of variance component parameters in |

`X_design` |
(Optional) Design matrix for fixed effects. Will be extracted from |

`Z_design` |
(Optional) Design matrix for random effects. Will be extracted from |

`block` |
(Optional) Factor variable describing the blocking structure. Will be extracted from |

`times` |
(Optional) list of times used to describe AR(1) structure. Will be extracted from |

`returnModel` |
(Optional) If true, the fitted input model is included in the return. |

A list with the following components

`p_beta` | Numerator of effect size |

`r_theta` | Squared denominator of effect size |

`delta_AB` | Unadjusted (REML) effect size estimate |

`nu` | Estimated denominator degrees of freedom |

`kappa` | Scaled standard error of numerator |

`g_AB` | Corrected effect size estimate |

`V_g_AB` | Approximate variance estimate |

`cnvg_warn` | Indicator that model did not converge |

`sigma_sq` | Estimated level-1 variance |

`phi` | Estimated autocorrelation |

`Tau` | Vector of level-2 variance components |

`I_E_inv` | Expected information matrix |

Pustejovsky, J. E., Hedges, L. V., & Shadish, W. R. (2014).
Design-comparable effect sizes in multiple baseline designs: A general modeling framework.
*Journal of Educational and Behavioral Statistics, 39*(4), 211-227. \Sexpr[results=rd]{tools:::Rd_expr_doi("10.3102/1076998614547577")}

```
data(Laski)
Laski_RML <- lme(fixed = outcome ~ treatment,
random = ~ 1 | case,
correlation = corAR1(0, ~ time | case),
data = Laski)
summary(Laski_RML)
g_REML(Laski_RML, p_const = c(0,1), r_const = c(1,0,1), returnModel=FALSE)
data(Schutte)
Schutte$trt.week <- with(Schutte, unlist(tapply((treatment=="treatment") * week,
list(treatment,case), function(x) x - min(x))) + (treatment=="treatment"))
Schutte$week <- Schutte$week - 9
Schutte_RML <- lme(fixed = fatigue ~ week + treatment + trt.week,
random = ~ week | case,
correlation = corAR1(0, ~ week | case),
data = subset(Schutte, case != 4))
summary(Schutte_RML)
Schutte_g <- g_REML(Schutte_RML, p_const = c(0,0,1,7), r_const = c(1,0,1,0,0))
summary(Schutte_g)
```

Embedding an R snippet on your website

Add the following code to your website.

For more information on customizing the embed code, read Embedding Snippets.