Existing tutorial examples supplied as ready-to-load tables. Load one with
data("mfrmr_example_gtheory_crossed", package = "mfrmr"), then assign it
to a shorter name such as ratings. No simulation code is needed to run
the G-study or D-study examples.
Format
Two data frames with five columns: Person, Rater, Task,
Content, and Organization. Each row is one Person-by-Rater-by-Task
rating cell; the last two columns are continuous numeric scores.
mfrmr_example_gtheory_crossed has 1,920 rows (40 Persons, eight common
raters and six common tasks). mfrmr_example_gtheory_nested has 540 rows
(30 Persons, six tasks and three raters per task).
Source
The existing crossed G-theory tutorial (seed 24) and nested
G-study help example (seed 2027). Their unchanged generation recipes are
retained in data-raw/make-teaching-data.R in the source repository.
Details
In the crossed example, the same rater ID denotes the same rater on every
task. In the nested example, R1, R2, R3 are local labels: the pair
Task and Rater identifies a rater, giving 18 raters in total. Declare
nesting = c(Rater = "Task") for that example. Persons are crossed with
all tasks and rater teams in both examples.
The scores were constructed from independent normal random effects for every source in the respective design. The nested example correlates the two score components within each source. These are fictional unbounded continuous scores, not ordered rubric categories or MFRM ability estimates. Their design and sample sizes demonstrate the functions; they do not establish that an assessment's assumptions or estimation precision are adequate.
Examples
data("mfrmr_example_gtheory_crossed", package = "mfrmr")
ratings <- mfrmr_example_gtheory_crossed
head(ratings)
#> Person Rater Task Organization Content
#> 1 1 1 1 0.5045747 0.5475769
#> 2 2 1 1 1.3440571 -0.9615514
#> 3 3 1 1 -0.4715306 3.7099968
#> 4 4 1 1 -0.5425612 0.5795908
#> 5 5 1 1 -0.3337672 2.8174334
#> 6 6 1 1 1.6440213 0.1798085
g <- mfrm_g_study(ratings, scores = c("Content", "Organization"),
person = "Person", rater = "Rater", task = "Task")
g$component_diagnostics
#> Source MinimumScaledEigenvalue Tolerance PositiveSemidefinite Rank
#> 1 Person 0.12937911 1.490116e-08 TRUE 2
#> 2 Rater 0.07993328 1.490116e-08 TRUE 2
#> 3 Task 0.03814013 1.490116e-08 TRUE 2
#> 4 Person:Rater 0.13711601 1.490116e-08 TRUE 2
#> 5 Person:Task 0.12143229 1.490116e-08 TRUE 2
#> 6 Rater:Task 0.14198058 1.490116e-08 TRUE 2
#> 7 Residual 0.14284035 1.490116e-08 TRUE 2
#> Dimension
#> 1 2
#> 2 2
#> 3 2
#> 4 2
#> 5 2
#> 6 2
#> 7 2
# Inspect component estimates before projecting future plans.
d <- mfrm_d_study(g,
design_grid = data.frame(Raters = c(2, 4), Tasks = 3),
weights = c(Content = .5, Organization = .5))
summary(d)
#> Scenario Raters Tasks Kind Score UniverseVariance
#> 1 1 2 3 Score Content 0.9492992
#> 2 1 2 3 Score Organization 0.9307520
#> 3 1 2 3 Composite Composite 0.4257850
#> 4 2 4 3 Score Content 0.9492992
#> 5 2 4 3 Score Organization 0.9307520
#> 6 2 4 3 Composite Composite 0.4257850
#> RelativeErrorVariance AbsoluteErrorVariance G Phi RelativeSEM
#> 1 0.9844251 2.1219180 0.4909175 0.3090954 0.9921820
#> 2 0.9712316 1.5667193 0.4893586 0.3726778 0.9855108
#> 3 0.4393618 0.9932672 0.4921535 0.3000488 0.6628437
#> 4 0.6456467 1.4837718 0.5951921 0.3901650 0.8035214
#> 5 0.6290735 0.9708793 0.5967026 0.4894492 0.7931415
#> 6 0.2846264 0.6629832 0.5993499 0.3910704 0.5335039
#> AbsoluteSEM Status GStatus PhiStatus RelativeSEMStatus AbsoluteSEMStatus
#> 1 1.4566805 Available Available Available Available Available
#> 2 1.2516866 Available Available Available Available Available
#> 3 0.9966279 Available Available Available Available Available
#> 4 1.2181017 Available Available Available Available Available
#> 5 0.9853321 Available Available Available Available Available
#> 6 0.8142378 Available Available Available Available Available
#> ComponentPSD ComponentTreatment
#> 1 TRUE raw
#> 2 TRUE raw
#> 3 TRUE raw
#> 4 TRUE raw
#> 5 TRUE raw
#> 6 TRUE raw