Package {roundRobinR}

Title:	Manipulate and Analyze Round Robin Dyadic Data
Version:	2.0.0
Description:	Provides utilities for processing and analyzing dyadic data collected using a round-robin design, in which each person in a group rates or interacts with every other person on at least one variable. Data manipulation functions prepare datasets for dyadic data analysis by creating the actor and partner dummy variables required by the social relations model (SRM). Analysis functions implement the SRM using multilevel modeling via a custom 'nlme' covariance class ('pdSRM'), following the approach of Snijders and Kenny (1999) <doi:10.1111/j.1475-6811.1999.tb00204.x> and Knight and Humphrey (2019) <doi:10.1037/0000115-019>. The package estimates group, actor, partner, and relationship variance components along with generalized and dyadic reciprocity correlations, and supports both null and fixed-effects models.
License:	MIT + file LICENSE
Encoding:	UTF-8
LazyData:	true
Depends:	R (≥ 3.5.0)
Imports:	data.table (≥ 1.14.0), Matrix, nlme (≥ 3.1-150), stats
Suggests:	numDeriv, testthat (≥ 3.0.0), knitr, rmarkdown
Config/testthat/edition:	3
VignetteBuilder:	knitr
URL:	https://github.com/andrewpknight/roundRobinR
BugReports:	https://github.com/andrewpknight/roundRobinR/issues
Config/roxygen2/version:	8.1.0
NeedsCompilation:	no
Packaged:	2026-09-17 18:29:12 UTC; apk
Author:	Andrew Knight [aut, cre, cph]
Maintainer:	Andrew Knight <knightap@wustl.edu>
Repository:	CRAN
Date/Publication:	2026-09-17 20:40:08 UTC

Extract Coefficients from a pdSRM Object

Description

Internal method called by coef.pdMat to extract the three underlying SRM parameters: actor standard deviation, partner standard deviation, and actor-partner correlation.

Usage

## S3 method for class 'pdSRM'
coef(object, unconstrained = TRUE, ...)

Arguments

Value

a named numeric vector with elements std. dev-a, std. dev-p, and corr.

Examples

d <- createDummies(
  group.id = "groupId", act.id = "actId", part.id = "partId",
  d = sampleDyadData[sampleDyadData$timeId == 1, ],
  merge.original = TRUE
)
o <- nlme::lme(
  liking ~ 1,
  random = list(groupId = nlme::pdBlocked(list(
    nlme::pdIdent(~1),
    pdSRM(~ -1 + a1 + a2 + a3 + a4 + p1 + p2 + p3 + p4)
  ))),
  correlation = nlme::corCompSymm(form = ~1 | groupId / pdSRM_dyad_id),
  data = d,
  na.action = stats::na.omit
)

Extract Correlation Matrix from a pdSRM Object

Description

Internal method called by corMatrix to reconstruct the correlation matrix from the compressed SRM parameters.

Usage

## S3 method for class 'pdSRM'
corMatrix(object, ...)

Arguments

Value

the correlation matrix corresponding to the positive-definite matrix represented by object, with a "stdDev" attribute giving the standard deviations

Examples

d <- createDummies(
  group.id = "groupId", act.id = "actId", part.id = "partId",
  d = sampleDyadData[sampleDyadData$timeId == 1, ],
  merge.original = TRUE
)
o <- nlme::lme(
  liking ~ 1,
  random = list(groupId = nlme::pdBlocked(list(
    nlme::pdIdent(~1),
    pdSRM(~ -1 + a1 + a2 + a3 + a4 + p1 + p2 + p3 + p4)
  ))),
  correlation = nlme::corCompSymm(form = ~1 | groupId / pdSRM_dyad_id),
  data = d,
  na.action = stats::na.omit
)

Fit the Co-Partner Social Relations Model from Raw Long-Format Data

Description

cpsrm() is a friendly, tidyverse-style entry point for the Co-Partner SRM. Give it raw long-format data and the names of the outcome, actor, and group columns; it builds the actor/partner dummy matrices with create_cp_dummies and fits the model with cpsrm_run. Use cpsrm_run directly when you need to supply your own dummy matrices (e.g. dummies built once and reused across several model calls).

Usage

cpsrm(
  data,
  dv,
  actor_id,
  group_id,
  location_id = NULL,
  weight_partners = TRUE,
  ...
)

Arguments

Value

an object of class "cpsrm"; see cpsrm_run for the full return value. The dummy-construction output from create_cp_dummies is attached as $dummies for reference (e.g. to inspect $dummies$group_size_table).

See Also

cpsrm_run for the full-control interface, create_cp_dummies for the dummy-matrix construction this wrapper calls internally

Examples

# cpsrm() expects one row per PERSON per GROUP, with every other group
# member treated as a simultaneous co-partner (e.g. one row per player
# per team, as in three-person golf teams). sampleDyadData is a
# round-robin/directed-dyad dataset (multiple rows per actor per group,
# one per rated partner) and is NOT the right shape for this function --
# see create_dummies()/srm_run() for that design instead. This example
# simulates a small, correctly-shaped dataset: 30 three-person teams
# drawn from a pool of 90 players.
set.seed(1)
team_dat <- do.call(rbind, lapply(1:30, function(g) {
  data.frame(player = sample(1:90, 3), team = g)
}))
team_dat$score <- rnorm(nrow(team_dat), mean = 70, sd = 3)

fit <- cpsrm(
  data         = team_dat,
  dv           = "score",
  actor_id     = "player",
  group_id     = "team",
  zero_rho     = TRUE,     # keep the toy example fast/well-behaved
  se_method    = "none",
  stage1_maxit = 100,
  stage2_maxit = 100
)
print(fit)

Deprecated: use cpsrm_run instead

Description

cpsrmRun has been renamed to cpsrm_run. Key argument changes: course.id -> location_id; actor.dummies -> actor_dummies; partner.dummies -> partner_dummies; zero.rho -> zero_rho; zero.actor -> zero_actor; zero.partner -> zero_partner; zero.group -> zero_group; zero.course -> zero_location; weight.partners -> weight_partners; fixed.effects -> fixed_effects; stage1.maxit -> stage1_maxit; stage2.maxit -> stage2_maxit; se.method -> se_method; rho.boundary -> rho_boundary; verbose.every -> verbose_every.

Usage

cpsrmRun(
  dv,
  actor.id,
  group.id,
  actor.dummies,
  partner.dummies,
  data,
  course.id = NULL,
  zero.rho = FALSE,
  zero.actor = FALSE,
  zero.partner = FALSE,
  zero.group = FALSE,
  zero.course = FALSE,
  weight.partners = TRUE,
  fixed.effects = NULL,
  stage1.maxit = 1000L,
  stage2.maxit = 1000L,
  se.method = c("hessian", "none"),
  rho.boundary = 0.99,
  verbose.every = 50L,
  ...
)

Arguments

Value

see cpsrm_run

Run the Co-Partner Social Relations Model

Description

Fits the Co-Partner Social Relations Model (CP-SRM) using restricted maximum likelihood (REML). The "block" method uses the Woodbury matrix identity for fast computation on large datasets; the "loop" method constructs the full covariance matrix by iterating over groups.

Usage

cpsrm_run(
  dv,
  actor_id,
  group_id,
  data,
  actor_dummies,
  partner_dummies,
  group_sizes = NULL,
  location_id = NULL,
  method = c("block", "loop"),
  zero_rho = FALSE,
  zero_actor = FALSE,
  zero_partner = FALSE,
  zero_group = FALSE,
  zero_location = FALSE,
  weight_partners = TRUE,
  fixed_effects = NULL,
  start = NULL,
  maxit = 200000L,
  stage1_maxit = 1000L,
  stage2_maxit = 1000L,
  tol = 1e-10,
  se_method = c("hessian", "none"),
  rho_boundary = 0.99,
  optimizer = c("Nelder-Mead", "L-BFGS-B"),
  verbose = FALSE,
  verbose_every = 50L
)

Arguments

Value

an object of class "cpsrm"

Reporting partner variance: RAW vs. COMBINED

Because each row sums multiple independent partner effects (its group_size - 1), \sigma_P^2 can be reported on two different, equally valid bases when weight_partners = FALSE:

• RAW (\sigma_P^2 itself): the variance of the partner-effect parameter, on the same per-person basis as \sigma_A^2. Use RAW whenever comparing the magnitude of partner effects to actor effects (e.g. tests of generalized reciprocity).

• COMBINED (partner's row multiplied by group_size - 1): partner identity's total contribution to the variance of a single observed score. Use COMBINED only when reading a full variance decomposition in which all components must sum to the outcome's total variance.

Mixing the two conventions in the same comparison is a common source of error: because COMBINED mechanically multiplies partner's numerator while RAW does not, a naive COMBINED-vs-RAW comparison can make partner effects look several times more (or less) important than actor effects even when the two are, per person, identical in magnitude. print.cpsrm/summary.cpsrm report both whenever weight_partners = FALSE and group size is constant across the data; when group size varies, or weight_partners = TRUE (partner dummies already rescaled to average rather than sum partner effects), only RAW is shown.

Testing variance components: use the boundary-corrected LRT

cpsrm_run reports a Wald z-statistic (estimate / SE) for each variance component, but variance components are bounded below by zero, so the usual two-sided reference distribution is not correct for testing whether a variance component is zero. The standard correction (Self & Liang, 1987; Snijders & Bosker, 2012; Stram & Lee, 1994) is to fit the model with and without the component of interest (e.g. zero_partner = TRUE vs. FALSE), form the likelihood-ratio statistic from each fit's fit["reml_loglik"] (LRT = -2(\ell_{restricted} - \ell_{full})), and compute a ONE-SIDED p-value:

p <- 0.5 * pchisq(LRT, df = 1, lower.tail = FALSE)

Testing TWO variance components jointly at their boundary (e.g. actor and partner both zero) is NOT a simple halving of the df = 2 chi-square, because both parameters are boundary-constrained; it requires the three-part mixture (Self & Liang, 1987):

p <- 0.25 * pchisq(LRT, df = 0, lower.tail = FALSE) +
     0.50 * pchisq(LRT, df = 1, lower.tail = FALSE) +
     0.25 * pchisq(LRT, df = 2, lower.tail = FALSE)

(the df = 0 term is always 0 for any LRT > 0). Applying the plain, uncorrected pchisq p-value, or halving the df = 2 joint test the same way as a single-component test, both misstate significance and are not valid inferential procedures for this model.

See Also

create_cp_dummies

Examples

# Prepare dummies using create_cp_dummies(), bind onto data, then call
# cpsrm_run(). See create_cp_dummies() documentation for a full example.

Deprecated: use create_dummies instead

Description

createDummies has been renamed to create_dummies. The argument names have also changed: group.id -> group_id, act.id -> act_id, part.id -> part_id, d -> data, include.self -> include_self, merge.original -> merge_original.

Usage

createDummies(
  group.id,
  act.id,
  part.id,
  d,
  include.self = FALSE,
  merge.original = FALSE
)

Arguments

Value

see create_dummies

Create Actor and Partner Dummy Matrices for the Co-Partner SRM

Description

Generates the actor and partner dummy matrices required by cpsrm_run. Unlike create_dummies, which produces dummies for the standard Social Relations Model, this function is designed for the Co-Partner SRM where each observation involves one actor and all other members of the group acting simultaneously as partners. Variable group sizes (2, 3, 4, or mixed) are fully supported.

Usage

create_cp_dummies(
  data,
  actor_id,
  group_id,
  weight_partners = TRUE,
  prefix_actor = "A",
  prefix_partner = "P"
)

Arguments

Value

a list with elements:

actor_mat

matrix of actor dummies (n x n_persons)

partner_mat

matrix of partner dummies, optionally weighted (n x n_persons)

actor_names

character vector of actor dummy column names

partner_names

character vector of partner dummy column names

group_sizes

integer vector of length n giving the group size for each observation; pass this to cpsrm_run() via the group_sizes argument

person_levels

character vector of person IDs (column order)

n_persons

integer; number of unique persons

n_groups

integer; number of unique groups

group_size_table

table of group size frequencies

See Also

cpsrm_run, create_dummies

Examples

# create_cp_dummies() expects one row per PERSON per GROUP, with every
# other group member treated as a simultaneous co-partner (e.g. one row
# per player per team, as in three-person golf teams). sampleDyadData is
# a round-robin/directed-dyad dataset (multiple rows per actor per
# group, one per rated partner) and is NOT the right shape for this
# function -- see create_dummies()/srm_run() for that design instead.
# This example simulates a small, correctly-shaped dataset: 30
# three-person teams drawn from a pool of 90 players.
set.seed(1)
team_dat <- do.call(rbind, lapply(1:30, function(g) {
  data.frame(player = sample(1:90, 3), team = g)
}))

dummies <- create_cp_dummies(
  data        = team_dat,
  actor_id    = "player",
  group_id    = "team"
)
cat("Persons:", dummies$n_persons, "\n")
cat("Groups:", dummies$n_groups, "\n")

Create Dummy Variables for the Social Relations Model

Description

Generates the actor and partner dummy variables required to fit the Social Relations Model (SRM) using multilevel modeling, following the approach of Snijders and Kenny (1999). The function produces N actor dummies and N partner dummies, where N is the maximum group size in the dataset.

Usage

create_dummies(
  group_id,
  act_id,
  part_id,
  data,
  include_self = FALSE,
  merge_original = FALSE
)

Arguments

Value

a data.frame containing:

pdSRM_act_id

integer internal actor identifier

pdSRM_part_id

integer internal partner identifier

pdSRM_act_num

integer actor slot number within group

pdSRM_part_num

integer partner slot number within group

pdSRM_dyad_id

integer undirected dyad identifier

a1, a2, ...

actor dummy variables

p1, p2, ...

partner dummy variables

If merge_original = TRUE, all original variables are appended.

References

Snijders, T. A. B., & Kenny, D. A. (1999). The social relations model for family data: A multilevel approach. Personal Relationships, 6, 471–486. doi:10.1111/j.1475-6811.1999.tb00204.x

Examples

d_out <- create_dummies(
  group_id = "groupId",
  act_id   = "actId",
  part_id  = "partId",
  data     = sampleDyadData
)
head(d_out)

Construct pdSRM Object

Description

Internal method called by pdConstruct to initialize a pdSRM object. Enforces the SRM constraint of equal actor variances, equal partner variances, and a single actor-partner covariance.

Usage

## S3 method for class 'pdSRM'
pdConstruct(
  object,
  value = numeric(0),
  form = stats::formula(object),
  nam = nlme::Names(object),
  data = sys.frame(sys.parent()),
  ...
)

Arguments

Value

a pdSRM object representing the SRM positive-definite matrix

Examples

d <- createDummies(
  group.id = "groupId", act.id = "actId", part.id = "partId",
  d = sampleDyadData[sampleDyadData$timeId == 1, ],
  merge.original = TRUE
)
o <- nlme::lme(
  liking ~ 1,
  random = list(groupId = nlme::pdBlocked(list(
    nlme::pdIdent(~1),
    pdSRM(~ -1 + a1 + a2 + a3 + a4 + p1 + p2 + p3 + p4)
  ))),
  correlation = nlme::corCompSymm(form = ~1 | groupId / pdSRM_dyad_id),
  data = d,
  na.action = stats::na.omit
)

Extract Matrix or Square-Root Factor from a pdSRM Object

Description

Internal method called by pdMatrix to reconstruct the full covariance matrix from the three stored SRM parameters (actor SD, partner SD, actor-partner correlation).

Usage

## S3 method for class 'pdSRM'
pdMatrix(object, factor = FALSE)

Arguments

Value

if factor is FALSE, the positive-definite matrix represented by object; if TRUE, an upper triangular Cholesky factor with a logDet attribute

Examples

d <- createDummies(
  group.id = "groupId", act.id = "actId", part.id = "partId",
  d = sampleDyadData[sampleDyadData$timeId == 1, ],
  merge.original = TRUE
)
o <- nlme::lme(
  liking ~ 1,
  random = list(groupId = nlme::pdBlocked(list(
    nlme::pdIdent(~1),
    pdSRM(~ -1 + a1 + a2 + a3 + a4 + p1 + p2 + p3 + p4)
  ))),
  correlation = nlme::corCompSymm(form = ~1 | groupId / pdSRM_dyad_id),
  data = d,
  na.action = stats::na.omit
)

Construct the pdSRM object

Description

Creates the positive-definite matrix class used to specify the actor-partner covariance structure of the Social Relations Model within lme. This class enforces the SRM constraint that all actors share a single variance, all partners share a single variance, and a single actor-partner covariance (generalized reciprocity) is estimated.

Usage

pdSRM(
  value = numeric(0),
  form = NULL,
  nam = NULL,
  data = sys.frame(sys.parent())
)

Arguments

Value

a pdMat object representing a positive-definite matrix conforming to the SRM covariance structure

References

Knight, A. P., & Humphrey, S. E. (2019). Dyadic data analysis. In S. E. Humphrey & J. M. LeBreton (Eds.), The Handbook for Multilevel Theory, Measurement, and Analysis (pp. 423–447). American Psychological Association. doi:10.1037/0000115-019

Snijders, T. A. B., & Kenny, D. A. (1999). The social relations model for family data: A multilevel approach. Personal Relationships, 6, 471–486. doi:10.1111/j.1475-6811.1999.tb00204.x

Examples

d <- createDummies(
  group.id = "groupId", act.id = "actId", part.id = "partId",
  d = sampleDyadData[sampleDyadData$timeId == 1, ],
  merge.original = TRUE
)
o <- nlme::lme(
  liking ~ 1,
  random = list(groupId = nlme::pdBlocked(list(
    nlme::pdIdent(~1),
    pdSRM(~ -1 + a1 + a2 + a3 + a4 + p1 + p2 + p3 + p4)
  ))),
  correlation = nlme::corCompSymm(form = ~1 | groupId / pdSRM_dyad_id),
  data = d,
  na.action = stats::na.omit
)

Print a cpsrm Object

Description

Print a cpsrm Object

Usage

## S3 method for class 'cpsrm'
print(x, digits = 4, ...)

Arguments

Value

x, invisibly

Sample Round Robin Dataset

Description

A simulated directed dyadic dataset in long format, suitable for demonstrating the functions in this package. The dataset contains observations from a round-robin design in which each person in a group rated every other person. Ratings were collected at two time points.

Usage

sampleDyadData

Format

A data frame with 1548 rows and 16 variables:

groupId

integer; group identification variable

actId

integer; actor (rater) identification variable

partId

integer; partner (ratee) identification variable

timeId

integer; time point identifier (1 or 2)

contact

integer; directed dyad-level measure of contact frequency

liking

integer; directed dyad-level measure of liking

actEx

numeric; actor-level measure of extraversion

actAg

numeric; actor-level measure of agreeableness

actMale

integer; actor-level binary gender indicator (1 = male)

actAge

integer; actor-level age in years

partEx

numeric; partner-level measure of extraversion

partAg

numeric; partner-level measure of agreeableness

partMale

integer; partner-level binary gender indicator (1 = male)

partAge

integer; partner-level age in years

groupCohesion

numeric; group-level measure of cohesion

groupEfficacy

numeric; group-level measure of efficacy

Source

Simulated data created for package demonstration purposes.

Deprecated: use srm_pseudo_rsq instead

Description

srmPseudoRSq has been renamed to srm_pseudo_rsq. The argument names have also changed: null.model -> null_model, predict.model -> predict_model.

Usage

srmPseudoRSq(null.model, predict.model)

Arguments

Value

see srm_pseudo_rsq

Deprecated: use srm_run instead

Description

srmRun has been renamed to srm_run. The argument names have also changed: groupId -> group_id, actId -> act_id, partId -> part_id, feVars -> fe_vars.

Usage

srmRun(dv, groupId, actId, partId, feVars = NULL, data)

Arguments

Value

see srm_run

Deprecated: use srm_var_pct instead

Description

srmVarPct has been renamed to srm_var_pct.

Usage

srmVarPct(object)

Arguments

Value

see srm_var_pct

Calculate Pseudo R-Squared Values for the Social Relations Model

Description

Computes pseudo R-squared values for each SRM variance component by comparing a null model (intercept only) with a predictor model (with fixed effects). The pseudo R-squared for each component is (null - predicted) / null, reflecting the proportion of each variance component explained by the fixed effects.

Usage

srm_pseudo_rsq(null_model, predict_model)

Arguments

Value

a data.frame with three columns and four rows (Group, Actor, Partner, Dyad):

null

variance component from the null model

predict

variance component from the predictor model

pseudoR2

pseudo R-squared: (null - predict) / null

Examples

d <- create_dummies(
  group_id = "groupId", act_id = "actId", part_id = "partId",
  data = sampleDyadData[sampleDyadData$timeId == 1, ],
  merge_original = TRUE
)
null_mod <- nlme::lme(
  liking ~ 1,
  random = list(groupId = nlme::pdBlocked(list(
    nlme::pdIdent(~1),
    pdSRM(~ -1 + a1 + a2 + a3 + a4 + p1 + p2 + p3 + p4)
  ))),
  correlation = nlme::corCompSymm(form = ~1 | groupId / pdSRM_dyad_id),
  data = d,
  na.action = stats::na.omit
)
pred_mod <- nlme::lme(
  liking ~ actEx + partEx + contact,
  random = list(groupId = nlme::pdBlocked(list(
    nlme::pdIdent(~1),
    pdSRM(~ -1 + a1 + a2 + a3 + a4 + p1 + p2 + p3 + p4)
  ))),
  correlation = nlme::corCompSymm(form = ~1 | groupId / pdSRM_dyad_id),
  data = d,
  na.action = stats::na.omit
)
srm_pseudo_rsq(null_model = null_mod, predict_model = pred_mod)

Run the Social Relations Model Using Multilevel Modeling

Description

A wrapper function that fits the Social Relations Model (SRM) on a directed dyadic dataset using restricted maximum likelihood via lme. The function creates the necessary actor and partner dummy variables, constructs the SRM covariance structure using pdSRM, and returns both the raw lme output and a formatted variance decomposition table.

Usage

srm_run(dv, group_id, act_id, part_id, fe_vars = NULL, data)

Arguments

Value

a named list with two elements:

lme.output

the full lme model object

srm.output

a data.frame from srm_var_pct giving variances, percentages, and reciprocity correlations

References

Knight, A. P., & Humphrey, S. E. (2019). Dyadic data analysis. In S. E. Humphrey & J. M. LeBreton (Eds.), The Handbook for Multilevel Theory, Measurement, and Analysis (pp. 423–447). American Psychological Association. doi:10.1037/0000115-019

Snijders, T. A. B., & Kenny, D. A. (1999). The social relations model for family data: A multilevel approach. Personal Relationships, 6, 471–486. doi:10.1111/j.1475-6811.1999.tb00204.x

Examples

o <- srm_run(
  dv       = "liking",
  group_id = "groupId",
  act_id   = "actId",
  part_id  = "partId",
  fe_vars  = c("actEx", "partEx", "contact"),
  data     = sampleDyadData[sampleDyadData$timeId == 1, ]
)
o$srm.output

Extract Variance Decomposition from a Fitted SRM

Description

Extracts and formats the variance components and reciprocity correlations from an lme object fitted with the pdSRM covariance structure. Returns group, actor, partner, and dyadic (relationship) variances as both raw values and percentages of total variance, along with generalized reciprocity (actor-partner correlation) and dyadic reciprocity.

Usage

srm_var_pct(object)

Arguments

Value

a data.frame with two columns and six rows:

variances.and.covariances

Group, Actor, Partner, and Dyad variances; Generalized Reciprocity covariance; Dyadic Reciprocity covariance

percents.and.correlations

variance percentages for the four components; Generalized Reciprocity correlation; Dyadic Reciprocity correlation

References

Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic Data Analysis. Guilford Press.

Examples

d <- create_dummies(
  group_id = "groupId", act_id = "actId", part_id = "partId",
  data = sampleDyadData[sampleDyadData$timeId == 1, ],
  merge_original = TRUE
)
o <- nlme::lme(
  liking ~ 1,
  random = list(groupId = nlme::pdBlocked(list(
    nlme::pdIdent(~1),
    pdSRM(~ -1 + a1 + a2 + a3 + a4 + p1 + p2 + p3 + p4)
  ))),
  correlation = nlme::corCompSymm(form = ~1 | groupId / pdSRM_dyad_id),
  data = d,
  na.action = stats::na.omit
)
srm_var_pct(o)

Summarize a cpsrm Object

Description

Summarize a cpsrm Object

Usage

## S3 method for class 'cpsrm'
summary(object, ...)

Arguments

Value

object, invisibly

Summarize a pdSRM Object

Description

Internal method that produces a summary.pdMat representation of a pdSRM object, used by lme when printing model output.

Usage

## S3 method for class 'pdSRM'
summary(object, structName = "Social Relations Model", ...)

Arguments

Value

an object of class summary.pdMat with additional attributes structName and noCorrelation

Examples

d <- createDummies(
  group.id = "groupId", act.id = "actId", part.id = "partId",
  d = sampleDyadData[sampleDyadData$timeId == 1, ],
  merge.original = TRUE
)
o <- nlme::lme(
  liking ~ 1,
  random = list(groupId = nlme::pdBlocked(list(
    nlme::pdIdent(~1),
    pdSRM(~ -1 + a1 + a2 + a3 + a4 + p1 + p2 + p3 + p4)
  ))),
  correlation = nlme::corCompSymm(form = ~1 | groupId / pdSRM_dyad_id),
  data = d,
  na.action = stats::na.omit
)