Skip to contents

Calculate SRM summary statistics

Source: R/srm_stats.R
srm_stats.Rd

Computes descriptive statistics and variance/covariance components from a round-robin network using the Social Relations Model framework. Requires a square (unipartite) adjacency matrix; for bipartite networks, use srm() instead.

Usage

srm_stats(
  mat,
  type = c("rowmeans", "colmeans", "totalmeans", "actor_var", "unique_var",
    "partner_var", "relationship_cov", "actor_partner_cov"),
  time = NULL
)

Arguments

mat

A square numeric matrix (or list of square matrices for longitudinal data). Diagonal values are set to zero internally. NAs are replaced with zero. Must not be an srm object (statistics are already in summary(fit)$variance_table).

type

One of:

  • "rowmeans": row means (average out-ties per actor).

  • "colmeans": column means (average in-ties per actor).

  • "totalmeans": grand mean of off-diagonal entries.

  • "actor_var": actor (sender) variance component.

  • "partner_var": partner (receiver) variance component.

  • "unique_var": unique dyadic variance component.

  • "relationship_cov": generalized reciprocity covariance.

  • "actor_partner_cov": covariance between sending and receiving.

Variance and covariance components require n >= 4 actors. For n <= 3, NA is returned.

time

Optional character vector of time-point names to subset when mat is a named list.

Value

A named numeric vector (for descriptive statistics) or scalar (for variance components). For a list input, a list of such values. See srm() for the unified interface.

Examples

# simple inline example
test_matrix = matrix(c(0, 1, 2, 2, 0, 1, 1, 2, 0), nrow = 3, ncol = 3)
rownames(test_matrix) = colnames(test_matrix) = c("A", "B", "C")
actor_totalmeans = srm_stats(test_matrix, type = "totalmeans")

data(classroom)
srm_stats(classroom, type = "rowmeans")
#>      Cady     Aaron  Gretchen     Karen     Janis    Damian     Kevin      Glen 
#> 4.4172727 3.6690909 3.4300000 4.5300000 3.1218182 5.8963636 3.7109091 2.9563636 
#>     Shane     Trang    Regina   Norbury 
#> 2.6563636 3.0945455 0.7109091 3.8990909 
srm_stats(classroom, type = "actor_var")
#> [1] 1.400046
srm_stats(classroom, type = "actor_partner_cov")
#> [1] -0.2524702

# longitudinal
data(trade_net)
srm_stats(trade_net, type = "colmeans", time = c("2015", "2019"))
#> $`2015`
#>       USA       CHN       DEU       JPN       GBR       FRA       KOR       IND 
#> 2.2188889 0.8955556 2.3422222 2.6511111 3.0111111 1.7422222 1.9833333 2.4411111 
#>       BRA       CAN 
#> 1.6144444 3.6844444 
#> 
#> $`2019`
#>      USA      CHN      DEU      JPN      GBR      FRA      KOR      IND 
#> 2.585556 2.523333 2.006667 2.728889 1.484444 3.736667 1.736667 2.053333 
#>      BRA      CAN 
#> 2.988889 3.095556 
#>