R/data_ison.R
ison_classmates.RdFriendship among 26 pupils aged 11 to 13 in a Dutch school class, measured at four time points over one school year, collected by Andrea Knecht for her dissertation. The data are widely used to illustrate the joint dynamics of network selection and behavioural influence.
Two layers of tie are recorded:
friends: friendship nominations, at each of the four waves
primary: whether the pupils knew each other from primary school. This is a time-invariant baseline relation and recorded at wave 1 only.
Four fixed nodal attributes are included:
sex: female or male
age: age at wave 1, in years
ethnicity: Dutch or non-Dutch
religion: religion of the father, one of Christian, none, other
Two further attributes change over the waves, and are held as nodal changes for waves 2 to 4:
delinquency: self-reported delinquent behaviour, measured at all four waves
alcohol: self-reported alcohol use,
not measured at wave 1 and so NA there
data(ison_classmates)#> -- # Knecht's Classmates -------------------------------------------------------
#> # A longitudinal, labelled, multiplex, directed network of 26 pupils and 460
#> friendship arcs and 86 primary arcs over 4 waves
#>
#> -- Nodes
#> # A tibble: 26 x 7
#> label sex age ethnicity religion delinquency alcohol
#> <chr> <chr> <int> <chr> <chr> <int> <dbl>
#> 1 a01 female 12 Dutch none 2 NA
#> 2 a02 female 12 Dutch none 1 NA
#> 3 a03 female 12 non-Dutch other 2 NA
#> 4 a04 male 12 Dutch none 2 NA
#> # i 22 more rows
#>
#> -- Changes
#> # A tibble: 150 x 4
#> time node var value
#> <int> <int> <chr> <list>
#> 1 2 1 alcohol 1<int>
#> 2 2 1 delinquency 2<int>
#> 3 2 2 na TRUE<lgl>
#> 4 2 3 alcohol 3<int>
#> # i 146 more rows
#>
#> -- Ties
#> # A tibble: 546 x 4
#> from to layer time
#> <int> <int> <chr> <int>
#> 1 3 1 friends 1
#> 2 9 1 friends 1
#> 3 12 1 friends 1
#> 4 3 2 friends 1
#> # i 542 more rows
#>Missing attribute values, coded -99 elsewhere, are recorded as NA here.
The source data also distinguish nominations that are missing (9) from
those that are structurally missing (10), where the pupil was not in the
class at that wave. The two are recorded differently here.
A missing nomination says that a pupil did not answer. One pupil did not
answer at wave 2 and two did not at wave 3, which is logged as a change of
their na status, and a change back at the wave they answer again.
Every nomination they would have given is then missing: 25 at wave 2 and 48
at wave 3, which as_missinglist() returns.
The nominations they received are not missing, since the other pupils
still answered. Recording them this way, rather than as the absence of a
tie, keeps a nomination nobody gave from being read as one somebody
withheld. impute_ties() treats them as absent ties, with
rule = "zero", or imputes them, with one of its other rules.
A structurally missing nomination says only that one pupil was not in the
class, from wave 3 onwards. That pupil leaving is logged as a change of
their active status at wave 3, which as_siena() renders as a
composition change, and no ties are recorded for them at either wave.
Nothing they might have been nominated for is missing at those waves
either, since there was nothing there to miss. This is why the two pupils
who did not answer at wave 3 have 24 missing nominations each and not 25.
Nobody was asked to nominate themselves, so the diagonal of each matrix is dropped. This drops a self-tie in each of the wave 2 friendship and primary-school matrices, and a self-nomination coded as missing in the wave 2 and wave 3 friendship matrices.
Knecht, Andrea. 2008. Friendship Selection and Friends' Influence. Dynamics of Networks and Actor Attributes in Early Adolescence. PhD dissertation, University of Utrecht.
Knecht, Andrea, Tom A. B. Snijders, Chris Baerveldt, Christian E. G. Steglich, and Werner Raub. 2010. "Friendship and Delinquency: Selection and Influence Processes in Early Adolescence". Social Development 19(3): 494-514. doi:10.1111/j.1467-9507.2009.00564.x