TY - CHAP
T1 - Interaction between participants
AU - Leppink, Jimmie
PY - 2019
Y1 - 2019
N2 - For the sake of introduction of all the concepts thus far, the example experiments in previous chapters are about individual participation of participants and a few other conditions (e.g., no twins, family members or natural groups otherwise) that likely do not invalidate the assumption of individual participants counting as independent observations. However, as in natural settings, increasing numbers of experiments have to deal with some kind of interaction between participants prior or during the experiment (e.g., twins, family members, students from the same learning groups, employees or patients from the same centre). In such cases, independence of observations is an unrealistic assumption. Whether a dependence results from natural ties or from an interaction component explicitly incorporated in an experiment, we need to account for that dependence in our statistical models. Three different types of situations are discussed in this chapter: dyads (e.g., couples), small-size groups (e.g., project teams), and larger groups or social networks. In each of these types of situations, individuals are treated as actors nested within higher-level actors or units (e.g., pairs, teams, centres or cliques). Also discussed in this chapter is how failing to account for this kind of data structure (dependence) can result in substantial distortions of our perspective on a treatment effect of interest.
AB - For the sake of introduction of all the concepts thus far, the example experiments in previous chapters are about individual participation of participants and a few other conditions (e.g., no twins, family members or natural groups otherwise) that likely do not invalidate the assumption of individual participants counting as independent observations. However, as in natural settings, increasing numbers of experiments have to deal with some kind of interaction between participants prior or during the experiment (e.g., twins, family members, students from the same learning groups, employees or patients from the same centre). In such cases, independence of observations is an unrealistic assumption. Whether a dependence results from natural ties or from an interaction component explicitly incorporated in an experiment, we need to account for that dependence in our statistical models. Three different types of situations are discussed in this chapter: dyads (e.g., couples), small-size groups (e.g., project teams), and larger groups or social networks. In each of these types of situations, individuals are treated as actors nested within higher-level actors or units (e.g., pairs, teams, centres or cliques). Also discussed in this chapter is how failing to account for this kind of data structure (dependence) can result in substantial distortions of our perspective on a treatment effect of interest.
U2 - 10.1007/978-3-030-21241-4_13
DO - 10.1007/978-3-030-21241-4_13
M3 - Chapter
SN - 978-3-030-21240-7
T3 - Springer Texts in Education
SP - 199
EP - 212
BT - Statistical methods for experimental research in education and psychology
PB - Springer
CY - Switzerland
ER -