TY - CHAP
T1 - Models for treatment order effects
AU - Leppink, Jimmie
PY - 2019
Y1 - 2019
N2 - In the experiments discussed thus far, treatment is a between-subjects factor: participants in one condition do not also participate in another condition. However, in some cases, different groups may receive different treatments at different occasions, the order of treatment varies across groups, and there is a measurement of an outcome variable of interest at each occasion. In the simplest setup, there are two treatments, A and B, which are taken in a different order by each of two groups: A-B in one group (with a measurement after A followed by a measurement after B), B-A in the other group (with a measurement after B followed by a measurement after A). In other cases, there are more treatments and more orders for a larger number of groups with it, or there are only a few—and perhaps only two—treatments which can vary in each of a larger number of trials. At each trial, there is a measurement of an outcome variable of interest. Consider students who are asked to read ten articles, each article is read in each of two possible formats determined in a random order, and after each article students are asked to rate on a VAS how much effort it took to read the article. These are all examples of situations where treatment varies both between and within participants and a measurement of an outcome variable of interest takes place in each trial (e.g., after each condition or for each article). As in Chaps. 14 and 15, the occasions or trials can still be viewed as stations to be passed by each of the participants, but there is something that varies from station to station that has to be accounted for in our models. This final chapter provides examples for how to do that.
AB - In the experiments discussed thus far, treatment is a between-subjects factor: participants in one condition do not also participate in another condition. However, in some cases, different groups may receive different treatments at different occasions, the order of treatment varies across groups, and there is a measurement of an outcome variable of interest at each occasion. In the simplest setup, there are two treatments, A and B, which are taken in a different order by each of two groups: A-B in one group (with a measurement after A followed by a measurement after B), B-A in the other group (with a measurement after B followed by a measurement after A). In other cases, there are more treatments and more orders for a larger number of groups with it, or there are only a few—and perhaps only two—treatments which can vary in each of a larger number of trials. At each trial, there is a measurement of an outcome variable of interest. Consider students who are asked to read ten articles, each article is read in each of two possible formats determined in a random order, and after each article students are asked to rate on a VAS how much effort it took to read the article. These are all examples of situations where treatment varies both between and within participants and a measurement of an outcome variable of interest takes place in each trial (e.g., after each condition or for each article). As in Chaps. 14 and 15, the occasions or trials can still be viewed as stations to be passed by each of the participants, but there is something that varies from station to station that has to be accounted for in our models. This final chapter provides examples for how to do that.
U2 - /10.1007/978-3-030-21241-4_16
DO - /10.1007/978-3-030-21241-4_16
M3 - Chapter
SN - 978-3-030-21240-7
T3 - Springer Texts in Education
SP - 243
EP - 254
BT - Statistical methods for experimental research in education and psychology
PB - Springer
CY - Switzerland
ER -