O n f e r e n c e

participants in this condition) that did not seem to change their work flow and 
collaboration intensity during the whole intervention. Elen (2013), who has investigated 
this kind of unexpected behaviour in online learning, coined the term "instructional 
disobedience" for it. The difference in the two conditions concerning adherence to 
instruc
tional support is corroborated by the fact that the students in the “model
-
first” 
condition rated the usefulness of the support instruments higher in the post-hoc 
questionnaires. Whether this can be attributed to the sequence of the scaffolding 
instruments, remains to be explored. It might also be related to a difference in learning 
style, as the groups for the conditions were convenient samples from two different 
institutions of higher education (a University College and a University, respectively). 
Another important question that could not be answered satisfyingly by the collected data 
set is: To what extent a “learning
-by-
doing” effect has played a role in the development 
of the collaborative process, in addition to the instructional support? In order to answer 
this question, a control setting with at least 2 unsupported sessions would be required, 
which, in turn, seems daunting in terms of motivation. 
Finally, a pending dilemma regarding statistical testing of the results should be 
mentioned: As the data have a high degree of interdependency (observations over time 
nested within students nested within triads), multilevel analysis would be required. On 
the other hand, the sample size at the highest level, i.e., the number of triads in each 
condition (11 resp. 14), is too small to carry out this type of analysis in a reliable way 
(Hox, 2010). 
To conclude, this intervention study evidenced a strong effect of (scaffolded) online 
collaboration on content elaboration which in turn is a key factor to successful synthesis 
writing. Some interesting details were revealed regarding instructional support for a 
complex online task in advanced foreign language writing. Benefits of 
both 
scripting and 
observational learning could be identified in the “learning to write and to collaborate” –
process.