Pair programming is a good example of the positive impacts that pairing can have on speed, efficacy and reliability of the outcome. Pair programming however, has been implemented and assessed only in company and classroom settings. It has never before been implemented for an online course.The contribution of this paper therefore is to provide an initial framework for implementing Pair programming called Peer Assisted Learning(PAL) in an online course such as a Massively Open Online Course(MOOC). A proof-of-concept has also been provided by implementing the framework on an online Software Engineering course offered on the Open Collaborative Learning(OCL) platform developed in-house in our research laboratory. A pairing algorithm has been described that works solely on the time at which the course participants are available online. Provided an optimal pairing exists, the pairing algorithm has been categorically proven to converge to that optimal pairing.The framework allows for addition of various other factors to generate an optimal pairing. The algorithm takes as input numeric values associated with every factor along with the rule for pairing to produce an optimal pairing as the output provided it exists. Examples of such factors may include prior experience of the course participants and probability that they will complete the course.Future directions in Peer Assisted Learning(PAL) may include evaluation of the effects of such an implementation, inclusion of more factors for generating an optimal pair and extending the same concept to the formation of optimal groups.
Sreecharan Sankaranarayanan, National Institute of Technology Karnataka, India