Development of an instrument for data collection in a multidimensional scaling study of personal Web usage in the South African workplace
Thesis (MBA)--Stellenbosch University, 2011.
In a relatively very short period the Internet has grown from being virtually unknown to becoming an essential business tool. Together with its many benefits, the Internet has unfortunately brought with it several new organisational challenges. One of these challenges is how to manage personal Web usage (PWU) in the workplace effectively. Although many managers see PWU as a form of workplace deviance, many researchers have pointed out its potential benefits such as learning, time-saving, employee well-being and a source of ideas. To help organisations manage PWU in the workplace more effectively, this research realised the need for a typology of PWU behaviours in the South African workplace. Multidimensional scaling (MDS) was identified as an objective method of creating such a typology. The objective of this research was therefore to develop an instrument to gather data for a multidimensional scaling study of PWU behaviours in the South African workplace. A questionnaire was designed that consists of three distinct sections. The first section contains seven pre-coded demographics questions that correspond with specific demographic variables, proven to have a relationship with PWU. The second section of the questionnaire is designed to gather dissimilarity data for input into an MDS algorithm. To begin with, 25 Web usage behaviours of South Africans were identified using Google Ad Planner. After weighing up various options of comparing the Web usage behaviours, the pairwise comparison method was selected. Ross sequencing was used to reduce positioning and timing effects. To reduce the number of judgements per participant, the 300 required judgments are split six ways, resulting in 50 judgements per participant. The last section of the questionnaire is designed to gather data to assist with interpreting the dimensions of the MDS configuration. Eight benefits and risks of PWU were identified. These are combined into a matrix together with the 25 Web usage behaviours. The data from this section will allow future research to use linear regression to discover the relationship between the Web usage behaviours (the objects), and the benefits and risks of PWU (the variables). It is believed that this design offers a fair compromise between the time and effort required of participants and the quality and integrity of the acquired data.