Abstract:
Analytical Hierarchical Process (AHP) relies on expert responses to capture a decision maker's point of view on
different application domains, which is fundamental to the credibility and quality of decisions made. Do we know what
experts/respondents take into consideration when they give responses? How does their knowledge influence the responses
they give? It seems obvious that opinions should be sought from experts in diverse fields in a Geographical Information
Systems (GIS) based oil pipeline routing using AHP. Is this an assumption that should be made? Do experts in the same area
of expertise make decisions based on their professional knowledge or do they make subjective judgment irrespective of their
profession? Their decisions, whether rational or subjective, will have an input on the final proposed pipeline route. This study
compared the weights of 13 variables to be considered in pipeline routing derived using AHP in a GIS based pipeline routing
process from the responses of six groups comprising of civil engineers, environmentalists, county administrators, local
residents of the study area, oil industry experts, and geoinformation experts. Comparison of the responses was done among
experts of the same group and between groups using the Spearman's rank correlation coefficient. Intraclass Correlation
Coefficients (ICCs) were used to measure the reliability or consistency of rating the different variables by the experts. Visual
comparison of the responses was done using scatter plots and bar graphs. Out of the six groups of experts, geoinformation
specialists gave closely related responses and county administrators gave moderate related responses amongst themselves,
while the rest gave relatively varied responses. It was shown that most individuals made subjective decisions, due to the large
variation of responses within groups of similar profession. There was little correlation within the groups of oil experts,
environmentalists and local residents. The consistency of rating of different variables by these groups was also low. In these
groups, there was lower reliability level if we were to seek responses from only one expert in each of the groups. From the
analysis between groups, consistency of rating of the variables by the different groups was high, but the reliability if we were
to ask one group was low. Therefore, it was concluded that responses should be sought from different groups of experts,
having the expertise required in pipeline routing, with each group having several respondents. However, experts should
respond based on their professional knowledge. Else, the need to seek responses from different experts in a group, and from
different professional groups loses its meaning.
