Bhalerao, Aneesh Madhao (2011-08). Statistical Analysis of the Fourth Case Study in Reverse Auction Research. Master's Thesis.
Participating in an auction and winning items by placing bids has been in practice since at least 500 B.C. Auctions have evolved since then and anyone can now participate in one online and buy items ranging from clothes, electronics, automobiles and homes using online auction websites, such as eBay. A Reverse Auction varies from the traditional style of Auction where items or services are won by placing successive higher bids until the auction ends. The study of Reverse Auction was first introduced to Texas A & M University in 2004 and continues today, using a SQL based web system.
This current research provides a detailed statistical analysis of the fourth case study in this long running work. This fourth case study involved the participation of five bidders who had no prior experience in Reverse Auctions. A Microsoft Access database system and ASP web based user interface was developed and used to conduct these initial studies. However, due to the limited capability of the Access system to handle more than a limited number of connections or bidders, a Microsoft SQL database and web system was developed in 2006 and has been used in all subsequent studies. Case studies have involved up to ten participants. The results from the fourth case study show that a Reverse Auction can result in an increase in the average cost of the job to the owner. Also, there is evidence of game play amongst the bidders and against the purchaser that causes their profits to rise as they gain proficiency in the game. This behavior has been termed as 'tacit collision', but it is considered a byproduct of the system and not illegal behavior. This study analyzes the fourth study data to investigate if the behavior termed "tacit collusion" is evident in the bidding data. This analysis is completed by performing a detailed statistical analysis of the bidding data.
Analysis of the profit percentages illustrates the different stages of the game play amongst the bidders. This game play behavior is illustrated by plotting average number of bids to the profit made by each bidder. The data clearly suggests that the players became efficient in their bidding strategy, although some bidders are more efficient than others. This observation negates the common conception that Reverse Auctions will result in lowering average costs for the owners.
The individual data of bidders for bids and profit reveal why some players were able to obtain higher than average results and why the others were not. This study can be taken further by analyzing the patterns of the successful and unsuccessful players to determine what causes them to gain or lose profits.