3. PROPUESTAS DE MEJORA
3.3. SUSTITUCIÓN DE LUMINARIAS DE BAJO RENDIMIENTO
The purpose of this study is to provide insights into the effects of information revelation on the process and outcomes of auctions. The Imaras system is served as a test bed to manipulate different rules of information revelation and the proposed hypotheses are then tested with an experimental approach. The experimental design, tasks and procedure for data collection are described in the following sections.
5.4.1 Experimental design
The study was carried out in two experiments: A controlled laboratory experiment and an online experiment. The former helps establish the causality between independent variables and dependent variables and ensures the internal validity of the study (Colquitt and Gainesville 2008), while the latter provides a more natural environment in order to increase the external validity of the study and strengthen the findings (Kerlinger and Lee 2000).
Table 5-1 shows the experimental design with the treatments on information rules in the two experiments.
Table 5-1. Experimental design and treatments (Study 2)
Information revelation Experiment 1
(Laboratory) Experiment 2 (Online) Rules Types Information
Rule B Type B ABS only Treatment B
Rule BDwin Type B + Dwin ABS + winning bids Treatment BDwin Treatment BDwin Rule BDall Type B + Dall ABS + all bidders’ bids Treatment BDall
Experiment 1 was conducted in the laboratory environment with a number of sessions. Each session lasted one hour and 40 minutes, including the time for preparation and questionnaires. The auction in each session lasted 50 minutes; it could end earlier if no bidders submitted bids or no more better bids available for the buyer. It was multi-round auction; each round took five minutes. The bidders obtained the auction updates on the rounds and the revealed information immediately from their computer screen.
Experiment 2 was conducted in an online setting, in which the auctions were running on the Internet and lasted 10 days. Each round was set as one day. The bidders could log on and off at any time during the auction. They were informed by emails about any auction updates, but they were required to log on to access the revealed information.
5.4.2 Experimental tasks
The experimental tasks were role-play with a business case that involved contracting between a milk producer and several transportation service providers. Three attributes of the transportation service
were concerned: (1) standard rate; (2) rush rate; and (3) penalty for delay. There were a number of option values for each attribute, and the possible ranges were known to each participant.
The bidders played the role of a sales manager for the service providers. They were competing with each other to win the contract. Each contract could be awarded to only one supplier. The service providers wanted to get the contract; however, the contract should bring their business profits rather than losses. Therefore, the bidders were given break-even points of the company they represented.
The context and background provided the public information that was known to every supplier, while the preferences were explained in the private information that was not known to the other suppliers. The reservation and aspiration levels of each company were also stated in their private information. A financial calculator is implemented in the system, and it could be used by the participants to calculate the revenue and profit for each contract alternative. Appendix D summarizes the business profiles and preferences of the companies involved in the transaction.
In both experiments, the utility function, reservation and aspiration levels, and the break-even point were kept unchanged for the buyer and the suppliers. The only difference was the number of alternatives due to the number of option values for each attribute. In Experiment 1, each attribute consisted of six option values, resulting in 216 alternatives in total; in Experiment 2, each attribute had 15 possible values; there were 3,375 alternatives.
5.4.3 Experimental procedure
The experiments involve several steps and a series of activities (Table E-1 in Appendix E). Before the experiment, the participants first signed up online and their demographical information was gathered via a registration form. The participants were then randomly matched up with different roles and assigned to one of the treatments.
During the experiment, the participants first read the case for the preparation, including the public information and the private information. This was followed by a case quiz and pre-
questionnaire before the auction started. Table F-1 in Appendix F shows a sample case quiz for one supplier, which was used to improve the participants’ understanding of the tasks. The participants were then asked to provide their perceptions of the transaction task as well as their aspiration and reservation levels. These measures were used to examine their understanding of the task, their expected bidding space (the alternatives they may bid), and the contract to achieve.
During the auction process, the participants constructed and submitted bids on behalf of the companies they represented. Once the auction was closed, they were asked to fill out a post- questionnaire that collected subjective responses of the process and outcomes. The participants reported their evaluation of the auction process in terms of performance, effort, and experience.
The participants’ activities during the transaction process were recorded in a database, which was used to analyze the transaction process and outcomes.
In the experiments, there was no change in the experimental procedure, except for the participants in Experiment 2 being provided with a video demonstration of the system before the participants actually used it. This was introduced following the feedback from the participants in Experiment 1. In addition, because Experiment 2 was conducted online, the participants could be neither given instructions similar to those in Experiment 1, nor explanations during the process. The purpose of introducing the video demonstration was to improve the participants’ knowledge and understanding of the system and the bidding process, which was similar to the business training for new technologies in the real world and also ensured that the participants had the same baseline experience in order to test the effects due to treated conditions (Kumar and Benbasat 2006; Kamis, Koufaris et al. 2008).