“No sale” items in auctions: do they really matter?

Bruce L. Alford, Otis W. Gilley, Charles M. Wood, Obinna Obilo

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

Research interest in auctions has grown steadily in nearly every business discipline because of the distinctive constructs and processes that the data analyses yield. Auction data present researchers with a particular challenge because of the presence of no sale items. Researchers have varied in their accommodation of no sale items and the analysis technique, which may lead to biases in the analyses. The first question is how should no sale items be treated? To address this, the type of data generated by auctions must be examined, since various methods handle no sale items differently. Additional questions are: What analysis technique is preferred for this type of data? What is the effect of using other analysis techniques? The results suggest that including no sale items improves estimation accuracy. Also, we found that using ordinary least squares with censored data provides results that are much better than using truncated data.

Original languageEnglish
Pages (from-to)155-168
Number of pages14
JournalMarketing Letters
Volume28
Issue number1
DOIs
StatePublished - Mar 1 2017

Keywords

  • Bid
  • Internet auction
  • No sale
  • Poisson
  • Regression
  • Tobit

Fingerprint

Dive into the research topics of '“No sale” items in auctions: do they really matter?'. Together they form a unique fingerprint.

Cite this