Abstract
In this technical paper, we estimate instrumental limits of an unexacting comparison of results reported in different Web log studies. We consider sensitivity of results of log analysis to 4 controllable factors: a log sampling technique, an observation period and two cut-off variables peculiar to the Web log analysis (a LAN cut-off to exclude local area networks clients and a temporal cut-off to detect temporal search sessions). It is shown that 3 of these factors may lead to multiple differences in the case of marginal values of the factors whereas an effect of usual factors combinations is limited by 30%. These limits overcover differences between the results of Excite and AltaVista studies.
| Original language | English |
|---|---|
| Title of host publication | Workshop on Logging Traces of Web Activity: The Mechanics of Data Collection |
| Publication status | Published - 2006 |
| Externally published | Yes |