I used an online survey in combination with data from the delicious system to investigate my pilot hypothesis (see instrumentation at http://www.erickaakcire.net/delicious/pilotarchive.htm). A survey was chosen because the independent variable, feedback, cannot be fully measured through content alone.

Feedback was operationalized in three ways. First there was the feedback from browsing the delicious web site itself. Pilot study questions 3-5 and 7-15 measured the frequency of each browsing feedback occurrence. Web syndication was another form of feedback that was measured by questions 16-21. The use of other feedback from tools developed by third parties was intended to be measured by pilot study question 23, however, a technical error prevented proper data collection. Although user collectivity was not the dependent variable in the pilot study, it was intended to be measured as a control variable by survey question 29. The level of system use was measured by collecting data from the delicious system. I collected the survey respondent's total number of links, total number of tags, date of first link, and date of 100th link to measure total use and frequency of use.

I recruited participants from the delicious-discuss mailing list. The discussion on this list focuses around design, technical, social, and intellectual issues related to the delicious system. The developer of delicious contributes by responding to user issues and asking for feedback on ideas. The list has over 800 of the most engaged delicious users. I sent one announcement and one follow-up clarification to the list and received 70 valid responses in four days.

The participants were overwhelmingly male, 92.5%. This extremely skewed distribution may result from the small sample size and the gender distribution of those with high technology interest. Also, the age distribution was very young, with 71% under 30, and occupations in the information technology industry and education or research predominated. Although the system-wide average length of time to accumulate 100 links is not known, the sample median of 36.5 days seems to indicate heavy recent usage.

In other descriptive aspects the participants were quite diverse. Of the 65 who indicated their world region 57% were from North America, 28% from Western Europe and the UK, and the remainder were from Central and Eastern Europe, the Pacific, South Asia, and the Middle East. South and Central America, Asia, South East Asia, and Africa were not represented. The date of the sample's first link, and by implication, the start date, was also diverse, with no clear predominance of a starting time period.