Okay, (the collective) we designed a schema, implemented it, and started collecting the data. Initial looks into it suggest the schema is valid -- in other words, doesn't introduce a sampling bias. If that is the case, we need to analyze the session lengths that we're logging. Perform standard survival analysis on the checkins in the TestSearchUserSatisfaction2 schema.
- We probably want to dig in more what the old and new EL schemas do and did.
- "and then later we would use statistical methodology for dealing with right-censored data." I don't know what this means, and I work here ;p.
- "bug (introduced on September 10th, 2015) in the Event Logging (EL) system" is all link. Shorter link?
- "whether were differences" - "whether there were differences"
- There's kind of a lot of TLAs in this. Could we define terms and then either use the full names or the abbreviations but not both?
Everything else looks good!