| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
Rebecca says: > Actually, even the so-called "random" failures can introduce skew. Definitely, especially if we create a model and call "random" anything the model doesn't cover. The consent decree in Common Cause v. Jones didn't depend on the Votomatic errors having any systematic bias. They could. But even if they didn't, their error rate would have a systematic civil rights impact and a systematic effect on the total vote count. Dave.