3 ms·
Even if the tech was the same, couldn't they just run the less accurate test several times to increase the confidence?
by shawnz 8y ago
Even if the tech was the same, couldn't they just run the less accurate test several times to increase the confidence?
- hexane360 8y agoThat will reduce the random error, yes. But there's still the concern of systematic error, which has to be dealt with separately. If you're taking a biased sample to begin with, it doesn't matter how many times you repeat the test; the bias will still be there.
- shawnz 8y agoBut wouldn't systematic error affect everyone's results in the same way, meaning it couldn't result in any particular pair of people becoming more or less likely to be matched as the same person?