While sampling error was considered part of margin of error by the largest proportion of respondents (69%), significant numbers also thought "margin of error" implied question wording errors (66%), errors in developing a representative base or weighting errors (45%), mistakes by interviewers (45%) and question placement errors (40%).
These results can be interpreted as evidence that decreased lexical activation with nonword primes increased the probability of assigning the dominant stress, and of making stress errors on words with the less frequent stress pattern.
Two circumstances can be treated as errors of manual word segmentation; firstly, if a ridiculous segment appears in segmentation results, it can be treated as an error (type I error); and secondly, if an indivisible segment is divided in segmentation results, it can also be treated as an error (type II error).
"I've noticed that some blind people depend on their iPhones a lot," says Chris Parsons, a technology instructor at CCB. Parsons often warns her students not to become too dependent on smartphones, since the apps can at times be unreliable, and because some students will use voice to text messaging and not bother to correct the inevitable errors of the words.
Percentage of errors to words and percentage of dominant stress responses to nonwords were measured.
The performance of the G2P conversion system was expressed in two average error rates (over the fivefold): average error rate of phonemes (PER) and average error rate of words (WER).
It could be that applying stress sub-lexically generically increases the probability of errors for words.
The analysis of stress errors to word targets (i.e., non-dominant stress words pronounced with the dominant stress, i.e. "regularisation" errors) showed that, in agreement with the predictions, these errors were significantly more likely after nonword primes (13.4%) than after word primes (3.6%), [t1 (23) = 5.02, p<.001; t2 (82) = 2.44, p<.05].
Experimental results show that, in spite of high error rates, a "bag-of-words" representation of TV contents can be used as description, for the purposes of information retrieval and navigation.
This was done by using inter-rater reliability i.e. students' writing papers were collected by three different raters who were asked not to count errors of capitalization and word choice (e.g. buy vs. purchase) unless they impeded meaning.
