We’re interested in “researching” whether using Google Translate works better or worse for one language versus another when using round-trip translations. To do this, we will use headlines from articles published on the New Zealand Herald website. The explanatory variable will be a categorical variable language_code, which has two levels (the language codes for the two languages you have selected to compare for your investigation in Q1). In Q3, you will be able to use an app to obtain a random sample of these headlines. For each headline, a translation_similarity_score will be determined by calculating the percentage of words that are the same after a round-trip translation (RTT) with your selected languages. Punctuation is also taken into account when calculating this score. This numeric variable will be our response variable and the units are “points”. Write a suitable research question that: starts with “For headlines from all articles published on the New Zealand Herald website, how does the mean translation similarity score for …. “ is open (two-sided) requires sample-to-population inference to answer can be investigated using the variables and sample data described above.

Would it be appropriate for someone who is not a Māori speaker and has read the Google Translate translation only to claim that they have read the text? Would your analysis change if the capabilities of Google Translate improved (as likely to happen in the future)?

Which evaluation metric is commonly used to assess the performance of machine translation systems?Question 11Answera.Accuracyb.BLEU (Bilingual Evaluation Understudy)c.Precisiond.F1 Score

Translating material from a reliable foreign language source is good research.Group of answer choicesNoOnly from Māori.Only if you do the translation yourself, rather than e.g., google translateYes

Topic: Discuss one of the following hypotheses:Translators should always work into their L1The faster the translator works, the worse the translationPost-editing is always faster and better than fully human translationTranslations done with translation memories are of better quality than translations done from scratch  You must: 1) include examples from translation between Chinese and English; 2) include references to at least three published articles or books (not lectures or videos), ideally presenting arguments both for and against the hypothesis.You may 1) present the results from the work you or your colleagues have done in class, or 2) present the results of a small experiment you have organized yourself.On the basis of the empirical evidence and the weight of the argument, you should present your own conclusion at the end. You can either 1) report on an empirical study or 2) write a position paper. Report on an empirical studyIf you are reporting on an experiment, then the essay should be organized as follows:IntroductionLiterature review (at least three references)MethodologyResultsDiscussion and conclusion ReferencesHere is an example: sample essay 1 ActionsPosition paperIf you are presenting a range of arguments and considerations, then the structure might be as follows:Introduction (present and explain the hypothesis)Arguments and results against the hypothesesArguments and results for the hypothesis ConclusionReferences

How accurate is translation software?Group of answer choicesb. up to 80% or 90%a. usually 100%d. very inaccuratec. up to 50%