**** ETNI on the web http://www.etni.org.il http://www.etni.org **** This was intended to be sent to the list - but I missed the ReplyAll button, so it turned into an offlist exchange. I'm posting it to teh list so as to share it with the rest of its subscribers. Lev ----------- 1. Ellen Schur wrote: If you delete every nth word from a passage, that means that some of your deletions will be function words, i.e., determiners, prepositions, etc. Others will be content words with a fixed form (nouns, adjectives, and adverbs); still others will be verbs which would then need to be conjugated and/or put in the correct tense. On a limited item test, this variation raises the question of what you are testing: grammar? Spelling? collocational knowledge? the student's ability to retrieve and produce a particular word in its correct form? And we could all point to additional linguistic elements as well that get included in the cloze soup. ---snip--- 2. I responded: That's exactly the point. The way cloze is used today, it is not "every nth word" removed - it's mostly content words (well, determiners, prepositions and conjunctions sometimes, too!). And if used properly, cloze can test knowledge in an intergrative manner - grammar, syntax, spelling, collocational knowledge, the student's ability to retrieve and produce a particular word in its correct form, as well as some additional linguistic elements. In diagnostic testing this sort of instrument would probably be of lesser importance, as we'd want to know exactly what the person taking the test can achieve in every one of these fields separately. But in a summative test like Bagrut, cloze is probably the method of choice. No-one says cloze should replace other methods. But to remove the cloze from the Bagrut was irresponsible, to say the least. ----------------------------------- 3. Ellen wrote back to me: Yes, I know that it isn't every nth word that's deleted. The problem is the mostly and the sometimes. Unless you give me a test, i.e., with enough items per category to be reliable, that says x percentage of the items are of this type and y percentage of another, how do I know what is being tested? On the other hand if you do a selected deletion with only content words of a particular frequency band, then I can make some meaningful statement of a student's semantic knowledge at a certain level. But before I gave them this type of test, I would want them to work with clozes with a bank of content words so that (1) they are familiar with the format and (2) I know that they have reviewed the vocabulary in context. --------------------------------- 4. I responded: You know this. I know this. A few others do. The question is, what do those who write actual Bagrut tests know? I have a feeling these tests are written by people who could use some courses in theory and practice of language testing. No doubt properly qualified specialists are available; the question is what to do with the tenured Ministry employees? No-one can write an ideal test. But trying to get as close to that point as possible is commendable. The Ministry should simply try harder... ------------------------- PS: I also attached a cartoon which will not pass the ETNI listserv; if you want to view it click here: http://schiffman.biz/gallery/view_album.php?set_albumName=Try-Harder Lev --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.726 / Virus Database: 481 - Release Date: 22/07/2004 ##### To send a message to the ETNI list email: etni@xxxxxxxxxxxxx ##### ##### Send queries and questions to: ask@xxxxxxxx #####