Loading ...
Sorry, an error occurred while loading the content.

vocabulary acquisition from ER: what is a realistic number?

Expand Messages
  • Glenski
    As the spring semester is about a month away from starting, I am beginning to polish my opening lectures for a reading skills class. I want to provide students
    Message 1 of 9 , Mar 4 5:29 PM
    • 0 Attachment
      As the spring semester is about a month away from starting, I am beginning to polish my opening lectures for a reading skills class. I want to provide students with some information about what ER can do for them, and one of those items concerns how much vocabulary they might expect to pick up. Naturally, this is a real rat's nest of a topic!

      I just read in The Backseat Linguist (http://www.tprstories.com/images/ijflt/IJFLTMarch2012.pdf) an article tearing down a 2003 paper by M Hill and B Laufer where they claimed to have calculated that readers need to read 8 million words in order to acquire 2,000 words. TBL explains quite clearly (to me) how the logic in Hill and Laufer's paper is flawed. Moreover, TBL states that although the acquisition rate for words met only once in a book is low, it isn't trivial, and then cites Pellicer-Sanchez & Schmitt (2010, Reading in a Foreign Language) to show that "the meaning of 29% of the words that appeared only once in the text they used to measure incidental acquisition were recognized...in a post-test". TBL also then cites Waring and Takaki (2003, Reading in a Foreign Language) who showed 16% retention after 3 months. I have only skimmed the Waring & Takaki article, but it has a wealth of information in it pointing out flaws in methodology which is meant to measure this incidental acquisition.

      Waring & Takaki nicely list some issues common in research on vocabulary gains. A couple that stood out to me were the following:

      1) They said that it is "quite rare...to ask how long these gains will last". I am surprised. Isn't long-term retention something we should be striving for, and therefore trying to measure?

      2) The type of test one chooses to measure post-reading gain is critical, and multiple-choice tests are "notoriously difficult to construct reliably". My question is, what is a GOOD test? Waring and Takaki simply state that researchers should use more than one.

      3) Frequency of the words in a text is not always taken into account. Heck, it makes just intrinsic sense that if a word is encountered more times it will be remembered better, so why can't researchers keep this in mind? Waring and Takaki state that other than one report (Horst, et. al, 1998), "few studies have looked at what types of words are learned in the reading" [meaning that how often words occurred in the text is not taken into account].

      One thing that neither TBL nor Waring & Takaki mentioned as far as I could see was student study/practice on vocabulary (but I only skimmed W&T's paper so far). Is there any data to show fairly long-term retention based on whether students studied the words later versus did not study (purely incidental learning)? I'm sure this opens a huge can of worms, too, about what type and frequency of practice/studying generates what level of retention, but I'd just like to know where to start looking. Right now, all I can say to my students is that they should make lists of new words and study them, and then show them the forgetting curve.

      How does everyone on this list start their semester with regard to explaining to students that ER alone will not greatly improve vocabulary? I'm curious because that seems to be the only real benefit some students want in any English class, aside from learning to speak better. They ALL want more vocabulary.

      Glenski
    • Rob Waring
      Hi all, Sorry I don t have much time to go into this but I d broadly agree with Hill and Laufer. Another way to tackle this is to get a big corpus of learner
      Message 2 of 9 , Mar 4 5:57 PM
      • 0 Attachment
        Hi all,

        Sorry I don't have much time to go into this but I'd broadly agree with Hill and Laufer. 

        Another way to tackle this is to get a big corpus of learner materials and set acquisition rates at say 10, 20, 30 meetings and see how many words one needs to meet in order to learn say the first 1000, 2000 etc. at those re-exposure rates. My own research shows this (excerpted from a recent unpublished article of mine)

        "the entire text (225,000 running words including audio and workbook activities) of a typical five-level course book series Sequences (Thomlinson, Waring and Woodall, 2009) was entered into a computer for analysis. It is a fairly standard 4 skills thematic-based course with readings, listening, speaking activities, and so on. The text was analyzed to determine which words were likely to be known at the end of the two and a half year course.

         

        There is considerable research (e.g. Pellicer-Sánchez and Schmitt, 2010; Nation, 2001, Waring and Takaki, 2003) to show that that it takes 20-30 meetings with a word to learn it well. Therefore, words met over 20 times can be considered ‘learnt’ at least receptively. The knowledge of a word that has been met between 5-19 times is probably partial or incomplete and slow to access. If the word is met fewer than five times over the two and a half years it is likely to be forgotten.

         

        Description: Macintosh HD:Users:Rob:Desktop:Screen Shot 2012-02-20 at 6.16.06 PM.png

        Table 1: The number of words likely to be known from a course book and graded readers

        Table 1 shows that only around 962 words of the 3,275 words in the five books, meet the ‘known’ criteria of 20 meetings with many of these being function words. Another 1052 would be partially known, while 1261 are likely to be forgotten. If however, we add one or two graded readers (about 1-2 hours) per week at their level of difficulty, then the learners’ vocabulary would more than double to 2,119 words ‘learnt’ and a further 1,571 words partially known, probably because their input quadrupled to over a million words which allows them to pass the threshold of ‘learning’. However, these figures do not include increases in reading speed and fluency, nor heightened awareness of one’s ‘sense’ of language, collocation and colligation, text structure and many aspects of discourse that stem from reading at one’s ability level. In addition, Nishizawa, Yoshioka and Fukuda in a four-year study of the impact of ER in reluctant learners found considerable gains in overall language ability. Dozens of other studies were found showing similar results in the meta-analysis mentioned above. This is clear evidence that graded readers can, and do, build and consolidate vocabulary knowledge."

        What these data UNDERrepresents learning gains because it does not show the learning of collocations, colligations, 'sense ' of language etc. I.e it under represents more learning.  However it also OVERrepresents learning because the computer analysis was done on word families not each form ( i.e. it assumes that meeting any member of the 'use' family means the whole family of 15 -20 items has been learnt.)

        Rob


        On Mar 5, 2012, at 10:29 AM, Glenski wrote:

         

        As the spring semester is about a month away from starting, I am beginning to polish my opening lectures for a reading skills class. I want to provide students with some information about what ER can do for them, and one of those items concerns how much vocabulary they might expect to pick up. Naturally, this is a real rat's nest of a topic!

        I just read in The Backseat Linguist (http://www.tprstories.com/images/ijflt/IJFLTMarch2012.pdf) an article tearing down a 2003 paper by M Hill and B Laufer where they claimed to have calculated that readers need to read 8 million words in order to acquire 2,000 words. TBL explains quite clearly (to me) how the logic in Hill and Laufer's paper is flawed. Moreover, TBL states that although the acquisition rate for words met only once in a book is low, it isn't trivial, and then cites Pellicer-Sanchez & Schmitt (2010, Reading in a Foreign Language) to show that "the meaning of 29% of the words that appeared only once in the text they used to measure incidental acquisition were recognized...in a post-test". TBL also then cites Waring and Takaki (2003, Reading in a Foreign Language) who showed 16% retention after 3 months. I have only skimmed the Waring & Takaki article, but it has a wealth of information in it pointing out flaws in methodology which is meant to measure this incidental acquisition.

        Waring & Takaki nicely list some issues common in research on vocabulary gains. A couple that stood out to me were the following:

        1) They said that it is "quite rare...to ask how long these gains will last". I am surprised. Isn't long-term retention something we should be striving for, and therefore trying to measure?

        2) The type of test one chooses to measure post-reading gain is critical, and multiple-choice tests are "notoriously difficult to construct reliably". My question is, what is a GOOD test? Waring and Takaki simply state that researchers should use more than one.

        3) Frequency of the words in a text is not always taken into account. Heck, it makes just intrinsic sense that if a word is encountered more times it will be remembered better, so why can't researchers keep this in mind? Waring and Takaki state that other than one report (Horst, et. al, 1998), "few studies have looked at what types of words are learned in the reading" [meaning that how often words occurred in the text is not taken into account].

        One thing that neither TBL nor Waring & Takaki mentioned as far as I could see was student study/practice on vocabulary (but I only skimmed W&T's paper so far). Is there any data to show fairly long-term retention based on whether students studied the words later versus did not study (purely incidental learning)? I'm sure this opens a huge can of worms, too, about what type and frequency of practice/studying generates what level of retention, but I'd just like to know where to start looking. Right now, all I can say to my students is that they should make lists of new words and study them, and then show them the forgetting curve.

        How does everyone on this list start their semester with regard to explaining to students that ER alone will not greatly improve vocabulary? I'm curious because that seems to be the only real benefit some students want in any English class, aside from learning to speak better. They ALL want more vocabulary.

        Glenski


      • mrwwalmsley
        Just quick thought... One of the big benefits of ER is enhancing existing vocabulary knowledge. Since ideally 95-98% of the words encountered during ER are
        Message 3 of 9 , Mar 5 2:59 AM
        • 0 Attachment
          Just quick thought...

          One of the big benefits of ER is enhancing existing vocabulary knowledge. Since ideally 95-98% of the words encountered during ER are supposed to be "known" to the learner, learners are probably not going to learn a whole lot of new words WELL, since most of the 2-5% unknown words are likely to be low frequency words.

          However, as Rob pointed out, ER is great for enhancing collocation knowledge etc. of the 95-98% of the known words in the text.

          --- In ExtensiveReading@yahoogroups.com, Rob Waring <waring_robert@...> wrote:
          >
          > Hi all,
          >
          > Sorry I don't have much time to go into this but I'd broadly agree with Hill and Laufer.
          >
          > Another way to tackle this is to get a big corpus of learner materials and set acquisition rates at say 10, 20, 30 meetings and see how many words one needs to meet in order to learn say the first 1000, 2000 etc. at those re-exposure rates. My own research shows this (excerpted from a recent unpublished article of mine)
          >
          > "the entire text (225,000 running words including audio and workbook activities) of a typical five-level course book series Sequences (Thomlinson, Waring and Woodall, 2009) was entered into a computer for analysis. It is a fairly standard 4 skills thematic-based course with readings, listening, speaking activities, and so on. The text was analyzed to determine which words were likely to be known at the end of the two and a half year course.
          >
          > There is considerable research (e.g. Pellicer-Sánchez and Schmitt, 2010; Nation, 2001, Waring and Takaki, 2003) to show that that it takes 20-30 meetings with a word to learn it well. Therefore, words met over 20 times can be considered `learnt' at least receptively. The knowledge of a word that has been met between 5-19 times is probably partial or incomplete and slow to access. If the word is met fewer than five times over the two and a half years it is likely to be forgotten.
          >
          >
          > Table 1: The number of words likely to be known from a course book and graded readers
          > Table 1 shows that only around 962 words of the 3,275 words in the five books, meet the `known' criteria of 20 meetings with many of these being function words. Another 1052 would be partially known, while 1261 are likely to be forgotten. If however, we add one or two graded readers (about 1-2 hours) per week at their level of difficulty, then the learners' vocabulary would more than double to 2,119 words `learnt' and a further 1,571 words partially known, probably because their input quadrupled to over a million words which allows them to pass the threshold of `learning'. However, these figures do not include increases in reading speed and fluency, nor heightened awareness of one's `sense' of language, collocation and colligation, text structure and many aspects of discourse that stem from reading at one's ability level. In addition, Nishizawa, Yoshioka and Fukuda in a four-year study of the impact of ER in reluctant learners found considerable gains in overall language ability. Dozens of other studies were found showing similar results in the meta-analysis mentioned above. This is clear evidence that graded readers can, and do, build and consolidate vocabulary knowledge."
          >
          > What these data UNDERrepresents learning gains because it does not show the learning of collocations, colligations, 'sense ' of language etc. I.e it under represents more learning. However it also OVERrepresents learning because the computer analysis was done on word families not each form ( i.e. it assumes that meeting any member of the 'use' family means the whole family of 15 -20 items has been learnt.)
          >
          > Rob
          >
          >
          > On Mar 5, 2012, at 10:29 AM, Glenski wrote:
          >
          > > As the spring semester is about a month away from starting, I am beginning to polish my opening lectures for a reading skills class. I want to provide students with some information about what ER can do for them, and one of those items concerns how much vocabulary they might expect to pick up. Naturally, this is a real rat's nest of a topic!
          > >
          > > I just read in The Backseat Linguist (http://www.tprstories.com/images/ijflt/IJFLTMarch2012.pdf) an article tearing down a 2003 paper by M Hill and B Laufer where they claimed to have calculated that readers need to read 8 million words in order to acquire 2,000 words. TBL explains quite clearly (to me) how the logic in Hill and Laufer's paper is flawed. Moreover, TBL states that although the acquisition rate for words met only once in a book is low, it isn't trivial, and then cites Pellicer-Sanchez & Schmitt (2010, Reading in a Foreign Language) to show that "the meaning of 29% of the words that appeared only once in the text they used to measure incidental acquisition were recognized...in a post-test". TBL also then cites Waring and Takaki (2003, Reading in a Foreign Language) who showed 16% retention after 3 months. I have only skimmed the Waring & Takaki article, but it has a wealth of information in it pointing out flaws in methodology which is meant to measure this incidental acquisition.
          > >
          > > Waring & Takaki nicely list some issues common in research on vocabulary gains. A couple that stood out to me were the following:
          > >
          > > 1) They said that it is "quite rare...to ask how long these gains will last". I am surprised. Isn't long-term retention something we should be striving for, and therefore trying to measure?
          > >
          > > 2) The type of test one chooses to measure post-reading gain is critical, and multiple-choice tests are "notoriously difficult to construct reliably". My question is, what is a GOOD test? Waring and Takaki simply state that researchers should use more than one.
          > >
          > > 3) Frequency of the words in a text is not always taken into account. Heck, it makes just intrinsic sense that if a word is encountered more times it will be remembered better, so why can't researchers keep this in mind? Waring and Takaki state that other than one report (Horst, et. al, 1998), "few studies have looked at what types of words are learned in the reading" [meaning that how often words occurred in the text is not taken into account].
          > >
          > > One thing that neither TBL nor Waring & Takaki mentioned as far as I could see was student study/practice on vocabulary (but I only skimmed W&T's paper so far). Is there any data to show fairly long-term retention based on whether students studied the words later versus did not study (purely incidental learning)? I'm sure this opens a huge can of worms, too, about what type and frequency of practice/studying generates what level of retention, but I'd just like to know where to start looking. Right now, all I can say to my students is that they should make lists of new words and study them, and then show them the forgetting curve.
          > >
          > > How does everyone on this list start their semester with regard to explaining to students that ER alone will not greatly improve vocabulary? I'm curious because that seems to be the only real benefit some students want in any English class, aside from learning to speak better. They ALL want more vocabulary.
          > >
          > > Glenski
          > >
          > >
          >
        • Rob Waring
          Hi Yes that s probably so but we have very little experimental evidence of this. It s mostly based on (strong) convictions and common sense. Still, it d be
          Message 4 of 9 , Mar 5 3:10 AM
          • 0 Attachment
            Hi

            Yes that's probably so but we have very little experimental evidence of this. It's mostly based on (strong) convictions and common sense. Still, it'd be great to find out what rate collocations and other (non form-meaning)  aspects of vocab knowledge are picked up.

            PhD anyone?

            Rob
            On Mar 5, 2012, at 7:59 PM, mrwwalmsley wrote:

             

            Just quick thought...

            One of the big benefits of ER is enhancing existing vocabulary knowledge. Since ideally 95-98% of the words encountered during ER are supposed to be "known" to the learner, learners are probably not going to learn a whole lot of new words WELL, since most of the 2-5% unknown words are likely to be low frequency words.

            However, as Rob pointed out, ER is great for enhancing collocation knowledge etc. of the 95-98% of the known words in the text.

            --- In ExtensiveReading@yahoogroups.com, Rob Waring <waring_robert@...> wrote:
            >
            > Hi all,
            >
            > Sorry I don't have much time to go into this but I'd broadly agree with Hill and Laufer.
            >
            > Another way to tackle this is to get a big corpus of learner materials and set acquisition rates at say 10, 20, 30 meetings and see how many words one needs to meet in order to learn say the first 1000, 2000 etc. at those re-exposure rates. My own research shows this (excerpted from a recent unpublished article of mine)
            >
            > "the entire text (225,000 running words including audio and workbook activities) of a typical five-level course book series Sequences (Thomlinson, Waring and Woodall, 2009) was entered into a computer for analysis. It is a fairly standard 4 skills thematic-based course with readings, listening, speaking activities, and so on. The text was analyzed to determine which words were likely to be known at the end of the two and a half year course.
            >
            > There is considerable research (e.g. Pellicer-Sánchez and Schmitt, 2010; Nation, 2001, Waring and Takaki, 2003) to show that that it takes 20-30 meetings with a word to learn it well. Therefore, words met over 20 times can be considered `learnt' at least receptively. The knowledge of a word that has been met between 5-19 times is probably partial or incomplete and slow to access. If the word is met fewer than five times over the two and a half years it is likely to be forgotten.
            >
            >
            > Table 1: The number of words likely to be known from a course book and graded readers
            > Table 1 shows that only around 962 words of the 3,275 words in the five books, meet the `known' criteria of 20 meetings with many of these being function words. Another 1052 would be partially known, while 1261 are likely to be forgotten. If however, we add one or two graded readers (about 1-2 hours) per week at their level of difficulty, then the learners' vocabulary would more than double to 2,119 words `learnt' and a further 1,571 words partially known, probably because their input quadrupled to over a million words which allows them to pass the threshold of `learning'. However, these figures do not include increases in reading speed and fluency, nor heightened awareness of one's `sense' of language, collocation and colligation, text structure and many aspects of discourse that stem from reading at one's ability level. In addition, Nishizawa, Yoshioka and Fukuda in a four-year study of the impact of ER in reluctant learners found considerable gains in overall language ability. Dozens of other studies were found showing similar results in the meta-analysis mentioned above. This is clear evidence that graded readers can, and do, build and consolidate vocabulary knowledge."
            >
            > What these data UNDERrepresents learning gains because it does not show the learning of collocations, colligations, 'sense ' of language etc. I.e it under represents more learning. However it also OVERrepresents learning because the computer analysis was done on word families not each form ( i.e. it assumes that meeting any member of the 'use' family means the whole family of 15 -20 items has been learnt.)
            >
            > Rob
            >
            >
            > On Mar 5, 2012, at 10:29 AM, Glenski wrote:
            >
            > > As the spring semester is about a month away from starting, I am beginning to polish my opening lectures for a reading skills class. I want to provide students with some information about what ER can do for them, and one of those items concerns how much vocabulary they might expect to pick up. Naturally, this is a real rat's nest of a topic!
            > >
            > > I just read in The Backseat Linguist (http://www.tprstories.com/images/ijflt/IJFLTMarch2012.pdf) an article tearing down a 2003 paper by M Hill and B Laufer where they claimed to have calculated that readers need to read 8 million words in order to acquire 2,000 words. TBL explains quite clearly (to me) how the logic in Hill and Laufer's paper is flawed. Moreover, TBL states that although the acquisition rate for words met only once in a book is low, it isn't trivial, and then cites Pellicer-Sanchez & Schmitt (2010, Reading in a Foreign Language) to show that "the meaning of 29% of the words that appeared only once in the text they used to measure incidental acquisition were recognized...in a post-test". TBL also then cites Waring and Takaki (2003, Reading in a Foreign Language) who showed 16% retention after 3 months. I have only skimmed the Waring & Takaki article, but it has a wealth of information in it pointing out flaws in methodology which is meant to measure this incidental acquisition.
            > >
            > > Waring & Takaki nicely list some issues common in research on vocabulary gains. A couple that stood out to me were the following:
            > >
            > > 1) They said that it is "quite rare...to ask how long these gains will last". I am surprised. Isn't long-term retention something we should be striving for, and therefore trying to measure?
            > >
            > > 2) The type of test one chooses to measure post-reading gain is critical, and multiple-choice tests are "notoriously difficult to construct reliably". My question is, what is a GOOD test? Waring and Takaki simply state that researchers should use more than one.
            > >
            > > 3) Frequency of the words in a text is not always taken into account. Heck, it makes just intrinsic sense that if a word is encountered more times it will be remembered better, so why can't researchers keep this in mind? Waring and Takaki state that other than one report (Horst, et. al, 1998), "few studies have looked at what types of words are learned in the reading" [meaning that how often words occurred in the text is not taken into account].
            > >
            > > One thing that neither TBL nor Waring & Takaki mentioned as far as I could see was student study/practice on vocabulary (but I only skimmed W&T's paper so far). Is there any data to show fairly long-term retention based on whether students studied the words later versus did not study (purely incidental learning)? I'm sure this opens a huge can of worms, too, about what type and frequency of practice/studying generates what level of retention, but I'd just like to know where to start looking. Right now, all I can say to my students is that they should make lists of new words and study them, and then show them the forgetting curve.
            > >
            > > How does everyone on this list start their semester with regard to explaining to students that ER alone will not greatly improve vocabulary? I'm curious because that seems to be the only real benefit some students want in any English class, aside from learning to speak better. They ALL want more vocabulary.
            > >
            > > Glenski
            > >
            > >
            >


          • Glen Hill
            Rob, Thanks for that rather lengthy answer. I m not sure if it directly answers my question about the Hill and Laufer critique by The Backseat Linguist. I
            Message 5 of 9 , Mar 5 3:31 PM
            • 0 Attachment
              Rob,
              Thanks for that rather lengthy answer. I'm not sure if it directly answers my question about the Hill and Laufer critique by The Backseat Linguist. I would really appreciate if you could make a little time (no rush, though) to look over that TBL article and see how justified it is in finding the flaws in Hill and Laufer's paper (and the related one by Schmitt). Thanks.
              Glenski

              On Mon, Mar 5, 2012 at 8:10 PM, Rob Waring <waring_robert@...> wrote:
               

              Hi


              Yes that's probably so but we have very little experimental evidence of this. It's mostly based on (strong) convictions and common sense. Still, it'd be great to find out what rate collocations and other (non form-meaning)  aspects of vocab knowledge are picked up.

              PhD anyone?

              Rob

              On Mar 5, 2012, at 7:59 PM, mrwwalmsley wrote:

               

              Just quick thought...

              One of the big benefits of ER is enhancing existing vocabulary knowledge. Since ideally 95-98% of the words encountered during ER are supposed to be "known" to the learner, learners are probably not going to learn a whole lot of new words WELL, since most of the 2-5% unknown words are likely to be low frequency words.

              However, as Rob pointed out, ER is great for enhancing collocation knowledge etc. of the 95-98% of the known words in the text.

              --- In ExtensiveReading@yahoogroups.com, Rob Waring <waring_robert@...> wrote:
              >
              > Hi all,
              >
              > Sorry I don't have much time to go into this but I'd broadly agree with Hill and Laufer.
              >
              > Another way to tackle this is to get a big corpus of learner materials and set acquisition rates at say 10, 20, 30 meetings and see how many words one needs to meet in order to learn say the first 1000, 2000 etc. at those re-exposure rates. My own research shows this (excerpted from a recent unpublished article of mine)
              >
              > "the entire text (225,000 running words including audio and workbook activities) of a typical five-level course book series Sequences (Thomlinson, Waring and Woodall, 2009) was entered into a computer for analysis. It is a fairly standard 4 skills thematic-based course with readings, listening, speaking activities, and so on. The text was analyzed to determine which words were likely to be known at the end of the two and a half year course.
              >
              > There is considerable research (e.g. Pellicer-Sánchez and Schmitt, 2010; Nation, 2001, Waring and Takaki, 2003) to show that that it takes 20-30 meetings with a word to learn it well. Therefore, words met over 20 times can be considered `learnt' at least receptively. The knowledge of a word that has been met between 5-19 times is probably partial or incomplete and slow to access. If the word is met fewer than five times over the two and a half years it is likely to be forgotten.
              >
              >
              > Table 1: The number of words likely to be known from a course book and graded readers
              > Table 1 shows that only around 962 words of the 3,275 words in the five books, meet the `known' criteria of 20 meetings with many of these being function words. Another 1052 would be partially known, while 1261 are likely to be forgotten. If however, we add one or two graded readers (about 1-2 hours) per week at their level of difficulty, then the learners' vocabulary would more than double to 2,119 words `learnt' and a further 1,571 words partially known, probably because their input quadrupled to over a million words which allows them to pass the threshold of `learning'. However, these figures do not include increases in reading speed and fluency, nor heightened awareness of one's `sense' of language, collocation and colligation, text structure and many aspects of discourse that stem from reading at one's ability level. In addition, Nishizawa, Yoshioka and Fukuda in a four-year study of the impact of ER in reluctant learners found considerable gains in overall language ability. Dozens of other studies were found showing similar results in the meta-analysis mentioned above. This is clear evidence that graded readers can, and do, build and consolidate vocabulary knowledge."
              >
              > What these data UNDERrepresents learning gains because it does not show the learning of collocations, colligations, 'sense ' of language etc. I.e it under represents more learning. However it also OVERrepresents learning because the computer analysis was done on word families not each form ( i.e. it assumes that meeting any member of the 'use' family means the whole family of 15 -20 items has been learnt.)
              >
              > Rob
              >
              >
              > On Mar 5, 2012, at 10:29 AM, Glenski wrote:
              >
              > > As the spring semester is about a month away from starting, I am beginning to polish my opening lectures for a reading skills class. I want to provide students with some information about what ER can do for them, and one of those items concerns how much vocabulary they might expect to pick up. Naturally, this is a real rat's nest of a topic!
              > >
              > > I just read in The Backseat Linguist (http://www.tprstories.com/images/ijflt/IJFLTMarch2012.pdf) an article tearing down a 2003 paper by M Hill and B Laufer where they claimed to have calculated that readers need to read 8 million words in order to acquire 2,000 words. TBL explains quite clearly (to me) how the logic in Hill and Laufer's paper is flawed. Moreover, TBL states that although the acquisition rate for words met only once in a book is low, it isn't trivial, and then cites Pellicer-Sanchez & Schmitt (2010, Reading in a Foreign Language) to show that "the meaning of 29% of the words that appeared only once in the text they used to measure incidental acquisition were recognized...in a post-test". TBL also then cites Waring and Takaki (2003, Reading in a Foreign Language) who showed 16% retention after 3 months. I have only skimmed the Waring & Takaki article, but it has a wealth of information in it pointing out flaws in methodology which is meant to measure this incidental acquisition.
              > >
              > > Waring & Takaki nicely list some issues common in research on vocabulary gains. A couple that stood out to me were the following:
              > >
              > > 1) They said that it is "quite rare...to ask how long these gains will last". I am surprised. Isn't long-term retention something we should be striving for, and therefore trying to measure?
              > >
              > > 2) The type of test one chooses to measure post-reading gain is critical, and multiple-choice tests are "notoriously difficult to construct reliably". My question is, what is a GOOD test? Waring and Takaki simply state that researchers should use more than one.
              > >
              > > 3) Frequency of the words in a text is not always taken into account. Heck, it makes just intrinsic sense that if a word is encountered more times it will be remembered better, so why can't researchers keep this in mind? Waring and Takaki state that other than one report (Horst, et. al, 1998), "few studies have looked at what types of words are learned in the reading" [meaning that how often words occurred in the text is not taken into account].
              > >
              > > One thing that neither TBL nor Waring & Takaki mentioned as far as I could see was student study/practice on vocabulary (but I only skimmed W&T's paper so far). Is there any data to show fairly long-term retention based on whether students studied the words later versus did not study (purely incidental learning)? I'm sure this opens a huge can of worms, too, about what type and frequency of practice/studying generates what level of retention, but I'd just like to know where to start looking. Right now, all I can say to my students is that they should make lists of new words and study them, and then show them the forgetting curve.
              > >
              > > How does everyone on this list start their semester with regard to explaining to students that ER alone will not greatly improve vocabulary? I'm curious because that seems to be the only real benefit some students want in any English class, aside from learning to speak better. They ALL want more vocabulary.
              > >
              > > Glenski
              > >
              > >
              >



            • Karen Rowan
              The International Journal of Foreign Language Teaching has just been published. It is a free, on-line, peer-reviewed journal. There is an article in this
              Message 6 of 9 , Mar 5 3:36 PM
              • 0 Attachment
                The International Journal of Foreign Language Teaching has just been published.  It is a free, on-line, peer-reviewed journal.  There is an article in this edition on SSR and another by Dr. Krashen on the acquisition of academic vocabulary through reading.   You will also find previous editions on-line and there are a great many articles on Extensive Reading, including some by authors I have noticed are on this listserv. =)

                I hope you enjoy it.  Please consider submitting articles to the next edition of the Journal --- either research articles for review --- or articles about teaching strategies that you use that might be helpful to other teachers.  We have a Teacher to Teacher section that this time contains an article on Embedded Reading.

                I look forward to hearing what those on this listerv thing of the Journal.  We started in 2004.

                Karen Rowan
                Director, Fluency Fast Language Classes
                www.FluencyFast.com / karen@...
                Founded in 2004
                Karen Rowan Workshops, Inc
                www.tprstories.com / karen@...
                Founded in 1996
                Editor, The International Journal of Foreign Language Teaching
                www.IJFLT.com
                Founded in 2004
                1-866-WWW-FLUENCY / 719-633-6000

                Have you seen our new book, Isabela captura un congo?  Available now for $5.  
                www.fluencyfast.com

                MISSION: The mission of Fluency Fast is to create and sustain a movement  that causes a global shift in consciousness by transforming communications among  individuals, communities, and countries and inspiring people to use language as a tool to build bridges with other cultures.  Our goal is to dispel the myth that learning languages is difficult and to inspire people to have fun learning Arabic, French, German, Mandarin, Russian and Spanish, easily, inexpensively, effectively and in a brief period of time. 

                Email sent to this address is automatically added to our periodic mailing list announcing upcoming classes.  Please request that we delete your email after our correspondence if you do not wish to receive our bi-annual upcoming class emails.  Fluency Fast is an equal opportunity educator and employer.  We do not discriminate on the basis of race, color, gender, creed, sexual orientation, disability or age.



              • mrwwalmsley
                Here s a few of my thoughts on the subject... I think the TBL post makes some good points. TBL points out a flaw in Hill and Laufer s CALCULATION to arrive at
                Message 7 of 9 , Mar 5 8:47 PM
                • 0 Attachment
                  Here's a few of my thoughts on the subject...

                  I think the TBL post makes some good points.

                  TBL points out a flaw in Hill and Laufer's CALCULATION to arrive at the statement that 8,000,000 words reading is required to acquire 2,000 words. TBL justifies his claim with support from Schmitt, and Warring and Takaki. I agree with TBL that there are flaws in H&Ls statement.

                  So if H&L's statement is wrong... How many words do you REALLY have to read to acquire 2,000 words? The answer is complex and depends on a number of factors. This is the same reason that estimates (in the literature) of the number of ER encounters required to 'acquire' unknown words range from 6-12+.

                  Here are several factors that need to be considered.

                  1. What does it mean to acquire an unknown word? The studies use different meanings for 'acquire', i.e. their test instruments that measure vocabulary acquisition measure different levels of vocabulary gain.

                  2. What is the ability level of the learner? The literature states that ER texts should contain 1-5% unknown words. As the % of unknown words in a text increases, the learners comprehension, and gains in vocabulary knowledge are diminished.

                  As an extreme example, the original Clockwork Orange study (Saragi et al., 1978) showed large (can't remember exact % off top of my head) gains in 'Nadsat' vocabulary when English native speakers read Clockwork Orange. From what I remember the % of Nadsat words in the text was around 4%-within the recommended 1-5%.

                  Other studies replicated the Clockwork Orange study... except they had ESL students read the text (which is not the easiest text for NATIVE ENGLISH SPEAKERS). Not surprisingly, these students 'acquired' FAR LESS 'Nadsat' vocabulary than the English natives in the original study.

                  3. Some words are easy to learn, others are difficult. For example, words with many senses (such as many high frequency verbs & nouns) require many more encounters to learn. Another example, is cognates, which are generally easier to pick up than non-cognates. Since Spanish has a lot more cognates with English than Japanese, a Spanish speaker will pick up vocabulary through ER at a much higher rate than the average Japanese student.

                  4. Another factor investigated in several studies is context. Some context that learners encounter a word will assist their learning of the word, others will have a negative effect.

                  5. Another factor (mentioned by Glenski) is supplementary language learning activities. There are a number of studies that have looked at the effect of pre, post, and during reading vocabulary exercises on vocabulary gains through ER. The conclusions are often weak. Many of these studies struggle to control all the key factors, or to adequately measure small incremental gains in vocabulary knowledge.

                  --- In ExtensiveReading@yahoogroups.com, Glen Hill <glenahill@...> wrote:
                  >
                  > Rob,
                  > Thanks for that rather lengthy answer. I'm not sure if it directly answers
                  > my question about the Hill and Laufer critique by The Backseat Linguist. I
                  > would really appreciate if you could make a little time (no rush, though)
                  > to look over that TBL article and see how justified it is in finding the
                  > flaws in Hill and Laufer's paper (and the related one by Schmitt). Thanks.
                  > Glenski
                  >
                  > On Mon, Mar 5, 2012 at 8:10 PM, Rob Waring <waring_robert@...> wrote:
                  >
                  > > **
                  > >
                  > >
                  > > Hi
                  > >
                  > > Yes that's probably so but we have very little experimental evidence of
                  > > this. It's mostly based on (strong) convictions and common sense. Still,
                  > > it'd be great to find out what rate collocations and other (non
                  > > form-meaning) aspects of vocab knowledge are picked up.
                  > >
                  > > PhD anyone?
                  > >
                  > > Rob
                  > >
                  > > On Mar 5, 2012, at 7:59 PM, mrwwalmsley wrote:
                  > >
                  > >
                  > >
                  > > Just quick thought...
                  > >
                  > > One of the big benefits of ER is enhancing existing vocabulary knowledge.
                  > > Since ideally 95-98% of the words encountered during ER are supposed to be
                  > > "known" to the learner, learners are probably not going to learn a whole
                  > > lot of new words WELL, since most of the 2-5% unknown words are likely to
                  > > be low frequency words.
                  > >
                  > > However, as Rob pointed out, ER is great for enhancing collocation
                  > > knowledge etc. of the 95-98% of the known words in the text.
                  > >
                  > > --- In ExtensiveReading@yahoogroups.com, Rob Waring <waring_robert@>
                  > > wrote:
                  > > >
                  > > > Hi all,
                  > > >
                  > > > Sorry I don't have much time to go into this but I'd broadly agree with
                  > > Hill and Laufer.
                  > > >
                  > > > Another way to tackle this is to get a big corpus of learner materials
                  > > and set acquisition rates at say 10, 20, 30 meetings and see how many words
                  > > one needs to meet in order to learn say the first 1000, 2000 etc. at those
                  > > re-exposure rates. My own research shows this (excerpted from a recent
                  > > unpublished article of mine)
                  > > >
                  > > > "the entire text (225,000 running words including audio and workbook
                  > > activities) of a typical five-level course book series Sequences
                  > > (Thomlinson, Waring and Woodall, 2009) was entered into a computer for
                  > > analysis. It is a fairly standard 4 skills thematic-based course with
                  > > readings, listening, speaking activities, and so on. The text was analyzed
                  > > to determine which words were likely to be known at the end of the two and
                  > > a half year course.
                  > > >
                  > > > There is considerable research (e.g. Pellicer-Sánchez and Schmitt, 2010;
                  > > Nation, 2001, Waring and Takaki, 2003) to show that that it takes 20-30
                  > > meetings with a word to learn it well. Therefore, words met over 20 times
                  > > can be considered `learnt' at least receptively. The knowledge of a word
                  > > that has been met between 5-19 times is probably partial or incomplete and
                  > > slow to access. If the word is met fewer than five times over the two and a
                  > > half years it is likely to be forgotten.
                  > > >
                  > > >
                  > > > Table 1: The number of words likely to be known from a course book and
                  > > graded readers
                  > > > Table 1 shows that only around 962 words of the 3,275 words in the five
                  > > books, meet the `known' criteria of 20 meetings with many of these being
                  > > function words. Another 1052 would be partially known, while 1261 are
                  > > likely to be forgotten. If however, we add one or two graded readers (about
                  > > 1-2 hours) per week at their level of difficulty, then the learners'
                  > > vocabulary would more than double to 2,119 words `learnt' and a further
                  > > 1,571 words partially known, probably because their input quadrupled to
                  > > over a million words which allows them to pass the threshold of `learning'.
                  > > However, these figures do not include increases in reading speed and
                  > > fluency, nor heightened awareness of one's `sense' of language, collocation
                  > > and colligation, text structure and many aspects of discourse that stem
                  > > from reading at one's ability level. In addition, Nishizawa, Yoshioka and
                  > > Fukuda in a four-year study of the impact of ER in reluctant learners found
                  > > considerable gains in overall language ability. Dozens of other studies
                  > > were found showing similar results in the meta-analysis mentioned above.
                  > > This is clear evidence that graded readers can, and do, build and
                  > > consolidate vocabulary knowledge."
                  > > >
                  > > > What these data UNDERrepresents learning gains because it does not show
                  > > the learning of collocations, colligations, 'sense ' of language etc. I.e
                  > > it under represents more learning. However it also OVERrepresents learning
                  > > because the computer analysis was done on word families not each form (
                  > > i.e. it assumes that meeting any member of the 'use' family means the whole
                  > > family of 15 -20 items has been learnt.)
                  > > >
                  > > > Rob
                  > > >
                  > > >
                  > > > On Mar 5, 2012, at 10:29 AM, Glenski wrote:
                  > > >
                  > > > > As the spring semester is about a month away from starting, I am
                  > > beginning to polish my opening lectures for a reading skills class. I want
                  > > to provide students with some information about what ER can do for them,
                  > > and one of those items concerns how much vocabulary they might expect to
                  > > pick up. Naturally, this is a real rat's nest of a topic!
                  > > > >
                  > > > > I just read in The Backseat Linguist (
                  > > http://www.tprstories.com/images/ijflt/IJFLTMarch2012.pdf) an article
                  > > tearing down a 2003 paper by M Hill and B Laufer where they claimed to have
                  > > calculated that readers need to read 8 million words in order to acquire
                  > > 2,000 words. TBL explains quite clearly (to me) how the logic in Hill and
                  > > Laufer's paper is flawed. Moreover, TBL states that although the
                  > > acquisition rate for words met only once in a book is low, it isn't
                  > > trivial, and then cites Pellicer-Sanchez & Schmitt (2010, Reading in a
                  > > Foreign Language) to show that "the meaning of 29% of the words that
                  > > appeared only once in the text they used to measure incidental acquisition
                  > > were recognized...in a post-test". TBL also then cites Waring and Takaki
                  > > (2003, Reading in a Foreign Language) who showed 16% retention after 3
                  > > months. I have only skimmed the Waring & Takaki article, but it has a
                  > > wealth of information in it pointing out flaws in methodology which is
                  > > meant to measure this incidental acquisition.
                  > > > >
                  > > > > Waring & Takaki nicely list some issues common in research on
                  > > vocabulary gains. A couple that stood out to me were the following:
                  > > > >
                  > > > > 1) They said that it is "quite rare...to ask how long these gains will
                  > > last". I am surprised. Isn't long-term retention something we should be
                  > > striving for, and therefore trying to measure?
                  > > > >
                  > > > > 2) The type of test one chooses to measure post-reading gain is
                  > > critical, and multiple-choice tests are "notoriously difficult to construct
                  > > reliably". My question is, what is a GOOD test? Waring and Takaki simply
                  > > state that researchers should use more than one.
                  > > > >
                  > > > > 3) Frequency of the words in a text is not always taken into account.
                  > > Heck, it makes just intrinsic sense that if a word is encountered more
                  > > times it will be remembered better, so why can't researchers keep this in
                  > > mind? Waring and Takaki state that other than one report (Horst, et. al,
                  > > 1998), "few studies have looked at what types of words are learned in the
                  > > reading" [meaning that how often words occurred in the text is not taken
                  > > into account].
                  > > > >
                  > > > > One thing that neither TBL nor Waring & Takaki mentioned as far as I
                  > > could see was student study/practice on vocabulary (but I only skimmed
                  > > W&T's paper so far). Is there any data to show fairly long-term retention
                  > > based on whether students studied the words later versus did not study
                  > > (purely incidental learning)? I'm sure this opens a huge can of worms, too,
                  > > about what type and frequency of practice/studying generates what level of
                  > > retention, but I'd just like to know where to start looking. Right now, all
                  > > I can say to my students is that they should make lists of new words and
                  > > study them, and then show them the forgetting curve.
                  > > > >
                  > > > > How does everyone on this list start their semester with regard to
                  > > explaining to students that ER alone will not greatly improve vocabulary?
                  > > I'm curious because that seems to be the only real benefit some students
                  > > want in any English class, aside from learning to speak better. They ALL
                  > > want more vocabulary.
                  > > > >
                  > > > > Glenski
                  > > > >
                  > > > >
                  > > >
                  > >
                  > >
                  > >
                  > >
                  >
                • Keith Folse
                  I joined this list to hear about research on ER, with my attention directed toward vocabulary growth, so I ve been interested in this recent discussion,
                  Message 8 of 9 , Mar 6 2:13 PM
                  • 0 Attachment
                    I joined this list to hear about research on ER, with my attention directed toward vocabulary growth, so I've been interested in this recent discussion, especially when posters have mentioned specific research results on this issue.

                    I would suggest that interested people  read 4 chapters from an OUTSTANDING book called Researching and Analyzing Vocabulary by Nation and Webb (2011).  The 4 chapters are:

                    Part 2. "Incidental Vocabulary Learning"
                    Chapter 5  Guessing Word Meaning from Context Clues
                    Chapter 6  Incidental Vocabulary Learning from Reading
                    Chapter 7  Vocabulary Learning Through Output

                    and

                    Part 3. "Corpus-based Research"
                    Chapter 9.  Analyzing Vocabulary Load and Opportunities for Learning


                    Keith Folse
                    University of Central Florida


                  • Karen Rowan
                    Thanks! I ll check those out. I would also recommend Developing Academic Language: Some Hypotheses, Dr. Stephen Krashen, page 8, The International Journal of
                    Message 9 of 9 , Mar 6 2:54 PM
                    • 0 Attachment
                      Thanks!  I'll check those out.

                      I would also recommend

                      Developing Academic Language: 
                      Some Hypotheses, Dr. Stephen Krashen, page 8, The International Journal of Foreign Language Teaching

                      It's free.

                      Extensive Reading is a frequent topic in this journal.  Beniko Mason published on study on ER in the most recent edition, also available at the same website.

                      Karen Rowan




                      On Mar 6, 2012, at 3:13 PM, Keith Folse wrote:

                       

                      I joined this list to hear about research on ER, with my attention directed toward vocabulary growth, so I've been interested in this recent discussion, especially when posters have mentioned specific research results on this issue.

                      I would suggest that interested people  read 4 chapters from an OUTSTANDING book called Researching and Analyzing Vocabulary by Nation and Webb (2011).  The 4 chapters are:

                      Part 2. "Incidental Vocabulary Learning"
                      Chapter 5  Guessing Word Meaning from Context Clues
                      Chapter 6  Incidental Vocabulary Learning from Reading
                      Chapter 7  Vocabulary Learning Through Output

                      and

                      Part 3. "Corpus-based Research"
                      Chapter 9.  Analyzing Vocabulary Load and Opportunities for Learning


                      Keith Folse
                      University of Central Florida




                      Karen Rowan
                      Director, Fluency Fast Language Classes
                      www.FluencyFast.com / karen@...
                      Founded in 2004
                      Karen Rowan Workshops, Inc
                      www.tprstories.com / karen@...
                      Founded in 1996
                      Editor, The International Journal of Foreign Language Teaching
                      www.IJFLT.com
                      Founded in 2004
                      1-866-WWW-FLUENCY / 719-633-6000

                      Have you seen our new book, Isabela captura un congo?  Available now for $5.  
                      www.fluencyfast.com

                      MISSION: The mission of Fluency Fast is to create and sustain a movement  that causes a global shift in consciousness by transforming communications among  individuals, communities, and countries and inspiring people to use language as a tool to build bridges with other cultures.  Our goal is to dispel the myth that learning languages is difficult and to inspire people to have fun learning Arabic, French, German, Mandarin, Russian and Spanish, easily, inexpensively, effectively and in a brief period of time. 

                      Email sent to this address is automatically added to our periodic mailing list announcing upcoming classes.  Please request that we delete your email after our correspondence if you do not wish to receive our bi-annual upcoming class emails.  Fluency Fast is an equal opportunity educator and employer.  We do not discriminate on the basis of race, color, gender, creed, sexual orientation, disability or age.



                    Your message has been successfully submitted and would be delivered to recipients shortly.