Hello dear CI Peek readers! We’ve been seeing a lot of discussion in our Facebook group lately on topic of selecting readers, and picking the “just right” fit for a given class or level, so we have decided to re-publish this blog post from last May. We hope it will answer a few questions for you! We’d love to hear what you think!
Join us on Facebook to share and discuss!
Fluency Matters Facebook page
Fluency Matters Comprehension-Based Readers Facebook group
iFLT Conference Facebook group
CI-Intermediate Students Facebook group
There was recently a lively discussion on the iFLT/NTPRS/CI Teaching Facebook Group about how TPRS novels / Comprehension-based readers are assigned a level. The discussion made me realize that there is currently no uniform (or even remotely similar) formula or system among the various publishers/authors for counting unique words and/or assigning a level. Some indicated that they only count word families or head words (the first word under which a set of related glossary entries appears), they do not count prepositions, and other function words, or they analyze text in other ways.
The Facebook thread certainly seemed timely, as just a week or two ago another publisher, Contee Seely from Command Performance Language Institute, wrote to me asking for help/information about how to identify and count unique words and how we determine the level of a text. When I saw the same questions/comments on Facebook, I thought this was a good time to write a blog post about how we at Fluency Matters determine a level for each book.
Much like assessing language proficiency, assigning a level to a read is for the most part ambiguous and messy. There are numerous internal and external factors that affect the level of a reader, so it is next to impossible to assign a GRADE level. Grade or class level is not reflective of language level. Furthermore, some reads or topics that might be considered more appropriate for younger learners, are often just as beneficial for inspiring SLA in older learners. Just because a text is considered “easy reading” does not mean it is not appropriate for higher level students. (You can never choose a book that is too easy! Adults rarely read at their true reading level.)
GRADE level is a misnomer. There are various proficiency levels within one grade level in both L1 and L2 classrooms, so it is a bit misleading to say that a book is appropriate for a specific grade level. According to a report by the Center for Applied Second Language Studies (CASLS), the majority of students studying a foreign language in a traditional high school program reach benchmark level 3 or 4 by end of the fourth year of study, regardless of the language studied. These levels are similar to the ACTFL levels Novice-High and Intermediate-Low. If students are first reaching intermediate-low after FOUR years of study, can they really handle (comprehend, process for acquisition and enjoy) “authentic texts” in novice level classrooms?
A college professor from Southern Illinois University, Edwardsville uses Carrie Toth’s Hija del sastre in his 300 level course to help students gain the skills they need to read their first native text. He says that even college students at the higher levels of intermediate-high/advanced-low find much more success reading unadapted native texts when they’ve had a firm foundation in leveled readers. Fluency Matters readers are the ladder that gets them to that level.
Fluency Matters uses 3 general labels for grading readers: beginner, advanced-beginner and intermediate. Each of these levels is analyzed in terms of unique word count, level of vocabulary used, content, word frequency, and sentence complexity.
The first factor we consider when assigning a level is total unique word count. To be clear, Fluency Matters counts ALL words, including articles, prepositions, connectors, verbs, nouns, adjectives, etc. We do not count a word as unique if it is a close derivative of another word form. For example, let’s use the word family GO or “ir” in Spanish. Each form would be counted as follows:
Since ‘va’ and ‘vas’ are easily recognizable, they count as one individual or unique word. Forms that are not recognizable or similar, however, count as a unique word. Cognates are not generally considered ‘unknown’ or unique words, UNLESS they are high level words in English. External factors, such as literacy level, background knowledge, how much previous reading has been done, etc. will affect the status of a cognate. Students who read less and have lower literacy levels generally know fewer words in English, and therefore, fewer cognates in the target language.
The overall level of vocabulary is another important indicator of level. For a book to be considered beginner, at least 90% of the words need to draw from the top 200 words in the language. To give you an idea what that means in terms of our titles, I’ll compare 3 chapters of Brandon Brown quiere un perro with two nonfiction titles, Esperanza and Felipe Alou.
Brandon Brown quiere un perro
Books like the Brandon Brown series contain very common (high-frequency) vocabulary. The topics in each story are based on everyday events and contexts with which students are very familiar, and there are tons of illustrations, making them exceptionally comprehensible. Brandon Brown quiere un perro is unequivocally considered a beginner book. In a beginner book like Brandon Brown quiere un perro, words that fall beyond the top 200 words in Spanish (words like tampoco [neither] and regalo [gift]) are footnoted within the text.
Nonfiction stories like Esperanza and Felipe Alou require content-specific vocabulary with which students might not be as familiar, and that explains why they have a larger percentage of words that fall outside the top 100 and 200 most frequently used words. In the case of these two readers, words that fall beyond the top 300 words are footnoted. Words that are also used just once or twice in a very short span of text may also be footnoted if the meaning cannot be ascertained from context. Although Felipe and Esperanza have low unique word counts, we consider them advanced-beginner, due to the nature of the vocabulary and the concepts that are required to understand the story.
That said, content and required background knowledge are another factor that we consider when assigning a level to a reader. Background knowledge is crucial for understanding any read whether it is in your first language or your second, and therefore, it is another factor that is taken into account.
Sentence complexity is another common consideration in analyzing text in L1, and it is no different in L2. You will notice that as Fluency Matters readers progress from one level to the next, so does sentence length and complexity. Our authors naturally incorporate richer language, longer phrases and more complex sentence structure in an effort to create texts which are as authentic as possible. There are some who confuse this use of complex language as an attempt to focus on grammatical features/structures, but it is NOT. It is the natural result of writing with more literary language, without the linguistic parameters that restrict writing at lower levels.
Frequency factor is one major difference between beginner and intermediate readers. Frequency factor is the number of times core words repeat throughout a text. The more times words repeat within a text, the more comprehensible a text will be. We not only analyze the number of times that words repeat throughout an entire book, we also analyze frequency within a span of words and pages. It should be noted, however, that frequency should not be noticeable! Frequency factor does not justify writing monotonous texts for the sake of frequency! There are times when words/phrases are overtly repeated for emphasis/effect, but overall, frequency should not be noticed by the reader.
To get an oversimplified, very rough picture of the frequency factor, divide the total number of words by the number of unique words. This is NOT empirical, but it does give a very rough idea of frequency of core structures throughout a text.
|Title||Total Word Count||Unique Word Count||Frequency Factor|
|Brandon Brown quiere un perro||4100||75||54.66|
|Brandon Brown dice la verdad||4500||104||43.26*|
*Frequency of various forms of quiere is 90+
Suffice it to say that all of these strategies are a work in progress. We are continually trying to learn more about proficiency benchmarks, ACTFL performance indicators, and language acquisition in general. Word counts and text analyses, however they are done, are really just a guide to help connect readers with the text that will best meet their needs. Although there is no official or accepted system for counting unique words and analyzing texts, knowing just how each publisher evaluates their books will help teachers select an ideal text to lead a powerful reader-based unit for their own students.
You may also find Martina Bex’s blog post helpful: “Is This Novel Really Level 1?”