Reading Level Placement and Assessment for ESL/EFL Learners: The Reading Level Measurement Method Aaron David Mermelstein, Ming Chuan University   Abstract For many ESL/EFL learners, reading is their main learning goal and may be the most important of the four language skills in the second language. For many ESL/EFL teachers, the issue of promoting proficient reading is important, and they look for additional or alternative methods to achieve this goal. However, when discussing reading as a primary goal of a curriculum, assessment becomes a critical element necessary for successful instruction. This article suggests that extensive reading (ER) should be used to achieve this goal and discusses a new method of reading assessment, the Reading Level Measurement Method (RLMM), as a practical means of assessing learners reading levels and accurately placing learners at their optimal reading levels to maximize their learning potential. Further, this article describes the step-by-step process of creating the RLMM, how to implement it in the classroom, and how to use it as a practical and economical assessment and placement tool. It also describes a successful quantitative study which adopted the RLMM and supports its effectiveness. Key Words: reading level, assessment, placement, extensive reading Introduction Stapleton, 2007) in spite of the current emphasis around the world on For many English as a second communicative language teaching language (ESL) and English as a foreign (CLT). When discussing reading as a language (EFL) students, reading is their primary goal of a curriculum, assessment main learning goal (Carrell, 1993; Day becomes a critical element necessary for & Bamford, 1998; Grabe, 2009; successful instruction. Postlethwaite Mermelstein, 2013; Nunan, 1999; and Ross (1992) found that regular Waring, 2006; Tamrackitkun, 2010), and assessment was a key factor associated for most of these students, reading will with students’ success in learning to be the most important of the four read. Assessment can help teachers language skills in the second language. determine if the instruction provided is Therefore, it is no surprise that in many resulting in adequate student progress. It countries, reading has become one of the can assist in identifying students who most emphasized skills in the ESL/EFL can benefit from a more accelerated classroom (Day & Bamford, 1998; instructional program and those who Grabe, 2009; Nation, 2001; need more intensive instructional Tamrackitkun, 2010; Tanaka, & 44   ORTESOL  Journal,  Volume  32,  2015 support. Essentially, assessment allows 2004; Pigada & Schitt, 2006). For the teachers to identify where their purpose of this article, ER is defined as students are academically, and where reading as much as possible within the they need to go. For example, if a learner’s peak acquisition zone, for the reading assessment determined that a purpose of gaining reading experience class of students was at a lower reading and general language skills and level, the teacher may decide to change obtaining pleasure from the texts. the course textbooks to use a graded readers or simplified texts instead of an Statement of the Problem original novel. Furthermore, if it was In order to better build up ESL/EFL determined that the same group of students’ English reading abilities, students’ reading level was low, the ESL/EFL teachers need to develop teacher may also decide to alter the English teaching programs that apply writing content of the course and provide teaching methodologies and reading more scaffolding for individual students materials that are efficient, effective, and if necessary. matched with the students’ abilities and For many ESL/EFL teachers and interests. Although many current administrators, the issue of promoting educational systems throughout the proficient reading is important. In order world are achieving some level of to do this, some teachers are looking for success, they may still be lacking additional or alternative methods to appropriate and/or alternative models for achieve this goal. In the field of teaching vocabulary and reading teaching ESL/EFL, there has been a (Mermelstein, 2013), such as student- trend for teachers to rely on skill- centered learning activities like ER. building textbooks that attempt to Therefore, the problem is not develop reading strategies for the learner necessarily the lack of effort teachers are to comprehend different genres of texts. putting into creating a reading These strategies are usually designed for curriculum for their students, but it may the purposes of teaching the reader how be a lack of knowledge regarding how to to find some sort of general or specific properly assess and place learners within information in the text. such a program. To further complicate Today, however, there is an the situation, most ESL/EFL courses are enormous amount of research spanning comprised of students with mixed decades promoting the effectiveness of abilities, which makes it even more extensive reading (ER) (i.e., Day & difficult for teachers to teach reading Bamford, 1998; Lee & Hsu, 2009; through traditional methods, like the Mermelstein, 2013; Nuttall, 1982; Sheu, Grammar-Translation method or direct 2003; Waring, 2006; Tamrackitkun, teaching of vocabulary. 2010; Yamashita, 2013). In fact, Nuttall Today there is a wide variety of (1982) stated that “an extensive reading reading assessment tests and assessment programme…is the single most effective formulas on the market (i.e. the Flesch- way of improving both vocabulary and Kincaid Grade Level Formula; the Fry reading skills in general” (p. 65). This Readability Graph), but most of these belief has also been echoed throughout are norm-referenced tests intended for more recent research (e.g. Cho, 2007; mainstream students and not for Day & Bamford, 2002; Day, & Hitosugi, 45   ORTESOL  Journal,  Volume  32,  2015 ESL/EFL learners. Further, they are approach in second and foreign language intended to compare students with what classrooms is Krashen’s (1985) Input is considered the normal reading level Hypothesis, which is based on the for each grade in school, which is also distinction between acquisition and not compatible with ESL/EFL students learning. Accordingly, the term because their age does not necessary acquisition is used to refer to an intuitive reflect their grade level in terms of or subconscious process of constructing English language ability. or “picking up” a language. The term learning, on the other hand, is used to ESL/EFL students are uniquely refer to a conscious, active effort to different from mainstream students understand information. Later, Krashen whose primary language (L1) is English (1991) created a more specific part of because they generally do not have the the Input Hypothesis known as the background and cultural knowledge to Reading Hypothesis, which states that perform as well on these types of comprehensible input in the form of reading assessments. For example, they reading can also stimulate language may not know the meaning behind terms acquisition. Since reading is the like traffic jam or sky scraper. They interaction of new information with old may also not possess the same linguistic knowledge, learners who understand knowledge of phonology, semantics, and most of the text can infer the meaning of syntax as mainstream students. new words as they read and then test ESL/EFL students may or may not have their hypothesis as they encounter these learned decoding skills, and students words repeatedly through reading. from different cultures may apply different strategies while reading, which One point that has not been may also be difficult to detect. In challenged when discussing ER is that addition, there is the problem of lexical vocabulary control is necessary. knowledge. ESL/EFL students may be However, there has been some debate able to sound out words in their minds or among researchers regarding the amount of aloud, but this does not mean that these vocabulary knowledge that is necessary for a words make sense to them. second language reader to accurately comprehend a text. In order to learn word Therefore, in order to assist teachers meanings incidentally through reading, it in developing English reading programs is important for learners to encounter a that apply methodologies and reading suitable number of unfamiliar words in a materials that are efficient, effective, and text. According to Liu and Nation appropriate to students’ abilities and (1985), Laufer (1987), and Hirsh and interests, this article suggests the use of Nation (1992), learners need to ER and offers a new and easy method of understand about 95% of the text in properly assessing learners and placing order to gain an adequate comprehension them within an ER program. and to accurately guess unknown words from the context. Hill and Thomas Literature Review (1988) suggest 90%, but the Extensive Reading Foundation (ERF) (2011) sets ER Theories & Programs the percentage much higher at 98%. At One of the major theories underlying 95% coverage means that there is the modern development of the ER approximately one unknown word in 46   ORTESOL  Journal,  Volume  32,  2015 every two lines of text, if each line of throughout the world (e.g. Day and text contains about ten words. It is vital Bamford, 1998; McQuillan, 2006; that learners know a sufficient amount of Mermelstein, 2013). There are several headwords and word families in order to differences between extensive reading understand 95% of the words in a text. and intensive reading. The first is the According to Hirsh and Nation (1992), a amount of reading materials that the word family is the base form of a word plus learner is required to read. The second the inflected and derived forms created from is the degree of intensity with which the affixes, which may include affixes like the materials are to be read. During third person –s, the superlative –est, -able, - intensive reading activities, learners are ness, etc. A headword is a word in which generally exposed to short texts which a group of related words would appear include specific lexical and/or syntactic with it in a dictionary. For example, if aspects of the language to learned, and one looks up the word automobile in a are usually followed by tasks to provide dictionary, it might be accompanied with reading strategy practice. The goal of related words like auto, car, motorcar, ER is different because ER attempts to vehicle, etc. Both Laufer (1997) and immerse the learner in large quantities of Nation (2001) suggest that a size of 3000 comprehensible input without any word families should cross the threshold specific linguistic task, except making and be enough for successful second meaning of the text. In other words, language reading. Liu and Nation learners are generally not asked to do (1985) demonstrated that words in a additional work related to their reading. low-density text, where there was only Day and Bamford (1998) offered the one unknown word out of twenty five following ten top principles of ER to words, were easier to guess than words help clarify the common characteristics in a high-density text, where there is one of successful ER programs: unknown word out of ten words. 1) Students read as much as possible. According to the ERF (2009), reading is at an “instructional” level when the 2) A variety of materials on a wide students know between 90% and 98% of range of topics are available. the words on a page. If the students 3) Students select what they want to know 98% or more of the words, then read. they are in the ER “sweet spot” and can 4) The purpose of reading is usually read quickly or at a constant pace related to pleasure, information, and because there are not too many unknown general understanding. words slowing them down. Further, at 98% it is most likely more enjoyable for 5) Reading is its own reward. the reader as well. Since vocabulary 6) Reading materials are well within the control has such a large impact and is so linguistic competence of the vital for reading comprehension and the students. acquisition of new vocabulary, proper assessment and placement into an ER 7) Reading is individual and silent. program is paramount. 8) Reading speed is usually faster rather Extensive Reading than slower. Extensive reading has been widely 9) Teachers orient students to the goals advocated for language learning of the program. 47   ORTESOL  Journal,  Volume  32,  2015 10) The teacher is a role model of a over the course of several years out of reader for students. the need to accurately assess and place mixed ability ESL/EFL students into ER can make a positive contribution reading programs and track their to the development of competence in a progress throughout the programs. The second language (Lee & Hsu, 2009; RLMM is specifically designed to work Mermelstein, 2013; Tamrackitkun, 2010; together with ER programs using graded Tudor & Hafiz, 1989; Yamashita, 2013) reader books. by providing learners several encounters with unknown words. It expands the There has been one published study learners’ interaction with the language and several as-yet unpublished studies by exposing them to different words in carried out using the RLMM. All of the different contexts so that learners can studies used the RLMM for the receive a more complete understanding participants’ placement in an ER of their meaning and use (Simensen, program, but two of them specifically 1987). The pedagogical value attached measured the reading level improvement to ER is based on the assumption that of the participants over the length of the having students interact with large studies. amounts of interesting, meaningful, and Mermelstein’s (2013) study was comprehensible language materials will conducted in Taiwan, using 4th year EFL produce positive effects on the learners’ university students. It was a 12-week abilities to use a second language. quantitative study involving 87 Several influential second language participants and its purpose was to studies involving ER have taken place examine the effects of ER on the reading over the past few decades and have levels of the students and to find an demonstrated several distinct benefits appropriate alternative to the traditional that language learners can receive teaching methodologies being used in through ER. They can acquire more Taiwan. After the initial reading level vocabulary knowledge, increase their placement, the students participated in a reading speed, improve their writing weekly in-class Sustained Silent Reading abilities, create more positive attitudes activity, supplemented by outside towards reading and the target language, reading, using graded reader books. The and develop their linguistic knowledge overall framework of this study was (Bell, 2001; Cho, 2007; Fernandez de based upon Day and Bamford’s (1998) Morgado, 2009; McQuillian, 2006; top ten principles for conducting a Mermelstein, 2013; Nation, 2008). successful extensive reading program, as listed above. First, a measure of the The RLMM means and standard deviations of the results of the pre and post RLMMs were The Reading Level Measurement taken and then a categorical analysis was Method (RLMM) is an assessment tool done on the reading level data using Chi- created by teachers and designed to square. The results indicated that the measure learners’ reading levels for the treatment group posted significantly purposes of placement within a reading higher gains than the control group, with program or for measuring reading level 1.01 levels gained vs. 0.46 levels gained progress over time. It was originally respectively. The Chi-square analysis, created by Aaron David Mermelstein using a two-way contingency table with 48   ORTESOL  Journal,  Volume  32,  2015 the two variables being the starting The rating scales for each series reading level and the ending reading differs, so it would be inaccurate to level, found the variables to be measure reading level based on book significantly different with Pearson Chi- level. Some series distinguish levels square (2, N = 87) = 115.72, p = 0.000, based on multiples of 100 word families Cramer’s V = 0.52. and others are based on multiples of 200 or more word families. In other words, RLMM Design one publisher’s level 3 book does not The design for creating the RLMM is directly correspond to another’s1. simple and not very time consuming. Therefore, one of the most important From start to finish, it takes factors for creating the RLMM and approximately one hour to create. towards assessing student development Teachers would create six sections of the is consistency within the texts, or only pre- reading level test (one for each using one graded reader series. reading level) to be used as assessment Once the RLMMs have been and/or placement and six sections of the completed, it is then important for the post-test to be used for assessment and teacher to determine the percentage of comparison. All of the RLMMs would unknown words that is acceptable at a be based on books from one specific students’ placement level. As previously graded reader series to measure mentioned, this is also one of the key improvement over the duration of the ER factors involved in using this RLMM program. If desired, the teacher could because the percentage of known or also create a midterm RLMM to match unknown words will have a huge impact needs or goals. The RLMM is based on on the learner and the learner outcomes. the principle that the book publishing For example, if a learner was placed at a companies have already done the most level where 90% of the words were difficult part for the teacher, that is, known, this would mean that the learner create graded reading levels based on the would not know 1 in 10 words, and it number of word families (e.g., help, would be difficult for the learner to helpless, helplessness, unhelpful, etc.) understand what they were reading. used in the text. Although it is not an They might need to frequently stop and exact science due to the actual number use a dictionary and would mostly likely of words in any text and the fact that feel frustrated as well. Therefore, there is no way to specifically measure selecting the right placement level for which words are being learned, in learners is paramount to success. principle, each company has already Diagram 1 explains the difference calibrated each level of its book series between a vicious circle of reading vs. a with a range of word families. virtuous circle, which is directly related Therefore, moving from one level to the to the percentage of unknown words next represents an increase in the while reading. learners’ reading level and the number of word families learned by the reader.                                                                                                                 How many word families are learned at 1  For a more complete comparison of each level depends on which graded graded reader scales by various reader series teachers have chosen as publishers, see the Extensive Reading well as which books learners have Foundation’s (ERF) (2011) graded selected. reader scale.   49   ORTESOL  Journal,  Volume  32,  2015 Diagram 1. Vicious vs. Virtuous reading cycle. Source: ERF (2013) Creating the RLMM English and select only books of a similar style throughout the levels. This The first and most important material will add to the validity and reliability of a teacher will need to get started is a set the RLMM. If students are unfamiliar of graded reader books using only one with British English texts, it is specific publisher’s series. While it is recommended to only use American not necessary to have a complete set of English texts in order to eliminate graded readers, it is important to have an confusion due to differences in spelling. adequate number of books available to students at all levels, so that they can The third step is to randomly select freely select books of interest from a full pages from each text with wide pool of genres. Most series have approximately the same number of levels 1-6. However, some series have words on each page. How many additional levels, like beginner, starter, RLMMs teachers want to make will or advanced. It may not be necessary to determine how many pages are selected create a starter level section of the from each book at each level. For RLMM, as students who find level 1 too example, if teachers only want to give a difficult would automatically be placed pre-test and a post-test, then they would at the starter level. select two pages from each text at each level, one to be used for each test. The second step is to select three Therefore, teachers would actually be books from each level for a total of 18 selecting a total of 6 pages for each books. When selecting books, teachers level. Using pages of the same texts for should be aware of the differences both the pre- and post-tests increases the between British English and American 50   ORTESOL  Journal,  Volume  32,  2015 validity and reliability of the assessment. underline all of the unknown words to Before finalizing the page selections, them. They should be instructed not to teachers read each one and determine if stop reading to use a dictionary. When it is of standard level quality, that is, if they finish reading each page, they they have a variety of word families on should count the number of unknown each page. If possible, it would be ideal words, only counting each unknown to have two other teachers analyze the word once, and then write the total test, so results can be normed. Having amount on the top of the page and circle two or three teachers agree on the it. standard would increase the statistical RLMMs are designed to take inter-rater reliability of the RLMM; approximately 50-60 minutes. However, however, this step is not necessary. each teacher may need to adjust the The final step is to piece all of the amount of time given for their own pages of the test together. The test students, since they know their students’ should start with the lowest level, and abilities best. More advanced readers there should be three pages of text for should be able to move through the each level for a total of 18 pages. lower sections of the test relatively Teachers should not worry if each page easily and quickly. Lower level students of the RLMM does not flow together as will move more slowly through the a single story. They should inform the RLMM and perhaps not finish. This is students of this fact prior to the test. It is okay and all of the students should be recommended to create a customized informed at the beginning that some of cover page for the RLMM where them might not finish. Since lower level students can write their names and other students would struggle through the important pieces of information. The higher levels and have a higher final version of the RLMM is also percentage of unknown words, it is not versatile, in that it may be given as a important whether they actually attempt hard copy, or it can be scanned and to read levels out of their range. given online. Either way, it is Assessing the RLMM Results recommended to scan the RLMM and keep an e-file which can be reused in the As indicated earlier, assessing the future, saving both time and money. RLMM for placement and comparison Another versatile aspect of this RLMM depends on what percentage of known or is that a teacher can adjust the RLMM to unknown words the teacher wants to set. assess as many levels as they deem In order to find the percentage of words, necessary, perhaps only creating an teachers should count the words on each RLMM with three levels. page. For some teachers, an approximate count may be adequate. So, Implementing the RLMM for example, if a page has approximately Implementing the RLMM is also 200 words and the teacher would like to simple for both teachers and students. set the percentage of known words at First, the teacher distributes the tests to 98%, then there would be approximately the students and indicates that they 4-5 unknown words underlined by the should not start reading until they are student on this page. Since the RLMM given the command. For each page of is designed to have three pages at each the test, the students are to read it and reading level, a simple average of the 51   ORTESOL  Journal,  Volume  32,  2015 three pages is all that is necessary to this book and select another book within identify difficulty level. However, it is the same level. If the second book is important to note that in this specific also too difficult for the student, or there example, if a student had identified 5 are too many unknown words present, unknown words on two of the pages at then the teacher can place the student at level 4 and 7-10 unknown words on the one level lower in the graded reader third page at the same level, this would series and make a note of the change. indicate that the student be placed at However, this should be done as soon as level 3. It is far better to allow the possible to maximize learning and student to read at a slightly lower minimize student frustration. reading level than a higher reading level For an overall assessment using the because the goal is to keep the reader as two RLMMs, all that is necessary is to close to the designated percentage of do a comparison of the starting level unknown words without going over. For with the ending level. Having a midterm example, if the goal is 98% of known RLMM is suggested since notifying the words and a learner is placed at a learners of any improvement may act as reading level of 95% of known words, a motivating factor increasing their the text may be too difficult for the reading outside of class reader to read without being able to infer the meaning of unknown words. Discussion Therefore, the learner would most likely need to stop reading to use a dictionary, While the RLMM does not measure which defeats the purpose and benefits any specific vocabulary gains, it does of ER. measure improvement or gains in reading levels. Therefore, a reasonable Placement can also be somewhat assumption would be that with an fluid. After the initial placement, the increase in one’s reading level comes an students should be informed that they increase in one’s sight vocabulary and/or should let the teacher know if the one’s acquisition of head words, and reading books they select are too easy or thus, an improvement of literacy. too difficult. This can also be Further, with an increase in one’s determined by the percentage of reading level, one’s reading fluency unknown words on randomly selected should also be improved (Mermelstein, pages from the texts that the students are 2015). reading. However, it should be noted that not all of the books within the same While many of the benefits and reading level will be equally challenging advantages of using the RLMM have for the individual student. One reason been discussed in the previous section, for this may be due to the familiarity of there may still be some teachers who the topic or the familiarity with specific may still ask why a classroom teacher vocabulary, but also because while there would use the RLMM instead of using is standardization within every graded one of the tests provided by the reader book level, each book involves publishers of graded reading books, for different vocabulary. Therefore, if a which there are several key reasons. selected book at any level is found to be First, publishers’ placement tests are too difficult for the student, they should meant to interact only with the specific also be told that they can stop reading books that they are trying to sell. They 52   ORTESOL  Journal,  Volume  32,  2015 are not compatible with any other graded access to the internet, and the speed or reader series. However, the RLMM is quality of the internet service. fully compatible by design to work with any graded reader series. In addition, Conclusions many of the publishers’ tests are only available for a fee or by purchasing one Considering the limitations of the of their programs. This can actually be classroom and the time available for quite expensive and not in the budget of teachers to directly interact with each most classroom teachers. In student, ER may be able to help second comparison, the RLMM is practically language learners become more free. In addition, most of the publishers’ autonomous learners and improve their reading comprehension tests are merely language skills in a multitude of ways, vocabulary tests (e.g. Oxford, Penguin, especially in ESL/EFL environments in and MacMillan), in which the words which exposure to the target language may be completely random and taken may be limited. However, assessing out of context for the learner. Often they learners’ reading levels and reading level are multiple choice tests and/or use placement poses very real problems for pictures for the test takers to guess their language teachers towards instructional answers. Since the test taker can guess planning and accountability. answers, the results of these tests are Misplacement generally means missed likely to be invalid and inaccurate, and learning opportunities, which can have a since the publishers do not provide any catastrophic impact on learner statistical data regarding their tests, one motivation. This can lead the learner to can only conclude their tests have not a negative view of themselves and the been statistically measured for validity language they are learning. It can also or reliability and there have not been any lead to misbehavior in the classroom. correlation studies done regarding test Stakeholders in the educational scores and placement. system want to accurately measure Another point to consider when learners and see improved results in comparing publishers’ tests with the reading. The RLMM offers a simple and RLMM is time. Each time a teacher or real solution to a real problem since it student wants to use one of the can be designed and implemented in publishers’ tests, it takes a considerable approximately two hours. Also, the amount of time to navigate through the flexibility of the RLMM allows it to be publishers’ website, and there is also no used and reused over and over again, way of storing information without either via a hard copy or an electronic printing it out or copying it to an e-file. version. It can be adapted to any graded Therefore, collecting and storing reader book series and include any students’ data can take a considerable reading level. In addition, for amount of time and effort. In addition, administrators and researchers, it can having students take an online test is also offer significant statistical data that also dependent on many outside factors: is both valid and reliable. the number of computers available, 53   ORTESOL  Journal,  Volume  32,  2015