READING FLUENCY: NORMATIVE DATA FOR THE ROMANIAN-SPEAKING POPULATION

The purpose of this study was to establish reading fluency norms in the Romanian language for first grade students at the end of the school year using the PEAFC (Proba de evaluare şi antrenare a fluenţei în citire/ The Assessment and Training of Reading Fluency Instrument, Bodea Haţegan & Talaş, 2014). A representative sample of 1977 first grade students (age between 7 and 8 years old) was assessed: girls (N g =954) and boys (N b =1023) from different counties of Romania (N c =11), attending rural (N r = 385) and urban (N u =1592) elementary schools. Oral reading fluency norms were established on the linguistic structure of the words (monosyllabic, disyllabic, three syllable words and text words). The results of this study offer a new perspective on the reading fluency levels at the end of the first grade and will help teachers and specialists to identify students at risk for dyslexia and to start effective early intervention programmes to improve reading fluency.


Introduction
Reading is a very complex academic skill that the students achieve throughout the first grades; this skill influences the academic progress of the students over the years. But there are only few recent studies in the Romanian language regarding reading normative data in elementary grades for teachers to follow and try to adapt different reading materials for elementary grades.
Since 2012, the Romanian school system has changed a lot, children start school when they are 6 years old in a preschool class, and they start first grade when they are 7 years old. With this major change, the curricula and the teaching literacy strategies have changed and there are no research works at the national level to offer specific information about reading skills.
And that is exactly how I view the book you are reading\u2014as the most complete review, within a single cover, of our expanding knowledge of: the history of the English alphabet, the controversies surrounding phonics instruction, issues and research in early reading instruction, basic perceptual and reading processes, the processes involved in identifying sounds, letters, words and meaning, and the processes involved in learning to read.(foreword conclude that fluent readers can concentrate more on comprehension if the readers do not spend a lot of time on the decoding process.

Reading Fluency Assessment Tools
Reading fluency can be measured using a timer and different instruments (lists with letters, lists with words, lists with pseudo-words, lists with expressions, lists with sentences, lists with paragraphs or texts) and the teacher records the number of correct sounds or words the students can read in a well-defined period of time. Most tests use the one minute reading session to record the number of sounds or words the student can read. Deeney (2010) concluded that One Minute Measures proved to be an effective tool for teachers working with both typical and struggling readers.
Another author (Speece & Case, 2001) present different reading tasks to improve reading fluency. The Letter Sound Fluency (LSF) is a probe where the teacher records the number of correct letter sounds the student identifies per minute (LSF) (Speece & Case, 2001). The same authors develop the Oral Reading Fluency probe and the teacher records the number of correct words the student can read in a text per minute. The length of a text depends on the student's grade level: for the first grade students the text has approximately 150 words and for the second grade students the text has approximately 200 words.
The Rapid Automatized Naming (further RAN) is another probe where the teacher records the number of seconds the child needs to name six familiar objects out of 36 objects displayed in four lines.
Also, the Test of Word Reading Efficiency (Torgesen et al., 1999;Torgeson et al., 2009) measured the number of words the student could read in 45 seconds. The authors used different lists of words with increasing the difficulty level.
The Oral Reading Fluency Scale from the National Assessment of Educational Progress (further NAEP) is a tool first developed in 1992 (White, 1995). This scale, up-dated in 2002, has four reading fluency levels starting with the word-by-word to meaningful phrase groups with expressive interpretation.
Recently, Biancarosa and Cummings (2015) and to link these findings to practical uses of reading curriculum-based measurement (R-CBM underline the importance of the reading curriculum-based measurement (R-CBM) tools in order to offer the best and most appropriate data about reading skills.

Oral Reading Norms
There is a real need for all teachers to have the reading norms for each grade level. Most of the research works focus on the English language, but these norms cannot be assumed by other languages. The specific structure of each language, the specificity of the teaching methods to teach reading, the curriculum structure are important factors that can influence the reading norms for each grade level (Griffith & Rasinski, 2004;Hasbrouck & Tindal, 2006;Gagliano et al., 2015).
The following paragraphs present some reading norms developed for the English language readers, in order for these reading fluency norms to be considered in discussing the data collected for the Romanian population, as we considered that transcultural studies and research are very relevant in the speech and language field and they can reveal several transcultural invariants related with speaking, reading and writing skills. Typical Oral Reading Rates, grades one to six compress the following norms: grade one, 30-70 words; grade two, 50-100; grade three, 70-120; grade four, 90-140; grade five, 100-150; grade six, 110-150 (Barr et al., 1995;Baar et al., 2007).
The national oral reading fluency norms for students of the first to sixth grade, in three different periods of time: autumn, winter and spring, are presented by Hasbrouck and Tindal (2006). At 50 percentile for grade one, the norms are 23 in winter and 53 in spring; grade two, 51 in autumn, 72 in winter, 89 in spring; grade three, 71 in autumn, 92 in winter, 107 in spring; grade four, 94 in autumn, 112 in winter, 123 in spring; grade five, 110 in autumn, 127 in winter, 139 in spring; grade six, 127 in autumn, 140 in winter, 150 in spring.
Other fluency goals for student grades of the first to sixth at the end of the school year are available (Shanahan, 2006). The norms are: grade one, 60 words per minute (wcpm); grade two, 90 wcpm; grade three, 120 wcpm; grade four 130 wcpm; grade five, 140 wcpm; grade six 150 wcpm.

Intervention Strategies to Improve Oral Reading Fluency
Over the years, teachers tried to create different strategies to improve reading fluency. Some authors (Lo et al., 2011) highlighted the following intervention strategies used in repeated reading: error correction, adult modelling, performance cueing and feedback, preview and practice of isolated words and repeated reading training materials.
The "repeated reading passages" is a very popular strategy included in many intervention programmes and demonstrated to be efficient in different research studies (Samules, 1979;Dowhower, 1997;Kuhn & Stahl, 2003;Ardoin et al., 2008;LeVasseur et al., 2008). Ardoin, Eckert & Cole (2008) demonstrate significant differences on oral reading fluency in intervention passages when using the repeated reading intervention compared to the multiple exemplars for students in the second and fourth grade.
In their meta-analysis, the authors (Morgan & Sideridis, 2006) present the importance of motivational stimulus using the repeated reading strategy. Goal settings and performance feedback are important motivational stimuli included in the repeated reading programme. Direction discrimination training and digital image enhancement improved reading fluency not only for children with dyslexia, but also for typical readers (Lawton, 2008) which means they are twice as likely to drop out of school. Previous research has found that children who are slow readers have reduced contrast sensitivity for detecting the direction of movement, and that improving their movement contrast sensitivity by training with sinusoidal gratings moving relative to fixed background gratings significantly increases their reading fluency. Since observers having reduced contrast sensitivity show much faster reading speeds when text is sharpened with digital filters, it is likely that children will also read filtered text more quickly than unfiltered text. Methods: Orientation discrimination contrast thresholds were measured for both dyslexic and normal readers in grades kindergarten through third grade and used to construct individualized digital image enhancement filters. Computer-based reading speeds were measured for both unfiltered and filtered grayscale text before and after training on direction discrimination. Following training, reading speeds for both unfiltered and filtered equiluminant colored text were measured as well. Results: Reading rates were twice as fast

READING FLUENCY: NORMATIVE DATA FOR THE ROMANIAN-SPEAKING POPULATION
Carolina Bodea Hategan, Dorina Talas, Carmen Costea-Bărluțiu, Raluca Trifu when utilizing filtered text to compensate for losses in orientation discrimination contrast sensitivity compared to unfiltered text, both before and after direction discrimination training. Both filtered and unfiltered colored text was read at least 30% more slowly than filtered or unfiltered equiluminant grayscale text. The effects of training on direction discrimination were also significant for both dyslexic and normal readers (p < 0.008. In addition (Keehn, 2003), there are alternative methods to improve reading fluency: rereading, modelling, explicit instruction and a manageable text. In the last decades, different technologies were involved in both teaching in general and in different intervention programmes for struggling readers. Feedforward video self-modelling proves to be an effective strategy to improve not only reading fluency, but also comprehension, accuracy and reader self-perception (Robson et al., 2015).

Aim and Objective
This research is based on the following main objectives, in order to establish norms regarding reading fluency for Romanian speakers: • to compute normative values for reading fluency in first graders with typical development; • to compare the values obtained, based on linguistic criteria, namely to establish if reading fluency of words is dependent either on the syllabic composition or the fact that the words are presented isolated or as parts of a text; • to set up some diagnostic values for reading fluency abilities for Romanian students at the end of the first grade; • to determine if there are any gender and age differences regarding reading abilities in first graders; • to investigate the association between the performance in two different subjects (Romanian Language and Mathematics) and first graders` reading fluency abilities; • to identify the differences between children in the first grade, based on the environment/ setting they are coming from (rural or urban).

Research Design
The data was collected in compliance with a specified procedure. Researchers paid attention to the selection of the material used for assessment, the training of the people involved in data collection and the manner the obtained data was collected and registered.

Method Participants
The selection criteria for the participants were: • to have a typical development (children with disabilities were not included in the research study); • to be enrolled in the first grade The initial number of participants in the research was 1977, coming from 11 counties around the country: Bistriţa-Năsăud (BN); Sibiu (SB); MS (Mureş); Caraş-Severin (CS), Timişoara (TM); Sălaj (SJ); Cluj (CJ); Suceava (SV); Vâlcea (VL); Dolj (DJ); Maramureş (MM). Thus, the data was collected at the national level. This aspect is relevant for our research as the norms we intend to offer at the end should to be representative for the whole Romanian population. The group of participants in the research can be described according to children`s age (7 and 8 years old), gender (male and female) and the environment they are coming from (urban or rural area). The characteristics of the group are presented in Table 1. Analysing the data from the Romanian National Institute for Statistics (Recensământul populaţiei şi al locuinţelor, 2011), we can conclude that the group of participants in our research is representative for the Romanian population based on gender (the number of female children is higher than the number of male children for around 3%, the aspect also illustrated in the composition of our group). The representativeness is not reached regarding the environment the participants are coming from. In order to reach this goal of representativeness, the collected data should contain almost equal rates from urban and rural participants. An explanation for not being able to collect more data from rural areas is that speech and language therapists in Romania are mostly working in urban areas and a small number in rural areas. As the professionals trained to assess the children were speech and language therapists, collecting data from rural areas was very difficult.

Instruments and Procedure Material Used during the Assessment Sessions
The materials used during the assessment sessions were selected from the PFEAC, Proba de evaluare şi antrenare a fluenţei în citire/ The Assessment and Training of Reading Fluency Instrument (Bodea-Haţegan & Talaș, 2014;Bodea Haţegan & Talaş, 2015). The materials used consisted of: one list of monosyllabic words, one list of disyllabic words, one list of three-syllable words and one list with a text (Annex 1). The linguistic material collected in the PEAFC was selected following carefully the linguistic features of the Romanian language.

READING FLUENCY: NORMATIVE DATA FOR THE ROMANIAN-SPEAKING POPULATION
Carolina Bodea Hategan, Dorina Talas, Carmen Costea-Bărluțiu, Raluca Trifu Thus, the linguistic material can be considered as representative/ typical for the Romanian language. The selection of the four specific lists used in the assessment of each participant was performed randomly from the whole PEAFC.

Training the Data Collectors
Speech and language therapists were involved directly in collecting data for the current research. They participated in training sessions regarding the the PEAFC using procedure. One specific training session regarded the discussion of the summary report the professionals had to fill in after collecting the data (Table 2). The report was used in order for the data to be collected in a unitary manner as there were 28 speech and language therapists who collected data for the research.

Procedure
The data was collected within a short period of time, from 15 April to 30 May, 2015. Speech and language therapists involved in the research were helped by children`s primary teachers to organise the assessment sessions. Each of the four lists was supposed to be read by children in one minute, speech and language therapists were asked to fill in the report summary of the number of the words read correctly and the number of the errors that the students made. After the data was collected, speech and language therapists sent the reports to the researchers. The collected data was converted to the SPPS 17 data base and a statistical analysis was performed.

Results 1
In order to reach the first objective of the study, descriptive statistics tools (from SPSS 17) were used, and the results obtained are illustrated in Table 3. Based on the values of the skewness and kurtosis parameters, our results did not match a normal distribution and we found that there were multiple outliers, which were eliminated from further analysis. Following this operation, the normality assumption of the data was reached and the subsequent norms are presented in Table 4. Results were also grouped into quartiles, so that they reflect some milestones in reading fluency and were summarised in a diagnostic chart (Table 5). The chart underlines diagnostic values representing milestones for high risk to absence of risk in reading fluency for children at the end of the first grade. In order to establish if reading fluency of words is dependent either on the syllabic composition or on the fact that the words are presented isolated or as parts of a text, correlational and comparison tests were applied (after the distribution of the results was normalized by eliminating the outliers). All the correlations were highly significant (p ≤ .01 level). The highest correlation was obtained between the number of monosyllabic and disyllabic words correctly read (r = .924), while the lowest correlation value was obtained between the number of monosyllabic words and the words from the text read correctly (r = .856). These results underline the fact that monosyllabic and dissyllabic words do not require very different fluency reading skills, compared with three syllable words or text words.
Comparison between means was performed using Student's t test, in order to establish the statistical significance of differences in reading fluency skills dependent on the linguistic material read (Table 8). Comparison test showed significant results between the numbers of the correctly read words, underlining the fact the nature of the linguistic material significantly influences reading skills.
The same results are obtained, at the general level by using the analysis of variance of the differences (by using One-way Anova). Further analysis of the differences was performed based on gender, age, environment and school grades of the participants. A summary of the results is presented in Table 10. According to our data, the only relevant demographic factor contributing to performance in reading fluency is the child's environment (with children coming from rural areas having much lower performance, than those from urban areas). We tested the correspondence between school performance, reflected in the children's grades (the Romanian language and mathematics), and reading fluency. The results for differences in reading fluency, depending on the children's grades in the Romanian language are presented in Table 11. 12.6 (16.5) I 1.5 (3) * significant ** highly significant 1 very good, 2 good, 3 sufficient, 4 insufficient -Romanian grading system for first to forth grades As it is presented in Table 11, the differences in children's performance in the Romanian language are reflected in their reading fluency skills, so the reading fluency is highly associated with their school performance. Post-hoc analysis performed on our data prove the significance of differences between all the grades, thus proving that the lower the child's school performance in the Romanian language, the lower the performance in reading fluency.
Next, we tested the significance of the differences in mathematics depending on the children's reading fluency. The results are shown in Table 12.  (20) 162.19 ** B 25.9 (16.1) S 12.5 (15.9) I 4.6 (7.4) * significant ** highly significant 1 very good, 2 good, 3 sufficient, 4 insufficient -Romanian grading system for first to forth grades The results detailed in Table 12 show that the differences obtained in the children's mathematic performance are reflected in their performance in reading fluency. Also, the posthoc analysis of our data shows that those children who are low in their school performance in mathematics are also low in their reading fluency performance.

Discussions
The results demonstrate that the average number of read words per minute in the Romanian language is 35.83 (this average value is expressed for the overall reading fluency (Further ORF), after the elimination of the outliers. Before eliminating, the outliers and the ORF mean value = 39.52. The obtained values prove that ORF in the Romanian language at the end of the first grade has a mean value between 35 and 40 words per minute, a value comparable with the reading rates indicated by Barr, Blachowicz, Katz and Kaufman (2002) (they indicated a reading rate between 30-70 words per minute), but less than Hasbrouck & Tindal (2006) (they indicated 53 words per minute being an average value for reading fluency skills at the end of the first grade) and less than Shanahan (2006) who indicated a mean value of the reading rate of 60 words per minute at the end of the first grade.
The milestones charts offered based on these findings in Table 5 is a valuable tool for establishing a proper difference between children with low reading abilities and children at risk for learning disabilities. This chart is even a more valuable tool for special education teachers, primary teachers and speech and language therapists as this is the first standardised output regarding this aspect for the Romanian language.
The correlation between demographic aspects (such as gender, age and environment) and reading fluency rates was another aspect which was also investigated. The obtained results proved that environment was the only demographic aspect that influenced reading fluency rates. Students from urban areas proved to register higher reading fluency rates than children from rural areas. This aspect reflects a negative aspect of the Romanian educational system, this aspect is often interpreted as a result of the fact that teachers from rural areas either do not have the fully necessary competences and educational training to offer a proper teaching of the reading skills or their work is less sustained by specialised support teachers who intermediate and facilitate reading fluency acquisition. The rate of absenteeism and the school failure can be another aspect which may differentiate between students coming from rural or urban areas. According to some Romanian authors (Apostu et al., 2012), there are no significant differences at the primary school level between children studying in urban and children studying in rural areas regarding the rates of absenteeism and school failure. It seems that children from urban areas rate higher than children from rural areas. In this situation the last possible explanation for our research findings is related to the reduced number of children from rural (N r =385) areas participating in this investigation, comparing with the higher number of children from urban areas (N u =1592).
The last aspects investigated reading the relation between reading fluency rates and school performance emphasise reading fluency rates are highly reflected in the grades students obtained both in the Romanian language and Mathematic classes. This aspect strengthens the perspective offered by the European Dyslexia Association, (further EDA), which defines dyslexia as an umbrella concept for all learning disabilities "a difference in acquiring reading, spelling and writing skills, that is neurological in origin. The cognitive difficulties that cause these differences can also affect organizational skills, calculation abilities, etc." (European Dyslexia Association-EDA 2014).

Conclusions
The present study is the first one that presents new norms for reading fluency and offers teachers and specialists in the speech and language field the first consistent milestones in reading fluency skills, ensuring the differentiation of the good readers and the poor readers and raising awareness regarding reading in specific learning difficulties domain.
This research is also valuable as it is based on the PEAFC, an assessment and intervention tool completely adapted to the Romanian language speakers. The use of this assessment tool has a great importance as cultural comparisons can be entirely covered and justified.
Oral reading fluency skills in first graders proved to be an important aspect that must be considered when assessing their general school performance. For future research the aim is to continue with establishing oral reading norms for school age readers, including the fourth graders. At the national curricular assessment we consider that the norms for oral reading fluency should be taken into consideration, and their value should be increased by oral reading comprehension norms.
A main goal proposed by the European Commission for a "European Policy Network of National Literacy Organizations" (2013/C 130/07) was "to identify good policy practices in raising literacy levels among children, young people and adults, particularly focusing on low achieving students and adults with in adequate levels of functional literacy" (Garbee et al. 2015). The present study can be considered a good practice example in this field.
Further research studies in reading fluency for students at the end of each school year in primary school are recommended in order to establish norms for reading fluency at the end of each grade level and to identify the best intervention strategies for children at risk for dyslexia. reading in specific learning difficulties domain. This research is also valuable as it is based on the PEAFC, an assessment and intervention tool completely adapted to the Romanian language speakers. The use of this assessment tool has a great importance as cultural comparisons can be entirely covered and justified.
Oral reading fluency skills in first graders proved to be an important aspect that must be considered when assessing their general school performance. For future research the aim is to continue with establishing oral reading norms for school age readers, including the fourth graders. At the national curricular assessment we consider that the norms for oral reading fluency should be taken into consideration, and their value should be increased by oral reading comprehension norms.
A main goal proposed by the European Commission for a "European Policy Network of National Literacy Organizations" (2013/C 130/07) was "to identify good policy practices in raising literacy levels among children, young people and adults, particularly focusing on low achieving students and adults with in adequate levels of functional literacy" (Garbee et al., 2015). The present study can be considered a good practice example in this field.
Further research studies in reading fluency for students at the end of each school year in primary school are recommended in order to establish norms for reading fluency at the end of each grade level and to identify the best intervention strategies for children at risk for dyslexia.