**1. Introduction**

Learning disabilities have numerous negative impacts on the educational progress of affected students. Not only is it assumed that school difficulties are consistent throughout the time of a student's school career [1–3], but the problems can also be observed up to adulthood [1]. Against the background of these findings, the proportion of children with problems in learning in Germany is very worrisome. In a Germany-wide survey in 2016, the reading performance of almost 30,000 fourth graders was examined against the background of nationwide standards [4]. In the sample, more than 12% of the children did not reach the minimum standard. These children are therefore able to read simple texts and understand their meaning, but the information must be explicitly specified in the text. To ensure a successful transition to secondary education, intensive support for these children is necessary. Only about 10% achieved the optimum standard, which means that these children can cope with clearly demanding requirements. They can think independently about texts, grasp topics and motives that are not explicitly given in the text, and draw complex conclusions. About 65% of the tested children reached the minimum standard in reading. The results underline the need for specific support for those affected as early as possible, as spontaneous remission cannot be assumed [5].

Similar findings in the USA led to far-reaching legal reforms. With the No Child Left Behind Act in 2001, various standards were enacted to optimize the quality of teaching and support. These include universal screenings on identifying children at risk, better-trained teachers and support staff, and the use of evidence-based instruction and intervention. Moreover, high-stakes assessments have been developed to evaluate the quality of teaching. In this context, curriculum-based measurements (CBM, see below) have gained in meaning. In the context of a response-to-intervention model, CBM provide ongoing information on the learning progress during the school year. On this basis, data-based decisions as well as a formative evaluation of instruction and student support can be made to optimize the teaching-learning framework [6].

The situation in Germany varies considerably [7]. Although there are regular, nationwide tests, whose results are used to assess the effectiveness of the school system, there are no consequences for individual schools (in the sense of direct assumption of responsibility). The approval of financial and human resources to support children with learning problems in Germany requires an official diagnosis of disabilities (e.g., dyslexia). Only then will children have access to additional educational resources. Accordingly, diagnostic measures serve to determine the need for support and do not indicate the need for the adaptation of instruction. Data-based decision-making processes and formative evaluation of the instruction-based on standardized data are unusual [8]. This is because teachers cannot judge the idea behind them and the significance of their use [8]. In addition, the selection of suitable progress monitoring instruments such as CBM is still limited, but constantly growing [9,10]. Against the background of these aspects, this paper presents the construction and evaluation of an item pool from which economic CBM instruments for primary schools in German-speaking countries can be generated. We assume that the word recognition skills can serve as an indicator to screen for low reading fluency skills.

#### **2. Reading Fluency**

One of the most important goals of school attendance for all primary school students is the successful acquisition of reading skills. From the transition to secondary school onwards, it is assumed that students will be able to independently extract and understand information from texts [11]. Historically and currently, much importance is attached to reading fluency in connection with the acquisition of reading skills [12,13]. The literature reveals a multitude of attempts to define reading fluency [14–16]. Current attempts see fluency as the result of a successful interplay of different basic competencies [13,17]. The US-American National Reading Panel [16] published a frequently cited definition. They describe fluency as the ability "to read orally with speed, accuracy, and proper expression" [16] (p. 5). Accordingly, reading fluency is a process of appropriate recoding and decoding of what has been read, the quality of which depends on various aspects, such as reading accuracy and phonological, orthographic, and morphological abilities [12,13,17–19].

Reading fluency, however, is not only the outcome of a successful combination of different partial competences, but also often seen as a prerequisite for higher reading skills [20,21]. In this context, emphasis is placed on the speed of reading [12,22,23]. The basic assumption is that insufficient word reading skills (slow, stagnant, and erroneous decoding) is an obstacle to the contraction of individual information into larger units of meaning. This in turn complicates the processes of activating prior knowledge, integrating new information into existing knowledge structures, and metacognitive control processes [21]. Only when the word reading process is automated do resources become available for higher forms of information processing, i.e., more complex reading processes [16,24–26]. Especially in the first years of school, clear connections between the ability to quickly recognize and decode words and reading comprehension can be empirically depicted [27,28]. With increasing reading experience, the students' mental lexicon expands, and frequent combinations of sounds, morphs, and words are stored and linked. Automated word recognition and rapid decoding thus form the basis for appropriate reading fluency at the sentence and text level, which is in turn necessary for the successful understanding of meaning [12,29].

Against the background of the different understandings of reading fluency listed above, it becomes clear that early interventions are necessary to promote reading fluency in order to prevent reading problems [20,30,31]. Reading fluency should therefore be understood in a development-oriented way [20]. Attention should be paid to it even at the beginning of reading acquisition. At this point, this concerns aspects of phonemic segmentation, alphabetic understanding, phonics, and orthography [12,30], as well as word recognition [32]. As with reading fluency, word recognition skills can be seen as an outcome, as well as a predictor. As a result of the interplay of letter and sound knowledge, as well as decoding abilities, word recognition skills serve as an outcome variable [16,33].

According to the National Reading Panel [16], fluency is the direct result of successful word recognition. Overall, word recognition skills can be assumed to be a potential indicator of reading fluidity [12,13]. In this sense, the assessment of word recognition skills (amount of words read identified correctly within a limited time span) over time plays an important role in preventing reading difficulties. According to the study by Speece and Ritchey [34], word recognition skills develop at the same time as the first word recognition processes and are therefore already important in grade 1. At the end of grade 2, most students should have acquired fundamental word recognition skills [35], and by the middle of grade 4 at the latest [36]. The assessment of basic reading skills, such as the precise recoding and decoding of words, should be a goal of instruction in the first grades at school [33]. Based on this data, further pedagogical decisions can be made.

#### **3. Assessing Reading Fluency with Curriculum-Based Measurements**

Curriculum-based measurements (CBM) [37] are a very prominent approach for progress monitoring of academic skills. CBM were developed in the USA and already have a long research tradition there, especially in the fields of reading, writing, and mathematics [37–39]. The original aim of the use of CBM was to provide teachers working in special education with reliable and valid data for assessing the development of students, in order to support instructional decisions [37]. These short test procedures can be used regularly at short intervals. Within a time limit of only a few minutes, the children have to solve as many tasks of a test as possible. Due to the repeated application, the monitoring of academic progress can be derived [40]. On the one hand, CBM can be easily implemented in school routines. On the other hand, the instruments must correspond to the current standards of psychological tests, so that the results can be clearly interpreted.

Depending on the domain of use, CBM may refer to separate competencies that are curricularly identified for the area (e.g., CBM for addition tasks in the numerical range up to 20) or that can be regarded as an indicator of general outcome (e.g., reading aloud as an indicator of general reading skills). Alternatively, they may bundle different partial competences relevant to the domain in a single instrument (CBM with mixed tasks for calculating sizes, for factual tasks, etc.) [37,40–43].

The origin of CBM research lies in the field of reading. Accordingly, many methods have been published in this domain [43]. Reading fluency is often assessed by reading aloud individual syllables, words, or texts [44]. The working time is limited to one minute. The test leader documents the correctly read syllables or words.

A large number of research findings show, in particular, that measures of fluency are relevant with regard to students' reading skills [45–48]. According to Fuchs et al. [44] and Reschly, Busch, Betts, Deno, and Long [49], oral reading fluency can be assumed to be a reliable indicator of overall reading competence.

While much attention has been paid to oral reading, there is a lack of research related to the silent reading of students [35,50]. One justification for this can be found in the conclusions of the National Reading Panel [16]. Accordingly, there is a lack of empirical research on the effectiveness of silent reading experiences [51]. Therefore, adequate time should rather be given to reading aloud in class [51–53]. In reality, however, silent reading is the most important form of reading from the first grade onwards [50]. Empirical findings indicate that there is a high correlation between oral reading and silent reading. This is particularly true for gifted readers and in higher grades [53–55].

#### **4. Research Questions**

Our research refers to the word recognition skills of elementary school students. In order to early identify struggling elementary school students, we want to generate CBM to assess their word recognition skills. To create di fferent CBM instruments, we designed an item pool, from which items can be flexibly selected according to content-related but also psychometric aspects. The aim of the study presented here is to test the psychometric suitability of the item pool.
