| dc.description.abstract | In recent years, research on literacy has shown a particular interest in the topic of fluency. Building fluency seems to be an essential part of the transition towards skilled reading. Similarly, the concept of fluency has received attention in the realm of numeracy research. Here, findings suggest that arithmetic fluency might be an important facet of skilled mathematical performance. Further, reading fluency and arithmetic fluency overlap, suggesting that development may depend on some of the same abilities. Rapid automatized naming (RAN) is considered one of the best predictors of reading fluency and has also been found to predict arithmetic fluency. Combining these findings, it would be interesting to see whether the prediction of RAN extends to the overlap of reading- and arithmetic fluency. In this study I will answer two research questions. The first question asks whether RAN predicts fluent performance in either of these academic domains. The second question asks whether RAN predicts common causes of the overlap between these types of fluency. The questions are: Does RAN predict either reading fluency or arithmetic fluency, separately, after controlling for working memory and processing speed? Does RAN predict the covariance of reading fluency and arithmetic fluency, after controlling for working memory and processing speed? This thesis is written in conjunction with the ongoing, longitudinal research project NumLit: Development of Numeracy and Literacy in children. NumLit is a project carried out by the Department of Education and the Department of Special Needs Education at the University of Oslo. For the final analyses, I use data obtained from 219 of the children participating in the project. These children have been tested annually since pre-school, on a range of measures relating to numeracy, literacy, as well as general cognitive abilities and language. The current study uses data from the first-, and the most recent wave of measurement; pre-school and third grade. The predictors: RAN, working memory, and processing speed, were measured in pre-school. While outcome measures of reading fluency and arithmetic fluency, were measured in third grade. To answer the research questions, we will first attend to descriptive statistics and bivariate correlations. After this, structural equation modelling (SEM) is used to create latent factors and specify relationships between the variables. This allows us to go beyond prediction of performance on single measures, and instead predict performance in more holistic constructs of fluency. SEM analysis was carried out in R-Studio (RStudio Team, 2021), with the lavaan package (Rosseel, 2018). Descriptive statistics and bivariate correlations were provided through jamovi (The jamovi project, 2021). Rapid automatized naming (RAN) provided a sizeable prediction of both reading fluency and arithmetic fluency, even after controlling for working memory and processing speed. When modelling the covariance of reading- and arithmetic fluency, as a second-order factor, RAN was the sole predictor of this factor. The prediction of RAN on the overlap of fluency, was found not to be due to demands for processing speed or working memory present in RAN performance. Processing speed instead predicted reading fluency directly, and working memory had a direct effect on arithmetic fluency. This means that processes involved in kindergarten RAN, are similar to processes common between the fluency outcomes in the third grade. | eng |