Phonic skills mediate improved literacy outcomes in at-risk primary school children: results from the Ready-to-Learn Programme

Background
Learning to read is a key goal during primary school: reading difficulties may curtail children’s learning trajectories. Controversy remains regarding what types of interventions are effective for children at risk for academic failure, such as children in disadvantaged areas. We present data from a complex intervention to test the hypothesis that phonic skills and word recognition abilities are a pivotal and specific causal mechanism for the development of reading skills in children at risk for poorer literacy outcomes.
Method
Over 500 pupils across 16 primary schools took part in a Cluster Randomised Controlled Trial from school year 1 to year 3. Schools were randomly allocated to the intervention or the control arm. The intervention involved a literacy-rich after-school programme. Children attending schools in the control arm of the study received the curriculum normally provided. Children in both arms completed batteries of language, phonic skills, and reading tests every year. We used multilevel mediation models to investigate mediating processes between intervention and outcomes.
Findings
Children who took part in the intervention displayed improvements in reading skills compared to those in the control arm. Results indicated a significant indirect effect of the intervention via phonics encoding.
Discussion
The results suggest that the intervention was effective in improving reading abilities of children at risk, and this effect was mediated by improving children’s phonic skills. This has relevance for designing interventions aimed at improving literacy skills of children exposed to socio-economic disadvantage. Results also highlight the importance of methods to investigate causal pathways from intervention to outcomes.

Machine learning classification of surgical pathology reports and chunk recognition for information extraction noise reduction

Background and aims: Machine learning techniques for the text mining of cancer-related clinical documents have not been sufficiently explored. Here some techniques are presented for the pre-processing of free-text breast cancer pathology reports, with the aim of facilitating the extraction of information relevant to cancer staging.

Materials and methods: The first technique was implemented using the freely available software RapidMiner to classify the reports according to their general layout: ‘semi-structured’ and ‘unstructured’. The second technique was developed using the open source language engineering framework GATE and aimed at the prediction of chunks of the report text containing information pertaining to the cancer morphology, the tumour size, its hormone receptor status and the number of positive nodes. The classifiers were trained and tested respectively on sets of 635 and 163 manually classified or annotated reports, from the Northern Ireland Cancer Registry.

Results: The best result of 99.4% accuracy – which included only one semi-structured report predicted as unstructured – was produced by the layout classifier with the k nearest algorithm, using the binary term occurrence word vector type with stopword filter and pruning. For chunk recognition, the best results were found using the PAUM algorithm with the same parameters for all cases, except for the prediction of chunks containing cancer morphology. For semi-structured reports the performance ranged from 0.97 to 0.94 and from 0.92 to 0.83 in precision and recall, while for unstructured reports performance ranged from 0.91 to 0.64 and from 0.68 to 0.41 in precision and recall. Poor results were found when the classifier was trained on semi-structured reports but tested on unstructured.

Conclusions: These results show that it is possible and beneficial to predict the layout of reports and that the accuracy of prediction of which segments of a report may contain certain information is sensitive to the report layout and the type of information sought.