chapter_5.tex

%   Filename    : chapter_5.tex 
\chapter{Conclusion and Recommendations}

\section{Conclusion}
The Philippines faces a significant learning crisis characterized by poor reading proficiency among elementary students. High percentage of Filipino children perform poorly in reading fluency and speed, and the country obtained low scores in global reading comprehension assessments. This issue hinders the growth and development of individuals, and has broader implications for the country's economic development. Although the Department of Education has designed the Phil-IRI program to alleviate the problem among elementary schools, challenges impede the program to produce positively significant outcomes. Phil-IRI’s manual implementation is arduous and time-consuming for teachers leading to poorly designed intervention programs.

CORA provides teachers and students with an application that can assist in solving the problem at hand. The application automates the Oral Reading Test component of the Phil-IRI program which measures a test taker’s word reading score, reading comprehension, and reading rate. It uses an automatic speech recognition and transcription service to transcribe test taker speech into text which is used to detect oral reading miscues. After obtaining the speech recording, the application provides the test taker a short quiz regarding the passage they read to assess the extent of their reading comprehension. CORA further calculates their reading rate by marking the time that a test taker started reading a passage and computing the number of seconds elapsed when they finished reading. The application then processes the data, detects the oral reading miscues, and generates a comprehensive oral reading profile of the students. To detect and count the number of reading miscues, CORA employed Needleman-Wunsch algorithm. The algorithm is an effective method to detect common oral reading miscues between the reference text (passage) and transcribed text. It can detect substitution, insertion, and omission. 

Additionally, the application allows teachers to view and assess students' results. Summarized results aid in planning for efficient and appropriate reading interventions. Since the Oral Reading Test component is divided into a pretest and a posttest, the application allows the teachers to toggle the availability of tests to ensure an orderly assessment.

With this application, teachers can avoid the tedious and inefficient implementation of the Oral Reading Test component of Phil-IRI. This will increase opportunities to ensure the provision of specialized reading instruction or intervention, which is the crucial part of the program.

\section{Recommendations}
The application is still in the early stages of development and is expected to undergo further improvements. Currently, the miscue detection algorithm generalizes the types of miscues identified in the Phil-IRI manual into three, namely substitution, insertion, and omission. Future research could further evaluate the reliability and validity of the miscue detection system by comparing its results with those obtained through manual assessment by trained assessors.

The developers recommend including the detection of other types of reading miscues or disfluencies, such as hesitations and interjections. An algorithm or framework that can specify the types of miscues from a speech would supplement effective assessment. 

Additional functionalities are also suggested, such as the automated conversion of results into other document file formats, such as .csv (comma-separated file), .xls, and/or .pdf. This could provide teachers with a physical copy of the student's results. Furthermore, the developers recommend automating other types of literacy tasks in the Phil-IRI program, such as the Silent Reading Comprehension and Listening Comprehension, and incorporating the Filipino language into the application, as CORA currently only supports Oral Reading Test in English.

Lastly, the developers recommend gathering feedback from teachers and students regarding the usability and effectiveness of the web application. It is also recommended to investigate the impact of using the automated system on the motivation and engagement of students.