The Online Teacher's Assistant : Using Automated Correction Programs to Supplement Learning and Lesson Planning
Year: 2014
Type: article
Abstract: This paper will focus on the use of several online software programs that streamline the process of finding and correcting student errors, helping educators pinpoint key areas for instruction. Using smartphones or computers, students input assignments into an online form, which is then sent to a spreadsheet. The teacher can then easily assess assignments manually as well as by using a variety of automated grammar/language tools. Based on the results, the teacher can tailor lesson plans specifically to address the more common errors of a given class. The information also assists the instructor in correcting submissions and can help students individually monitor and improve their writing. These programs can save educators time and resources as they quickly assess and identify problematic areas. This paper will help educators begin to implement these useful online tools in their own classes. Method During the course of their Freshman English class, a sample set of 17 Asia University Business Hospitality students completed a total of 135 book summaries containing 4,830 words and approximately 650 sentences over the course of the Spring 2013 semester. The information was gathered during the The Online Teacher’s Assistant 125 course of the Asia University Freshman English intensive reading course’s weekly reading assignment. This information was then assessed using several programs which checked for grammar, spelling, and word choice. In this course I require students to read seven to ten books throughout the semester and, using their smartphone or computer, electronically input and send their work to me. While the weekly book report assigned to the students includes a variety of questions, for this particular data gathering process only, one section of their weekly Book Report (Book Summary) was analyzed. The methodology for this paper comes from the relatively new field of Natural Language Processing. “Natural Language Processing (NLP) is an area of research and application that explores how computers can be used to understand and manipulate natural language text or speech to do useful things” (Chowdhury, 2003). Student data has been analyzed using three different web-based applications (grammarly.com, paperrater.com, and writewords.co.uk) which measure word frequency, phrase frequency, various grammar errors, and spelling mistakes. I collected the data using Google Forms and Microsoft Excel. Google Forms is a free online tool that allows users to input information and collects it in a spreadsheet. It is useful for teachers to help gather information from students for a variety of activities and assignments. In the intensive writing aspect of an ESL/Writing course, it provides numerous benefits, most notably the ability to access all student responses in one place. This allows educators to easily view and correct the information as well as adapt lesson plans to strengthen common class mistakes. The Online Teacher’s Assistant 126 Google Forms has a variety of organizational and visual features that make it ideal for classroom use. Time stamps for student-generated submissions can help simplify the grading process for teachers. Also, by using the Summary of Responses feature, the user can view a variety of graphs and information. For example, with multiple choice questions, Google Forms can create a graph of student responses based on each question. This can be used as a visual aid in the classroom to convey to students their problematic areas by showing a bar graph of the class’s answers. Using the graph, the teacher can immediately focus on questions which had the most incorrect answers. With the use of online error correction programs, teachers now have access to a variety of tools that can identify spelling, lexical, and grammar issues. This paper will take the sample set of student assignments and discuss the implications based on these three areas. Once the students submitted their weekly book report, the submissions were input into the Google Forms spreadsheet, which was then exported to several programs. The first, Grammarly, is an online proofreading website that can be used to scan documents for grammar mistakes. According to its website, “Grammarly scans your text for proper use of more than 250 advanced grammar rules, spanning everything from subject-verb agreement to article use to modifier placement” (Grammarly, 2013). It also offers style-specific correction for a variety of different types of writing. In this particular study, the “Student/Academia” setting was used to assess student writing samples. The site also offers “context optimized synonyms” and an “Adaptive The Online Teacher’s Assistant 127 SpellChecker” which claim to offer both spelling and word choice suggestions based on content. Another feature of the site is plagiarism detection, which checks writing against a database of eight billion web pages. By taking the entire data set and running it through Grammarly, I discovered patterns of common mistakes made by students. Grammarly sorted through the sample set, checking for spelling, grammar, punctuation, and style/word choice. From the fifteen pages of student-generated input, it created a forty-page analysis of areas to improve. By no means is it an exhaustive list of all errors created; however, it illustrated topics which the class could benefit reviewing. To obtain a more comprehensive analysis of the student's’ writing samples, I used writewords.org.uk, Microsoft Excel, and paperrater.com to discover word frequency and other pertinent data. Writewords.org.uk is a website designed for professional writers designed to help with proofreading. These tools sorted through the document to discover the most commonly used words, which I then classified manually by part of speech. I then aggregated the data using a Microsoft Excel spreadsheet to discover the average number of words per assignment students had submitted. Additionally, Writewords has a function that counts the frequency of phrases. The user can set the counter to find the most commonly used two-to-ten-word clusters. Combining these tools allows educators to assess student vocabulary and create activities (based on its results) to encourage expansion. Paperrater.com offers a similar service to Grammarly in that it also acts as an online proofreader. The Online Teacher’s Assistant 128 PaperRater is designed by linguistics professionals and graduate students. According to its website, “PaperRater combines the power of natural language processing (NLP), artificial intelligence (AI), machine learning, information retrieval (IR), computational linguistics, data mining, and advanced pattern matching (APM)” (paperrater.com). PaperRater provided interesting word usage statistics, including verb types used and parts of speech used to begin sentences, I chose to use Grammarly’s reports because its software highlights areas for improvement and sorts mistakes into categories based on type. These software tools, when combined, were able to process a semester’s worth of written class work in approximately two minutes. They created detailed summaries of student errors and habits. Over the course of about an hour, Grammarly was able to create individually tailored reports for the entire class. This is in contrast to the several hours (or potentially days) an instructor would need to manually correct and categorize errors. With these programs it is possible to tally and gather information that would prove to be monotonous and inordinately time-consuming otherwise. Data The study sample consisted of seventeen students who, over the course of sixteen weeks, electronically submitted 135 book summaries totaling 4,830 words. On average, students submitted forty-seven words per assignment to summarize their reading. Students were directed to write a minimum of three to four sentences per book summary. The most common errors, according to Grammarly, were in the areas of punctuation (178), spelling (79), sentence structure (55), articles (45), and passive voice (45). The Online Teacher’s Assistant 129 According to Writewords.org the most commonly used nouns were story (33), one (31), and day (28). The most common verbs (counting all tenses) were be (232), go (63), and have (33). The most commonly used adjectives were old (20), big (17), and strange (13). Lastly, the most commonly used adverbs were very (23), about (17), and finally (9). Writewords’s phrase frequency counter listed the most common threeword phrases as “This story is” (13), “to go to” (11), and “this book is.” The most common four-word phrases were “this story is a” (8), “there was a man” (4), and “story is a story” (4). Phrase frequency dropped as the number of word clusters increased past five. However, an interesting pattern began to develop of students who had either copied assignments or plagiarized from the book. Setting the phrase frequency counter to fiveor six-word clusters yielded several identical phrases, which were indications of duplicate submissions. (See the Discussion section for a more in-depth analysis of plagiarism patterns.) According to PaperRater, the vast majority of sentences were started with a pronoun (152). Considering the Grammarly software indicated the high incidence of punctuation errors in student writing, this is notable because it indicates students associate punctuation with pronoun usage. The second most common occurrence after pronouns was starting sentences with a conjunction (45). Paperwriter went a step further and dissected sentences, reporting that of the total work submitted, 11% of the writing included pronouns and 10% included a preposition. The Online Teacher’s Assistant
Author: Schraudner Michael
Language: English
Subfield: Education
Field: Social Sciences
Domain: Social Sciences
SDG: Quality education
Open Access status: green
OpenAlex ID: W267475890
