Cynthia Kilpatrick

Placement vs Proficiency: (How) is this Placement Test Right for Our Program?

Cynthia D. Kilpatrick, PhD

Interim Director, English Language Institute and Assistant Professor in Practice

Department of Linguistics and TESOL, College of Liberal Arts

This project examines whether the current placement test used at the English Language Institute (ELI) is predictive of success for students in different proficiency levels. The ELI English for Academic Purposes program prepares students with limited English proficiency for study in academic classrooms at US universities and colleges. The program consists of 7 levels, and students are able to enter fulltime academic study upon completion of the program. Students entering our program are placed into a particular level based on their initial placement test scores. However, the current placement test is relatively new to us, and our program has undergone significant changes over the past 2 years. Our goal is to place students as appropriately as possible, so that they can complete the program in a timely manner and move successfully into academic coursework.

Placement test scores for students who entered the program in each session since January 2016 were gathered and divided into distinct subsection scores for listening, grammar, vocabulary, and reading. These scores are then compared to program data, including the following:

Teacher impressions of student proficiency
Student grades and levels
Session End exam in Spring 2017

By comparing and correlating these data, we hope to gain a better understanding of how to better use our placement test to assign students to the most appropriate level for the highest degree of success in their language learning process.

J. T. Dellinger

Hi Cindy! Interesting project overall. One question - do you think that cultural heritage potentially plays a role in listening vs. reading scores for ELLs here at UTA? Also, did you notice any demographic differences that might have influenced results? How many students do you typically have at each level at a time? Thanks!
- Cindy
  
  I absolutely think that culture and the first language play a role in the scores we see! For many students who have had to learn a whole new alphabet as they learned English, listening and speaking skills tend to be higher, but that also depends on educational culture. We have not looked specifically yet at demographic data in relation to these specific scores, but it’s on the list! Our levels over the past year have ranged from 3-21 students, but the majority cluster toward our middle levels.
Regina Urban

I appreciate understanding the order of the subtests, and I agree with you that the student may experience exam fatigue by the time they get to the last test (reading). I also think that 35 minutes to complete 55 questions seems challenging. That’s only about 35 seconds for each question. If all of the reading questions are clustered at the end, I’m wondering if you have data on whether the questions are skipped / not answered because they ran out of time (affecting your reading score). Interesting! Thanks Cindy!
- Cindy
  
  We do need to look at questions skipped, but an even bigger issue is the random guessing that they do as they are running out of time, since we have no way of knowing if they guessed or just got it wrong.
Jenny Roye

Nice job, Cindy, always interesting to hear of current work in other colleges.
- Cindy
  
  Thanks!
Maria Trache

I think your project has many implications for practice. The placement tests are crucial to make sure students are not over- or under-prepared for the level they take. I assume they are tested many times and in multiple ways to keep them matched to the right instructional level. Not an easy task.
- Cindy
  
  It’s so much more complicated than we realize!
MEAGAN ROGERS

The appropriate use of placement tests is a fascinating topic, Cindy. Out of curiosity, is this topic gaining traction in the same way remedial education has been highlighted as adding to the “swirl” experienced by many college students today (low persistence, high student loans, etc). However, it seems as if language proficiency is far more complex than math proficiency!
- Cindy
  
  I think if we looked at placement scores on something like the Accuplacer/TSI we’d see more of a swirl, but since we are a relatively small non-credit program, students have less of a chance to experience that - we track attendance and grades pretty carefully, and because the majority of our students are on F1 visas, they don’t get any financial aid and cannot drop out or stop attending or they endanger their immigration status. We have only a few non-F1 students, and not enough to draw any conclusions from.