After imprecise Pre-Term Planning data led to another semester of unexpected rises and dips in course enrollment this spring, administrators are considering making changes to the tool.

Pre-Term Planning, which has been in use for three semesters, was developed to provide administrators and professors with a rough headcount of possible enrollees.

This semester’s Pre-Term Planning data led to a historic low in the number of additional teaching fellows that had to be hired after shopping week for Gen-Ed and Core classes, according to Dean of Undergraduate Education Jay M. Harris.

However, administrators are searching for ways to make the tool more predictive of actual enrollment across the board, Harris said.“We are well aware that the tool is horrible,” Harris said. “It’s not completely accurate. But for the most part, it’s working.”

For example, the Office of Undergraduate Education is considering pushing back the deadline for students to submit their Pre-Term Planning choices.

However, a later submission date for Pre-Term Planning might make it more difficult to implement classroom changes and to hire additional TFs, Associate Dean of Undergraduate Education Stephanie H. Kenen said.Harris’ office is also looking “to adopt the ease” of the popular course shopping tool created by students in Computer Science 50.

Harris said that the tool needs to be compatible with other systems within the university, while also ensuring that student data is stored securely. Administrators differed in their interpretations of the objective of the Pre-Term Planning tool.

According to Harris, the purpose of Pre-Term Planning “is not to manage the chaos of shopping week,” but to reduce the number of TFs hired after courses begin.

Kenen said she believes that reducing chaos is one of the objectives of the tool.

“The primary goal of Pre-Term Planning is to reduce the chaos of the first two weeks of classes and improve the educational experience for both students and instructors,” Kenen said.

Professors similarly had mixed impressions of the usefulness of Pre-Term Planning.

Head TF for Statistics 104: “Introduction to Quantitative Methods for Economics,” Kevin A. Rader, said that he finds estimates based on historical enrollment numbers more accurate than Pre-Term Planning numbers.

“We more or less ignore the Pre-Term Planning numbers when hiring TF’s,” Rader wrote in an email.

When Pre-Term Planning data indicated that 154 students—only six fewer than total enrollment last spring—showed interest in taking Statistics 107: “Introduction to Business and Financial Statistics,” Rader and Statistics senior lecturer Michael I. Parzen thought the number would be accurate. But 382 undergraduates enrolled in the class this spring—a 146 percent rise from last year.

“Ideally it’s useful for the COOP, but as far as I can tell, we don’t use it for anything useful,” Parzen wrote in an email. “I think we recognize that it’s a new system and not fully implemented yet. The numbers tell that story.”

In an effort to make Pre-Term Planning easier to use, Kenen said that she hopes students will be able to submit their choices this summer through the study card tool currently used for enrollment.

“I think the proposed changes might help,” Astronomy professor Dimitar Sasselov wrote in an email. “I am all in favor of experimenting. That’s the best approach to improving it.”