An Unusual Distribution

In August 2014, the first ever AP Physics 1 classes were begun. At Little Rock Central High, they replaced two courses at once: Pre-AP Physics, a class for mostly freshman, and AP Physics B, a second-level physics class for primarily sophomores, juniors and seniors.

As a result, the four sections of this class had a most unusual distribution of students. There was nearly an equal distribution of grade levels (20% Freshmen, 31% Sophomores, 22% Juniors, 27% Seniors). This is remarkable for most classes at Little Rock Central High, where our freshmen classes consist of primarily freshmen, and most other classes have students concentrated in one or two grade levels.

It was important to track the progress of students in this newly formed class, in order to determine whether grade level indicated a significant hurdle to excellence. The tests in this course were normalized about the median scores, and could present an unfair disadvantage to younger students who were competing against older ones with more experience and knowledge. Thus, in order to study the null hypothesis (that grade level played no part), a statistical analysis of each major test was conducted.

There were approximately 2 major tests per nine week term, in addition to a 1st semester final and an overall comprehensive exam. As the last test given was without seniors, it was omitted from the study.

Results of statistical analysis demonstrated the null hypothesis. Utilizing an alpha of <0.05 (95% certainty), it was found that there were very few statistically significant differences in average tests scores vs. grade levels. In only four circumstances (out of a possible total of 90 comparisons) were significant differences found between grade levels, and none were significantly different that the overall averages.

Of the four circumstances, every one pointed to a lower grade level performing better than the higher one, with three of the four being comparisons with seniors.

In conclusion, grade levels made no significant difference on performance, with more than 95% of comparisons between classes and overall averages showing no statistically significant variation.

Further studies to be done may compare normalized deviations between classes to see whether some classes had a greater range of assessed skills than others. It may also be informative to study the effect known as "senioritis," which may have played a crucial role in some of the minor variation observed.

Attached below is the raw data from which the comparisons were drawn, along with informative graphical analysis. All data has been deidentified, and this study was a part of normal classroom practices.

Patrick Foley

Physics Teacher

Little Rock Central High School