Analysis of Information Networks of Freshman Engineering Students

The main purpose of this study was to determine if social interaction within the University of Tennessee, Knoxville college freshman engineering classrooms correlates with academic performance. Also of interest was whether the interactions between genders had a significant affect on academic performance. Better academic performance is cited in the literature as improving retention and graduation rates; therefore, if factors that affect academic performance can be understood measures can be taken to help students perform better.

Five UT freshman engineering classes were surveyed to determine their level of involvement with the rest of the members in their class. Academic performance of the class as a whole and of each gender was retrieved from the class’s instructor at the end of the semester.

The demographic information revealed that there are significantly fewer females in engineering than males, however, the percent enrolled is consistent with that of the national average. Social networking analysis of the interactions within the class revealed that the women have a higher percent of interaction within the class than males do. However, classes overall do not have that much interaction.

The relationships did show that social interaction within a class could have an affect on student’s academic performance. While there was no significant relationship between the overall class grades and overall class density, strong relationships were revealed between overall class grades and gender-to-gender interaction and gender grades with respect to gender-to-gender interaction. A significant positive relationship was made between receiving A’s and an increase in male-to-female interaction (p = 0.038). As the A’s within a class increase, other grades will decrease; and thus, the class’s academic success rate increases. A weak positive relationship was made between the percent of males receiving A’s and the amount of male-to-female interaction; however, given a larger dataset, there may have been statistical significance.