关键词: 新生适应, 项目反应理论, 多层模型, 纵向Rasch模型, SAS GLIMMIX

Abstract: University attendance is a critical transition for many late adolescents. Many researchers have argued that successful adjustment, particularly during the first year, can predict students’ academic success, mental health and personal development. Thus, it is imperative for researchers to understand how adjustment problems emerge over time and vary across different students and develop prevention and how intervention efforts aid a smooth and productive transition. However, most of the traditional longitudinal analyses just focus on the general trend of change in people with time but ignore the individual differences. Moreover, few of the longitudinal studies consider the stability of instruments. The present longitudinal study aims to better understand the overall trend and individual differences of how academic, communication, and emotional adjustments of freshmen change over time, concurrently with an assessment of the invariance of the item location.
With an initial sample of 1940 participants recruited in 6 universities, the participants in the follow-up study sampled from only 1 university of the initial sample in convenience. New subjects were included in while some were lost during the following 3 waves of measurement. The final sample going to analysis comprised 1952 freshmen who were used for at least one measurement.
The adjustment of freshmen had been assessed 4 times during the first year, in Oct 2005, in Feb 2005, in Apr 2006, and in Jun 2006 using the Adjustment Questionnaire of College Freshman (AQCF). With SAS GLIMMIX procedure performing the analysis, longitudinal Rasch model was used to fit the polytomous repeated-measures data.
The following results were obtained: (1) Although the majority of students were well adjusted in academic study, emotion, and communication, academic transition is the most challenging for freshmen. (2) Academic and emotional adjustments had increased whereas communication adjustment declined during the first year in general. For academic adjustment has negative linear time effect and positive quadratic time effect, the change of academic adjustment over time appears a U-shape, with the lowest score at the end of the first semester. (3) There was significant variation in adjustment among freshmen on the entrance to college, but the trend of change with time was the same for all students.(4) The item locations of Academic scale were invariant over time, whereas some item locations on emotion and communication scales had changed linearly with time.
Student development professionals should pay special attention to students with poor adjustment. When they provide useful guidance to those students according to their specific needs during their first year, it should be noted that academic transition was the most challenging one and the academic adjustment may drop at the end of the first semester. When researchers use the AQCF questionnaire, the stability of emotion and communication sub-scales should be considered. Although longitudinal Rasch Model is useful in many ways, it still needs more investigation to popularize it

Key words: freshman adjustment, Item Response Theory, Multilevel Model, longitudinal Rasch model, SAS PROC GLIMMIX