In lieu of an abstract, here is a brief excerpt of the content:

• Heterogeneous Academic Achievement Profiles of Initially STEM-Intending Students Over the College Years
• Tong Li, Chris Kirk, and Leticia Oseguera (bio)

Academic achievement, often measured by GPA, has been extensively studied in the literature of science, technology, engineering, and mathematics (STEM) education, including its impact on student persistence and success in college (Rask, 2010). However, most research has only looked at students’ performance at a single point in time, such as their first-year GPA, and has not explored how their academic performance changes over time. As a result, there is a gap in our understanding of the longitudinal aspects of students’ intellectual journeys through college. While some evidence has suggested that students follow different paths of development or academic progress over time (Fesseha et al., 2020; Hong &You, 2012; Robinson et al., 2018), only a few studies have investigated features of college students’ academic achievement changes over the long term, particularly in STEM fields. The literature on college students’ academic performance over time has suggested that while it can be unstable, many students can improve their GPA throughout their college studies. Humphreys (1968) made this observation, and more recent longitudinal studies have supported this idea, including a study by Mabel and Britton (2018). Factors affecting the variability of college students’ academic performance over time, such as GPA slope and variance, have also been investigated. Cheng et al. (2012) found that female students performed better over time and experienced less GPA instability when they had higher levels of family social support. However, research has identified persistent disparities in academic achievement in specific student populations. Sharkness et al. (2010) found that graduating seniors who entered college with an interest in STEM majors exhibited a significant cumulative GPA difference between White students and their Black and Latino peers, even after controlling for precollege academic preparation, college experiences, and institutional contexts.

While these studies attempted to uncover the longitudinal features of college students’ academic performance, there has been relatively little research into the different change patterns, particularly among those who entered college with an interest in a STEM major. This study aimed to fill this gap by adding a nuanced understanding of the academic performance change patterns of a group of initially STEM-intending college students at State University and was guided by two research questions: (a) What are the change patterns in those students’ academic performance across different semesters? (b) How are these patterns related to their background and college experiences? [End Page 728]

METHODOLOGY

We conducted latent profile analysis (LPA) on eight semester-GPA scores of a group of initially STEM-intending students to identify the change patterns of students’ longitudinal academic profiles. LPA is a statistical method that aims to identify distinct groups of personal attributes based on continuous variables by fitting multiple models with an increasing number of profiles until a model that best fits the data is found (Spurk et al., 2020). Although there is no universally agreed-upon minimum sample size for conducting LPA, some scholars have suggested that a sample size of more than 500 is sufficient to detect the correct number of latent profiles in the data (Tein et al., 2013). Using this as a guideline, we selected a total of 625 students, which comprised a group of aspiring STEM students who were participating in a STEM support program, along with a comparable group of peers who were selected based on their similarity in terms of race, gender, and intended STEM majors at State University.

State University is a public, land-grant research university with an enrollment of more than 48,000 students as of fall 2022. It has an acceptance rate of approximately 50% and a 6-year graduation rate of around 85%. The university has a predominantly White student body, with over 60% of students identifying as White. Approximately 46% of students are women. The students included in this study were admitted to the university between 2012 and 2016 and reported an intention to pursue a STEM major at the time of admission.

The primary indicator we focused on in this study was students’ academic performance, which we operationalized as the composite GPA score at the end of each semester. Through a program...

中文翻译：

---

最初打算学习 STEM 的学生在大学期间的学业成绩差异

以下是内容的简短摘录，以代替摘要：

• 最初打算学习 STEM 的学生在大学期间的学业成绩差异
• 佟丽、克里斯·柯克和莱蒂西亚·奥塞格拉（简介）

学术成就通常以 GPA 来衡量，在科学、技术、工程和数学 (STEM) 教育文献中得到了广泛研究，包括其对学生坚持不懈和在大学取得成功的影响（Rask，2010）。然而，大多数研究只关注学生在单个时间点的表现，例如他们第一年的 GPA，并没有探究他们的学业表现如何随着时间的推移而变化。因此，我们对学生大学期间智力旅程的纵向理解存在差距。虽然一些证据表明，随着时间的推移，学生遵循不同的发展路径或学业进步（Fesseha et al., 2020; Hong & You, 2012; Robinson et al., 2018），但只有少数研究调查了大学生学业特征从长远来看，成就会发生变化，特别是在 STEM 领域。有关大学生学业成绩随时间变化的文献表明，虽然学业成绩可能不稳定，但许多学生可以在整个大学学习过程中提高 GPA。Humphreys (1968) 做出了这一观察，最近的纵向研究也支持了这一观点，包括 Mabel 和 Britton (2018) 的一项研究。影响大学生学业成绩随时间变化的因素（例如 GPA 斜率和方差）也得到了调查。程等人。(2012)发现，随着时间的推移，当女学生获得更高水平的家庭社会支持时，她们的表现会更好，并且 GPA 不稳定的情况也会更少。然而，研究发现特定学生群体的学业成绩持续存在差异。鲨鱼内斯等人。(2010) 发现，对 STEM 专业感兴趣的大学毕业生在白人学生与黑人和拉丁裔同龄人之间表现出显着的累积 GPA 差异，即使在控制了大学前的学术准备、大学经历和机构背景之后也是如此。

虽然这些研究试图揭示大学生学业成绩的纵向特征，但对不同变化模式的研究相对较少，尤其是那些对 STEM 专业感兴趣的大学学生。本研究旨在通过对州立大学一群最初打算学习 STEM 的大学生的学业成绩变化模式进行细致入微的了解来填补这一空白，并以两个研究问题为指导：(a) 这些学生的变化模式是什么' 不同学期的学习成绩？(b) 这些模式与他们的背景和大学经历有何关系？[完第728页]

方法

我们对一组最初打算学习 STEM 的学生的八个学期 GPA 分数进行了潜在概况分析 (LPA)，以确定学生纵向学术概况的变化模式。LPA 是一种统计方法，旨在通过将多个模型与越来越多的个人资料进行拟合，直到找到最适合数据的模型，从而基于连续变量来识别不同的个人属性组（Spurk 等人，2020）。虽然进行LPA没有普遍认可的最小样本量，但一些学者建议超过500的样本量足以检测数据中正确数量的潜在剖面（Tein et al., 2013）。以此为指导，我们总共选择了 625 名学生，其中包括一群有抱负的 STEM 学生，他们正在参加 STEM 支持计划，以及一组根据种族相似性选择的同龄人，性别，以及州立大学的 STEM 专业。

州立大学是一所公立、赠地研究型大学，截至 2022 年秋季在校学生超过 48,000 名。录取率约为 50%，六年毕业率约为 85%。该大学的学生群体以白人为主，超过 60% 的学生是白人。大约 46% 的学生是女性。本研究中的学生于 2012 年至 2016 年间被该大学录取，并在入学时表示有意攻读 STEM 专业。

我们在这项研究中关注的主要指标是学生的学业成绩，我们将其转化为每学期末的综合 GPA 分数。通过一个程序...