我有一个不平衡（每个观察单位的观测数量不等）面板数据集。该数据集包含学生的个人数据及其机构隶属关系。关键信息是他们是否达到学位或退学。
在下面的示例数据集（df_raw）中，有两个变量描述学生是否达到学位。我的挑战是，学位并非在所有情况下都在正确的学期注册（由于管理延迟。滞后可以达到2/3个学期）。因此，示例数据集中的学生1被注册以获得2012年春季学位。我希望将该值1移至2011年，因为那是学位完成的时间。
学生2在完成哈佛大学学位后改变了机构。学位于2012年春季注册，当时学生开始在斯坦福大学学习。在这种情况下，我想把这个值移到哈佛大学的最后一个学期。
想要的结果在df_complete（下面）中说明。我的数据集包含超过30 000名学生和20万个观察。同样：数据集是不平衡的，并且学位注册的滞后因具体情况而异。
请注意：哈佛和斯坦福仅作为示例使用，因为大多数读者都知道它们。
谢谢。
df_raw <- data.frame(
student = c(1,1,1,1,2,2,2,2,2,2),
year = c(2010,2011,2011,2012,2010,2011,2011,2012,2012,2013),
semester =
c("Fall","Spring","Fall","Spring","Fall","Spring","Fall","Spring","Fall",
"Spring"),
institution = c("Stanford","Stanford","Stanford","Stanford","Harvard","Harvard","Harvard","Stanford","Stanford","Stanford"),
level = c("Lower degree","Lower degree","Lower degree","Higher
degree","Lower degree","Lower degree","Lower degree","Lower degree","Lower
degree","Lower degree"),
degree_same = c(0,0,0,1,0,0,0,0,0,0),
degree_other = c(0,0,0,0,0,0,0,1,0,0))
df_complete <- data.frame(
student = c(1,1,1,1,2,2,2,2,2,2),
year = c(2010,2011,2011,2012,2010,2011,2011,2012,2012,2013),
semester = c("Fall","Spring","Fall","Spring","Fall","Spring","Fall","Spring","Fall", "Spring"),
institution = c("Stanford","Stanford","Stanford","Stanford","Harvard","Harvard","Harvard","Stanford","Stanford","Stanford"),
level = c("Lower degree","Lower degree","Lower degree","Higher degree","Lower degree","Lower degree","Lower degree","Lower degree","Lower degree","Lower degree"),
degree_same = c(0,0,1,0,0,0,0,0,0,0),
degree_other = c(0,0,0,0,0,0,1,0,0,0))