I. Why Focus on Delayed Graduation Rates

Choosing an advisor is essentially a decision made under information asymmetry.

For most students applying for a Professional Master of Finance (MF), the objective information available before officially confirming an advisor is extremely limited: official website profiles are stuck on research directions and academic achievements; word-of-mouth gathered through chat apps varies from person to person; and posts on Zhihu or forums are scattered and difficult to verify. Students face a structural information black box—even if they are rational enough, it is difficult to form a systematic basis for reference.

It is against this background that the author attempts to introduce a quantifiable reference dimension: the delayed graduation rate.

The so-called delayed graduation rate refers to the proportion of students under a certain advisor whose degree conferral date is significantly later than that of their peers in the same cohort. It is not a metric for measuring an advisor's academic level, nor does it directly reflect an advisor's mentoring style. However, the act of delaying graduation itself is recorded—recorded in the defense and degree conferral data publicly disclosed by the university every year, which can be organized and compared. In an environment lacking other objective data, it at least provides a factual dimension that can be verified.

The causes of delayed graduation are complex. The difficulty of the thesis topic, the student's personal state, active postponement due to employment pressure, and even administrative time differences in defense arrangements can all manifest as "delayed graduation" in statistics. The author has no intention of simply attributing the delayed graduation rate to the advisor, nor of using this to make moral judgments about anyone; it is left to the readers to judge for themselves.

II. Methodology—How the Delayed Graduation Rate is Calculated

The data used in this article comes from the degree conferral records publicly released by Fudan University. The author filtered out the defense records for the Professional Master of Finance degree, resulting in 1,410 valid data points after cleaning, covering 96 advisors and spanning from the class of 2017 to the present. The aforementioned data can be publicly obtained through official university channels and possesses basic verifiability.

Before formally calculating the delayed graduation rate, it is first necessary to define the operational definition of "delayed graduation." In reality, there are batch differences in degree conferral dates due to administrative processes. Within the same cohort, it is normal for those who defend earlier and those who defend later to have a difference of several weeks in their conferral dates. If a fixed calendar date is set as the "expected graduation time," it is easy to misidentify administrative approval time differences as delayed graduation, resulting in a large number of false positives.

To circumvent this issue, the author uses the median degree conferral date of the same cohort as the benchmark, rather than setting a fixed deadline universal across cohorts. There are two considerations behind this design: first, the graduation rhythm of different cohorts varies—for example, if a certain cohort generally postpones defenses due to curriculum adjustments, this is a systemic shift unrelated to the advisor's guidance; replacing an absolute date with a relative median effectively filters out such cohort-level interference. Second, the median is naturally resistant to extreme values; a few abnormal cases (such as suspension of schooling or special approval for early graduation) will not skew the benchmark.

Based on this benchmark, the author set the determination interval for delayed graduation: a degree conferral date more than 60 days but less than 300 days later than the cohort median.

The lower limit of 60 days is intended to exclude normal approval time differences—even if there are different defense batches within the same cohort, the gap usually does not exceed two months. Only exceeding 60 days is more likely to reflect substantive delays such as prolonged thesis revision cycles or postponed defenses. The upper limit of 300 days is used to filter out extreme situations, such as suspension of schooling, long-term sick leave, or active applications for extension, which are special circumstances relatively distant from the advisor-student relationship. For records exceeding 300 days, the complexity of their causes exceeds the scope of this article's analysis and is therefore not included in the delayed graduation statistics.

Ultimately, the formula for calculating the delayed graduation rate is:

Delayed Graduation Rate = Number of Delayed Students / Total Number of Students under the Advisor

Among them, the number of delayed students refers to the number of students whose degree conferral date falls within the interval of "more than 60 days but less than 300 days later than the cohort median"; the total number of students is the total number of Professional Master of Finance students who were supervised by and completed their degree conferral under that advisor within the statistical timeframe.

It should be pointed out that the aforementioned thresholds (60 days and 300 days) do not come from any official standard but are subjective settings made by the author based on data distribution and practical scenarios. Different threshold selections will affect the classification of some students at the boundaries, thereby changing the delayed graduation rate values for individual advisors. The author will further discuss the impact of this subjectivity on the reliability of the conclusions in Chapter IV. The transparency of the methodology is a basic principle of this article—if readers disagree with the threshold settings, they can maintain a moderate reservation about the conclusions accordingly.

III. Data Results

This chapter presents the delayed graduation rate calculation results for all 96 advisors, focusing on two groups: advisors with a sufficient sample size and a zero delayed graduation rate, and advisors with a sufficient sample size but a relatively high delayed graduation rate. The author consistently uses a two-dimensional framework of "sample size × delayed graduation rate" to evaluate the reliability of the data—a simple percentage figure has very limited statistical significance if there are only three or five samples behind it; only when the sample size reaches a certain scale does the delayed graduation rate provide a stable reference value.

High-Sample Advisors with Zero Delayed Graduation Rate

Among the 96 advisors, a total of 22 have a zero delayed graduation rate with a sample size of no less than 8 people. The full list is as follows, sorted by sample size from high to low:

It should be noted that a zero delayed graduation rate does not necessarily equate to the best quality of guidance. Factors affecting whether a student can graduate on time include the difficulty and cycle of the research topic, the student's personal academic foundation and psychological state, the pull of the external job market on the graduation rhythm, and even policy-based delays in specific years. These variables are all outside the direct control of the advisor. A zero delayed graduation rate is a positive signal worth noting, but it should not be used as the sole basis for evaluating an advisor.

Advisors with Higher Delayed Graduation Rates

Among advisors with a sample size of no less than 8 people, some have delayed graduation rates significantly higher than the overall level. The following are listed from high to low delayed graduation rates:

Similarly, both the sample size and the delayed graduation rate dimensions need to be considered simultaneously. In cases where the sample size is small, one more or one fewer delayed student will cause a large change in the percentage. Yang Changjiang (21 people, 23.81%) and Lan Xiaohuan (17 people, 17.65%) have relatively larger sample sizes and slightly higher data stability, but the absolute number of delayed students is only between 3 and 5.

Summary of Overall Distribution

Looking at the data for all 96 advisors, there are about 60 advisors with a sample size of 8 or more, of whom about one-third maintain a zero delayed graduation record. The overall median delayed graduation rate is roughly in the 5% to 10% range. There are 6 advisors with a delayed graduation rate exceeding 20% (sample size ≥ 8), which is a small proportion, but the sample sizes of some of these advisors have reached a certain scale, and the data is worth continued attention.

IV. Conclusion

There are several limitations to the analysis in this article, which must be clarified one by one before use.

First, the data can only reflect results and cannot trace causes. A student's delayed graduation may be due to the advisor's pace, the student's own planning adjustments, external employment timing, the difficulty of the thesis research itself, or purely administrative process differences. An advisor with a high delayed graduation rate may not necessarily have problems with their guidance method; an advisor with a low delayed graduation rate may not necessarily mean low pressure within the group. The data is unable to distinguish between these causes, and users must maintain this awareness themselves.

Second, there is a possibility of systemic fluctuations in different years. Cyclical changes in the job market and policy adjustments in the university's graduation process can cause the overall delayed graduation rate of certain cohorts to be higher or lower. Such macro fluctuations are beyond the control of individual advisors but will leave traces in the samples. When making horizontal comparisons across cohorts, this background interference cannot be ignored.

Third, the sample size for some advisors is too small to support stable conclusions. For advisors who have only recently started taking students or whose annual enrollment is limited, their delayed graduation rate may fluctuate significantly due to the special circumstances of one or two students. The author has already marked the sample sizes in Chapter III, but it must be emphasized again: for advisors with fewer than 8 students, their delayed graduation rate figures are for reference only and should not be used as a primary basis for judgment.

Based on the above limitations, the author suggests regarding the delayed graduation rate as one information dimension in the decision-making process of choosing an advisor, rather than a sufficient condition. A reasonable way to use it is to comprehensively compare the delayed graduation rate with research direction fit, word-of-mouth within the advisor's group, the advisor's recent research dynamics, and personal academic and career planning to form a relatively complete judgment. The delayed graduation rate is not the sole basis for choosing an advisor, but given the extreme lack of information, having one more verifiable reference dimension is always better than having to make a choice based only on rumors.

Appendix: Full Data

This article was generated with the assistance of Writer skills and edited/reviewed by humans.