| Course | GF582 Statistical Methods for Decision Making |
|---|---|
| Unit | Unit 6 |
| Paper type | Seminar reflection |
| Length | About 810 words, 5 pages double-spaced |
| Format | APA 7 student paper |
| School | Purdue University Global |
| Program | MS in Finance |
| Updated | October 2026 |
Free sample paper for GF582 Unit 6
Winners Repeat, Until the Closed Funds Come Back: A Unit 6 Seminar Reflection on Survivorship
[Student Name]
Purdue University Global
GF582: Statistical Methods for Decision Making
Unit 6 Seminar Reflection
[Instructor Name]
[Date]
The 223 funds and every count are composite and illustrative; the articles cited are real.
What I Believed Going In
I sit on my employer's retirement plan committee as an employee representative, and My notes from the night before the seminar stated the view plainly: the committee should favor funds with strong three-year records, because good managers keep being good. I expected the data to show that most winners stay winners, perhaps three in four.
The Data
Our facilitator gave us 223 composite large-company stock funds. Each was ranked by return over one three-year window and split into a top half and a bottom half. The question was how many top-half funds stayed in the top half over the next three years.
The First Test: Survivors Only
The fund database, like many, listed only funds that still existed at the end of the second window: 200 of the 223. Splitting the survivors into 100 winners and 100 losers in the first window, we found that 59 of the 100 winners were again in the top half in the second window.
My first reaction was satisfaction: 59 percent is better than a coin flip. A one-sided binomial test agreed, barely: if repetition were pure chance, at 50 percent, the probability of 59 or more repeats out of 100 is about 4.4 percent. But 59 percent was far below the three in four I had expected. The first failure was the size of the effect.
The Second Test: Closed Funds Restored
Then the facilitator added back the 23 funds that had closed or merged during the second window, using their returns up to closing and counting them in the bottom half, since funds are usually closed for poor performance. Most had been losers in the first window, but four had been winners.
Ranking all 223 funds in the first window gave 112 winners and 111 losers. Of the 112 winners, 60 stayed in the top half, about 53.6 percent. The probability of 60 or more out of 112 under pure chance is about 25 percent. The persistence was no longer distinguishable from luck.
The survivors-only data made past winners look more reliable than they were, because the funds that disappeared were disproportionately the ones that had stumbled, and removing them quietly removed much of the evidence against persistence.
Why the Closed Funds Mattered So Much
Leah, who works in fund operations, explained how closures happen. Fund companies merge or close funds after poor stretches to tidy up their lineups and remove weak records. A database that drops those funds reports a cleaner history than investors actually experienced. Research on fund databases has shown that survivorship bias raises average reported returns and can create the appearance of persistence where little exists (Brown et al., 1992; Elton et al., 1996).
Another classmate, Ravi, pointed out that the four winners that closed are just as revealing: even some first-window winners failed badly enough in the second window to be shut down.
What the Research Shows More Broadly
Afterward I read Carhart's study of fund performance, which traced most of the repeat performance among stock funds to shared exposure to momentum and to differences in costs, and found the clearest repetition at the bottom of the rankings, where costly funds kept trailing (Carhart, 1997).
How My Thinking Changed
I now think the committee should treat a strong three-year record as weak evidence on its own. Fees and the consistency of a manager's process are more reliable guides. A poor record, by contrast, deserves attention, since bad performance does seem to persist, often because of high costs.
Checking the Numbers Myself
After the session I reran the survivors-only test treating the 23 closed funds as if they had never existed in either window, which is what the original database did. The 59 percent repeat rate depended on that omission. When I instead assumed the closed funds had performed at the median before closing, the repeat rate among first-window winners fell to about 56 percent, still not significant.
A Habit I Will Keep
Before trusting any track record, I will ask what is missing from it. For funds, that means asking for the full history, including closed funds, before comparing a manager with its peers.
What I Told the Committee Chair
I sent our committee chair a short note after the session suggesting that our next fund review ask each provider for performance including merged and closed funds in the same category.
What I Still Want to Know
If persistence among winners is mostly luck, how long a record would the committee need to tell a genuinely skilled manager from a lucky one?
References
Brown, S. J., Goetzmann, W., Ibbotson, R. G., & Ross, S. A. (1992). Survivorship bias in performance studies. The Review of Financial Studies, 5(4), 553-580. https://doi.org/10.1093/rfs/5.4.553
Carhart, M. M. (1997). On persistence in mutual fund performance. The Journal of Finance, 52(1), 57-82. https://doi.org/10.1111/j.1540-6261.1997.tb03808.x
Elton, E. J., Gruber, M. J., & Blake, C. R. (1996). Survivor bias and mutual fund performance. The Review of Financial Studies, 9(4), 1097-1120. https://doi.org/10.1093/rfs/9.4.1097
GF582 Unit 6 instructions, in plain terms
Unit 6 of GF582 pairs a live statistical test in the seminar with a reflection on what it showed. State your belief before the session and why you held it. Describe the data and the test exactly, including how the sample was defined. Report each result with its test statistic or probability, and explain what changed between tests. Record classmates' contributions that explained the result, and connect them to research. Check one result yourself after the session. Say how your thinking changed and what you would do differently in practice. Close on something the session left unresolved. Speak in your own voice, but back each claim with the data or with published studies listed in APA 7, inside the word limit for the seminar.
Inside the GF582 Unit 6 example
The writer's role on a plan committee and a written prediction of three-in-four repetition open the sample. The data section describes 223 funds ranked in two windows. The first test on survivors finds 59 repeats with a binomial probability of about 4.4 percent, and the second, with closed funds restored, finds 60 of 112. The highlighted sentence explains how dropping closed funds removed evidence against persistence. Leah's explanation of fund closures follows, citing Brown, Goetzmann, Ibbotson and Ross and Elton, Gruber and Blake, then Ravi's point about closed winners. Reading from Carhart traces repetition to momentum and costs. A section on the writer's own recheck, a revised view, a habit to keep, a note to the committee chair and an open question close the reflection.
Reading the GF582 Unit 6 grading rubric
GF582 seminar reflections are graded on accurate reporting of the test and a clear account of how the result changed the writer's view. Graders check that the sample definition, counts and probabilities are stated correctly and that the difference between the two tests is explained. The largest share of credit goes to understanding the mechanism, here survivorship bias, and connecting it to research. Independent checking after the session shows initiative. Instructors in the MS in Finance program value practical consequences for the writer's own decisions. A closing question rooted in the session, credit given to each classmate by name, and the survivorship and persistence studies listed in APA 7 finish the marks.
Common GF582 Unit 6 mistakes, and how to avoid them
Reflections in GF582 often report the session's conclusion without the numbers that produced it, which makes the change of view impossible to evaluate. Others treat a barely significant result as strong evidence. The sample definition is left vague, so the role of missing funds goes unexplained. Classmates' contributions are summarized without attribution. Research is cited without connecting it to the data in the room. Practical consequences are missing. If your seminar tested a trading rule, a forecast or a hiring screen instead, keep the same structure: prior belief, data and test, results, mechanism, research, your own check, revised view and an open question.
Get a GF582 Unit 6 example written to your instructions
Whether you attended the live session in Unit 6 or are writing the alternative, send the problem it ran live, the dataset if one was shared, and the rubric. A custom reflection follows a prior view through the session's evidence and comes back within 24-48h, the first free, with your own remarks left bracketed. The paper above is an original model document written by our desk, not a submitted student paper and not an official Purdue University Global document.
More GF582 papers and MS in Finance samples
- GF582 Unit 1: Misleading Default Rate Discussion Post
- GF582 Unit 2: Bond Fund Descriptive Statistics
- GF582 Unit 3: Lending Probability Problem Set
- GF582 Unit 4: Card Portfolio Sampling Exercise
- GF582 Unit 5: Confidence Interval Workup in Two Parts
- GF582 Unit 7: Hypothesis Test Report
- GF582 Unit 8: Regression Output Memo
- GF582 Unit 9: Model Comparison Exercise
- GF582 Unit 10: Statistical Decision Brief
- GF581 Unit 6: Baht Float 1997 Seminar Reflection
- GF570 Unit 6: Norway Pension Fund Seminar Reflection
- GF561 Unit 6: Orange County Loss Seminar Reflection
- GF550 Unit 6: Roth Catch-Up Rule Seminar Reflection
GF582 Unit 6 questions, answered
How should a non-significant result be described in the reflection?
As an absence of evidence at that sample size, not as evidence of absence. Say what the test could and could not detect, and if the session discussed power, note it. The strongest reflections use the non-significant result as a turning point, asking what else in the data or the design might explain it, as the session in this example did.
Can I write the reflection if the live problem used data I had not seen?
Yes; most seminar problems are meant to be seen for the first time in the session. Record what you expected before the result appeared, even if that was only a hunch, and describe how the result and the discussion changed it. If your section provides a recording or slides, citing a specific table or exchange from them helps the reflection read as engaged.
What is survivorship bias in fund data?
The distortion created when a dataset includes only funds still operating at the end of the period. Funds that closed or merged were often poor performers, so dropping them makes the remaining group look better and can hide patterns among the losers. Research databases address it by keeping closed funds, and a careful analysis states whether its data do.
Where can I find a free GF582 Unit 6 sample paper?
A full GF582 Unit 6 seminar reflection on mutual fund persistence and survivorship bias is on this page, with both tests, classmates' explanations, a revised view and APA 7 references.
Why did restoring closed funds weaken the result?
Funds are usually closed after poor performance. Leaving them out removes many cases where past results did not repeat, so the survivors make persistence look stronger than it was.