The Dead Grandmother/Exam Syndrome

I draw your attention to this research paper from Professor Mike Adams from Eastern Connecticut State University titled The Dead Grandmother/Exam Syndrome, also published in The Annals of Improbable Research. In the paper, Adams investigates the phenomenon he summarizes as follows:

A student's grandmother is far more likely to die suddenly just before the student takes an exam than at any other time of the year.

He takes twenty years of historical data and confirms the existence of the phenomenon, thereby drawing attention to an important but overlooked national health problem: Increased mortality of women with grandchildren in college during the weeks leading up to exams.

  • If there is no exam imminent, the death rate is independent of how well the student is doing in class.
  • As a midterm nears, the death rate goes up by a factor of ten. As a final looms, it goes up by a factor of 19.
  • The effect is strongly dependent on how well the student is doing in class. Grandmothers of students doing poorly are at much greater risk. A grandmother of a failing student is 50 times more likely to die in the week prior to a final than a grandmother of a top student when there is no examination imminent.
  • Grandmothers are 24 times more likely to die than grandfathers.
  • The effect is independent of family size.

Adams develops theories which attempt to explain these phenomena and also has some proposals for addressing the effect.

A follow-up study by professor Lee Jussim of Rutgers University examined ways of addressing this enormous danger posed to grandmothers. In A Preliminary Report on an Intervention Designed to Reduce Grandmother Death Resulting From College Exams, Jussim found a way to save the lives of 4 out of 5 grandmothers during the lead-up to exams: Inform students that the make-up exam will be brutally difficult.

Bonus chatter: The articles are written tongue-in-cheek, but other less whimsical explanations for the observed behavior include

  • If a grandmother passes when no exam is imminent, the student will miss class without explanation.
  • Good students are less likely to ask for assistance with exams even if they lose a grandmother.
  • A student with an ill grandmother is more likely to have poor grades due to stress/worry.

Note also that the predicted grandmother popular collapse did not come to pass. One theory is that this was prevented due to another phenomenon: Grade inflation.

Comments (19)
  1. Gabe says:

    My grandmother actually did die right before finals in my first semester of college, so I have to believe that the research is correct. One can only imagine that worrying about how their grandkids will do on the tests is somehow more than their frail bodies can handle.

  2. Mike says:

    I wonder what the average mortality rate per student is, and whether it is greater or less than 4 :)

  3. SimonRev says:

    @Mike — actually the article addressed that to some degree.  It mentioned that the increase in divorce and remarriage rates in the US was speculated in some degree to compensate for this phenomenon by creating additional (step)grandmothers for each student to allow for > 2 deaths per student.  Note that only grandmothers seem to be subject to death-due-to-grandchild's-exam and not grandfathers.

    The article also discusses that it is good that only grandmothers are subject to this mortality, because otherwise smaller families wouldn't dare let their children go to college because they wouldn't have enough spare family members to sacrifice at exam time.

  4. SteveM says:

    My grandmother also died the week before my first year exams – my parents didn't tell me till afterwards!

  5. Jim says:

    What a great study! I know the world a little better now.

  6. Barbara Callow says:

    I vaguely recall a study or report about students having an excessive number of grandmothers/parents — enough to avoid multiple tests.

  7. Larry says:

    I used to work for a major Unnamed Air Line in their reservations center, holidays are very hard on grandmothers also.  The number of grandmothers that died just before Thanksgiving and Christmas was statistically significant.

  8. RandomStranger says:

    Shouldn't be exams abandoned for the sake of grandmothers then?

    Students would like such studies far more if that was the conclusion.

  9. Silly says:

    To defend students against any perception that they might be lying, here are some explanations that I just pulled out of my… head:

    1. The professors' experiences and studies took place when they lectured children with parents from the Baby Boomer generation. These children (GenX and maybe Y) would generally have a number of aunts and uncles due to the post-war sexy-time frenzy. As a result, the students' grandmothers could easily be at the end of their natural lifespans when the students are of college age. For example, if a baby-boomer's mother had eight children over a number of years then his or her child(ren) would be in college during a number of years +/- the end of the twentieth century. Additionally, women tend to live longer than men on average, so this results in more older student grandmothers than grandfathers. (Maybe Grandpa died just before high school finals.)

    2. Grandmothers are beautiful creatures of nature. Perhaps they sense the stress due to risk of failure in their grandchildren and subconsciously shuffle off their mortal coils in an attempt to assist in the success of their genetic lineage. This is similar the "worry themselves to death" conclusion, except it notes a possible underlying reason: helping to ensure the success of their line.

    3. Overzealous future and/or confused present physics students are killing their peers' grandmothers in an attempt  to observe the result of a spacetime paradox. For some reason (insert technobabble here) this does not eliminate the grandchildren or modify their memories, but instead results in a number of youths being labelled as 'lost causes'.

  10. Where I come from, the student must attend the exam session even if everyone in the universe including himself dies… or face received a zero for score.

  11. tocsa says:

    RandomStranger: "Shouldn't be exams abandoned for the sake of grandmothers then?"

    That is Adams's original paper's first candidate solution. I'm quoting Lee Jussim, he summarizes Adams's findings:

    "He also proposed three possible solutions:

    1. Stop giving exams.  Adams concluded that this would

    prevent colleges from evaluating the competence of people

    interested in becoming doctors, lawyers, engineers, MBAs,

    teachers, etc.  Therefore, colleges would no longer be able to

    train these and many other professionals.  As a result,

    the economy would come to a grinding halt.  This did not

    seem to be a good idea.

    2. Only admit orphans to college.  This would be a good idea except

    for the paucity of orphans.


  12. tocsa says:

    Thanks Raymond, this is a great material for my wife who is a professor. We could spin off another study, because I also noticed, that criminal activities rise too related to exams: cars get burglarized (and either some of the papers of the student gets stolen, and/or the student will be busy taken care of the car, and/or the student is not able to make it to the exam), homes get burglarized, and in general *weird* things start to happen suddenly. Very interesting!

    Also, if a professor wants to teach actually university level material, and demands some knowledge and reading, then the professor's rating deep dives. If the professor helps towards actual exam questions then the rating stabilizes. How to avoid grade inflation then?

  13. Neil says:

    I wonder what that 0xED character in the linked AIR page is supposed to be…

  14. George says:

    Forty years ago, wile teaching at UIUC, I had a student who lost three grandmothers in the course of three semesters. This phenomenon came up in a conversation at lunch among the instructors of different sections of the introductory course. When we called home, we found that they were all doing fine, and we noted that the student was eventually committed to a psychiatric hospital.

  15. Carl D says:

    One of my favorite classes from University oh so many years ago solved this problem right up front: This class had no final, no mid-term, but there was homework AND a test in every single session.  No excuses – you weren't there, you didn't get a score for that test, but with ~50 tests over the run of the class, missing one or two was rarely serious, so legitimate unexpected problems rarely caused any student a problem with their grade.

  16. Jerome says:

    My grandmother also died just before a final-semester exam at college. I went from top student to somewhere around the middle after spending time with her at the hospital and being too tired to study. I wonder how much of this has to do with the probability of older people dying, because they're *old*, while their grandchildren are in their twenties.

    [That would explain the ambient rate, but not why grandmothers are particularly affected by exams, or why grandfathers are 24 times more resilient than grandmothers. -Raymond]
  17. Brian_EE says:

    @Neil: "I wonder what that 0xED character in the linked AIR page is supposed to be…"

    0xED is obviously ED, which is short for Education.

  18. Yuri says:


    When I was in college I hated teachers who rated me on the number of homework I did. The best of 3 method of Raymond's father seems much better for good but lazy students like me. It actually rewards good student while solving the absence problem, just hit 100% on the first exam and let the other students cram while I'm done :D

    I was pretty p*&%$^ off to do a marathon of 50 homeworks instead of 1 exam, usually I skipped the homework and focus only on the exam.

  19. Darcy says:

    @Jerome &


    >Good students are less likely to ask for assistance with exams even if they lose a grandmother.

    >A student with an ill grandmother is more likely to have poor grades due to stress/worry.

    Both these observations (and Jerome's own experience) are suggested in the data already, interestingly.  Note the different no-exam FDRs for the A vs B groups: the latter is nearly twice the former, and both diverge markedly from the average (in different directions).

Comments are closed.

Skip to main content