Can Dying People Hang in There for the Holidays?

Most of us have heard stories about a dying loved one who clung to life just long enough to see their next birthday, or special holiday, or their child's wedding. It's a comforting, widely held belief that we have at least some control over what is perhaps the most unpredictable thing of all: death.

The phenomena seems widespread partly because of our psychological bias for selective attention. We tend to notice the times that something remarkable occurred, while subconsciously ignoring all the times that nothing remarkable happened. In psychology, this tendency is sometimes called "remembering the hits and ignoring the misses."

For every story we hear about someone who seems to have defiantly fended off death to make it to his or her 100th birthday, there are thousands of people who die just a few days short of that century mark. But their stories are not as interesting, and therefore less likely to be reported and remembered. (There is also a darker implication to this belief, suggesting that people who died before a special event didn't have enough will to live or somehow didn't try hard enough to live longer.)

But can dying people really will themselves to live long enough to experience a birthday, holiday, or special event?

A few scientific studies have suggested this, such as those of sociologist David Phillips at the University of California at San Diego. Phillips found that death rates among Jewish men declined just before Passover, and that pattern also held for elderly Chinese women and the Harvest Moon Festival.

The studies, however, have several drawbacks, including small sample sizes (which makes the results difficult to generalize to the larger population), and protocols that are very difficult to control. It's not clear, for example, that all Jewish men saw Passover as a significant life event. Even religious holidays may not have deep personal significance to all members of a given group; while many people look forward to Christmas as a special time of the year, others view it as a period of stress, depression, and anxiety. (One could argue that some people might prefer to die before Christmas instead of hanging on long enough to experience it.)

A given individual may in fact look forward to any number of things not tied to holidays, including a friend’s birthday, the end of an incompetent president's term, or the season finale of "Grey's Anatomy." However, since these factors would be personally specific (and knowable only to researchers through individualized questioning), large sample sizes for such a study are impractical.

The best research to date was a 2004 study that analyzed the records of 1.3 million people. Researchers, led by Ohio State University Cancer Center’s Donn Young, specifically examined the death rates of 300,000 cancer patients.

The study, published in the Dec. 22, 2004, Journal of the American Medical Association, found that there was no overall increase in mortality after Christmas, as would be expected if patients were able to "hold on" through the holidays. Nor was there a correlation between the date of death and Thanksgiving or the patients' birthdays. There may be a "will to live" effect, but it has yet to show up in careful scientific studies.

Benjamin Radford is a writer, filmmaker, and LiveScience's Bad Science columnist. His books can be found on his website.

Benjamin Radford
Live Science Contributor
Benjamin Radford is the Bad Science columnist for Live Science. He covers pseudoscience, psychology, urban legends and the science behind "unexplained" or mysterious phenomenon. Ben has a master's degree in education and a bachelor's degree in psychology. He is deputy editor of Skeptical Inquirer science magazine and has written, edited or contributed to more than 20 books, including "Scientific Paranormal Investigation: How to Solve Unexplained Mysteries," "Tracking the Chupacabra: The Vampire Beast in Fact, Fiction, and Folklore" and “Investigating Ghosts: The Scientific Search for Spirits,” out in fall 2017. His website is