Meta-analysis

In chapter 2 we introduce meta-analysis, a statistical procedure for examining the results of a large number of similar studies.

A large meta-analytic study was recently released by the World Cancer Research Fund. That study found links between body fat and six types of cancer. They are: cancer of the esophagus, the pancreas, the colon, the womb, the kidney, and the breast (post-menopause).

The study looked some 7,000 published research reports published since the 1960s (out of a possible half-million articles). Dr. Michael Marmot, an epidemiologist, directed the project.

The report recommends:

• ham, bacon, and other processed meats should be eaten rarely
  
• no more than a pound of cooked red meat should be eaten weekly
• drinking no more than a small glass of wine (or equivalent) per day
• mothers should breastfeed for at least six months because that lowers their incidence of breast cancer
• exercising at least 30 minutes per day and working up to 60 minutes per day
  
• minimizing salt use
• drinking water instead of drinks with sugar
• eating fruit, vegetables and fiber

Meta-analytic studies strive to see the "big picture" created by many smaller and narrower empirical studies. To see the full report from the World Cancer Research Fund click here.

Publish the Data

Science depends on planned data collection and its publication. Without publication many of the advantages of science disappear. So, it is shocking that NASA is refusing to publish the results of a major and expensive ($8.5M) telephone survey on airline safety.

The details of this story can be seen in this New York Times piece.

Over 8,000 pilots completed the survey which asked questions about safety issues during flights. Topics included the number of times birds struck airplanes and near-miss collisions (both on the ground and in the air). The survey was ordered in 1997 by a White House panel and was completed in 2005. Fearing that the data might be purged intentionally, Congress ordered that digital copies of the data be preserved.

NASA recently refused a Freedom of Information Act request to release the data from the Associated Press. The Associated Press responded, “Anxious to avoid upsetting air travelers, NASA is withholding results from an unprecedented national survey of pilots that found safety problems like near-collisions and runway interference occur more frequently than the government previously recognized.”

NASA maintained that release of the information would be detrimental to the public, the airlines, and to general aviation. They also said that the study was designed to investigate methods of collecting such data (i.e., a pilot study) not to collect data about safety.

As we point out in chapter 1, science depends on the honest publication of results. Researchers should not refuse to publish data. Many editorial columns are making the same argument. See Kansas City Star and The Boston Herald.

Children and Sleep Loss: Recent Research

This blog entry has several purposes. One is to highlight recent research on sleep loss and its effects. Another is to illustrate some of the research techniques used to document the effects of sleep loss.

A recent article in New York Magazine inspired this entry. Here's the link to that article. The article references survey research conducted by the National Sleep Foundation. Basically, children are sleeping less than they did 30 years ago. Very few (5%) high school seniors get eight hours sleep a night. The average sleep time for them is about six and a half hours. Here's a link to an article about this research.

Avi Sadeh, a researcher at Tel Aviv University, conducted a sleep study where he asked 4th and 6th graders to either sleep a little more or sleep a little less each night for three nights. The children were randomly assigned to either group.

Sadeh was worried that his treatment (which was about a half hour of more sleep or less sleep each night) would not be enough to detect. The dependent variable he used, a test of neurobiological functioning, detected large differences because of sleep loss. The sleepy sixth graders functioned worse than normal fourth graders. See: The effects of sleep restriction and extension on school-age children: What a difference an hour makes. Child Development, 74, 444-455. Sadeh need not have worried about his design. One hour made a big difference.

Much other research supports the relationship between sleep loss and academic problems. In response, some school districts are moving the start of the school day to a later time. Two school districts, one in Minnesota and one in Kentucky reported amazing results. In Minnesota, SAT scores went up. In Kentucky, driving accidents for teens went down.

Response to these data has not been what you might expect. Only a few school districts have started school later. Not many parents have made their children sleep more. The bottom line? We all need more sleep.

Scientific Societies: History and Psychological

The Royal Society

In chapter 12 (p. 384) we discuss the forming of the Royal Society, the first scientific society, founded in 1662. We also mention the American Psychological Association, the Association for Psychological Science, and the seven regional psychological associations.

Scientific societies hold an important role in science. Many publish journals and nearly all hold annual meetings. Scientific societies also help scientists identify themselves professionally and give them a venue for meeting other scientists who are interested in similar topics.

The Royal Society has the longest history, naturally, and this page tells the story of the Society. Before the actual founding of the society, English scientists began meeting in what they called an "invisible college" to discuss the ideas of Francis Bacon. Two of those early scientists were Christopher Wren and Robert Boyle. Robert Hooke and Isaac Newton were two other early members. Their correspondence led Newton to develop his theory of universal gravitation.

In 1679, Hooke, then the secretary of the Royal Society, wrote to Newton inviting him to correspond about scientific topics of mutual interest to them. (See: Cohen, I. B., (1981). Newton's discovery of gravity, Scientific American, 244 (3), 166-179 for a comprehensive account.) A few years earlier, Hooke and Newton had disagreed over Newton's work on optics. So, intense was Hooke's criticism that Newton nearly gave up science.

Their new correspondence was amicable and related to planetary motion and the predicted paths of objects in space as they were influenced by other objects. After a visit by Edmund Halley, the discoverer of Halley's comet, Newton's ideas became firmer. Halley urged Newton to publish his ideas and Newton did, writing De Motu.

Psychological Societies

The American Psychological Association was founded in 1892 in the home of G. Stanley Hall. Fernberger's account tells much about the first 50 years of the association. The Association for Psychological Science was founded in 1988. This link provides a timeline of APS's history.

Information on the regional psychological associations can be found here. As we mention in chapter 12, the seven regionals are: the Eastern Psychological Association (EPA), the Midwestern Psychological Association (MPA), the New England Psychological Association (NEPA), the Rocky Mountain Psychological Association (RMPA), the Southeastern Psychological Association (SEPA), the Southwestern Psychological Association (SWPA), and the Western Psychological Association (WPA).

The seven regionals differ in size. The WPA and MPA are the largest. NEPA is the smallest. The EPA and SEPA are mid-sized and the SWPA and RMPA are slightly smaller. All of the regionals meet annually in cities within their regions. The MPA nearly always meets in Chicago (at the Palmer House Hotel) while the others move from city to city.

Many students, both undergraduate and graduate, attend the meetings of regional and national societies. The meetings provide many learning opportunities for them. Check the links above for a meeting near you and try to attend.

Scientific Facts Do Not Speak for Themselves

In chapter 5 (p. 127) we write, "Although you may have heard that the data speak for themselves, this isn't true. It is the researcher's (and others') explanations that speak for the data."

Recently, Matthew Nisbet and Dietram Scheufele expanded on this thought in an article online. That article, "The facts never speak for themselves, which is why scientists need to "frame" their messages to the public," looks at science and how its methods and results are communicated by scientists and the media to the public.

They define framing as a way to "tailor messages in ways that make them personally relevant and meaningful to different publics." They argue that the older model of science, the popular science model, assumes "that the facts will speak for themselves and will win out, with no attention to how the facts are presented." We don't agree with popular science model either.

Here are some of the frames they identified in their research on science communication:

• morality/ethics
  
• social progress
  
• racing to find a cure
• economic competitiveness
• brain drain
• Pandora's box
• high tech inspired by nature
• asbestos
• Frankenfood

They point out that the media and advocacy groups often frame issues purposively while many scientists do not. They take time to differentiate framing from spin. Scientists should not spin their results; scientific integrity is still paramount.

As a "first step" they suggest that scientists must learn to be better spokespersons for their work. Scientific leaders, especially, must lead the way. They are the ones in positions that can effect positive change the quickest. They also suggest greater public dialog about science. However, they are not optimistic that great numbers of the public will attend such meetings.

They praise, E. O. Wilson's book, An Appeal to Save Life on Earth, because he frames his arguments scientifically, personally, and morally. That multiple framing, they say, has led to religious audiences reading and discussing his book.

They cite reasons why science and its results fare are so poorly communicated. People pay more attention to sources that confirm what they already believe. New and expanding media outlets such as the Internet and television offer the potential for better informing the public, yet paradoxically, they are not used by the public. Instead, only a small minority tune in to content about science. They rest look elsewhere.

They give specific examples of well known and little known instances of science and framing. Among the well known are evolution vs. intelligent design and stem cell research. Less well known (in the United States, but not in Europe) are plant biotechnology and nanotechnology.

As stated in the beginning, we agree with Nisbet and Scheufele. Scientists must be advocates for their data. We would add, that another strategy toward this goal is to teach budding scientists, from the outset, that the data do not speak for themselves.