Statistics: Then and Now

I must in a retrospective and nostalgic mood. I remember how I used to have to compute statistics back in graduate school. During my first year, I recall when the statistics class first walked into the department's calculator room. Imagine a room full of calculators that looked like the one below.

This is a Monroe brand rotary calculator. I doubt that I could use one today, but back then I learned to be pretty good at it. The calculations had to be recorded on large pieces of paper. I still have some of those calculations in my files. (But then, I also still have my first grade report cards too. I'm a bit of a pack rat.)

On the first day in the calculator room, we were all warned not to divide by zero. Recall that dividing any number by zero yields infinity. If a Monroe calculator was set to divide by zero, it would continue to chug until someone pulled the plug.

Before I went to graduate school, I invested in my own calculator, a Texas Instruments SR-10 model. It cost me $149, plus tax. The beauty of the SR-10, was that it could calculate square roots. Interestingly, I had to return my first SR-10; it had a missing decimal point in the display between the 1,000s and 10,000s. My roommate, an engineer, and I discovered the faulty decimal display when I could not figure out why my homework problem kept coming out wrong. Here is a picture of the SR-10. Note the square root button. The display was red LEDs. Here's a link to more information on the SR-10.

My roommate, the engineer, was an early adopter of a more expensive early calculator, the HP-35. He paid nearly $400 for it. The HP calculators used a different method for accepting numerical input: RPN or reverse Polish notation. I never liked having to use that system because it required me to learn a new way to think about math problems. Here is a link to a page on the HP-35.

The Texas Instrument calculators, on the other hand, used familiar algebraic rules for inputting data. I'm sure that alone led to many sales compared to the HP family.

Today, these calculator are only seen in museums. Statistical calculators, too, appeared and are still available today, but at much cheaper prices. Most statistical calculations now, however, are made on specific computer applications. SPSS, SAS, R, and Statistica are commonly used statistical computer applications.

It's interesting to me how quickly students (and faculty) adapt to new technologies. However, it seems that students are remarkably blasé about past technologies. So, I guess part of my job is to remind them.

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.