Bonner (2005) argued for the existence of two scientific methods, which he referred to as ‘‘method A’’ and ‘‘method B.’’ He described method A as the traditional approach, in which hypotheses guide the generation of predictions and the design of experiments to test them; hypotheses are supported or contradicted by comparing predicted results with observed results. In method B, hypotheses are generated only after collecting experimental data and serve to explain the data. In Bonner's view, recognizing that at least two scientific methods exist is relevant to teachers because many students lack the background knowledge needed to generate interesting hypotheses to test –– via method A. However, students can ask questions that can be answered through experimentation –– via method B. Thus, pedagogical progress can still be made.
Marshall Nirenberg's Research Method
In support of the existence and usefulness of method B, Bonner (2005) cited the Nobel Prize––winning research that Marshall Nirenberg conducted in the early 1960s. According to Bonner, Nirenberg's research followed method B and asked this descriptive question: ‘‘What amino acid does UUU code for?’’ At that time, biologists thought that the DNA code consisted of four letters (adenine –– A, guanine –– G, cytocine –– C, and thymine –– T). They also thought that the DNA code was first translated into an RNA code, also with four letters, but with uracil –– U substituting for thymine –– T. Hence, an RNA code consisting of combinations of A's, G's, C's, and U's somehow coded for the production of proteins by somehow stringing the 20 or so amino acids together. According to Bonner's interpretation of Nirenberg's research, there were 20 possible answers to his descriptive question (e.g., UUU codes for valine, UUU codes for serine, UUU codes for phenylalanine, and so on). In Bonner's view, Nirenberg harbored no hypotheses and advanced no predictions about which amino acid would be produced. Thus, the particular answer to his question was theoretically uninteresting. Nirenberg simply wanted to know which of the 20 amino acids UUU coded for. In other words, the fact that it turned out to be phenylalanine was just the way it turned out and was no more or less important than UUU coding for valine, serine, or any of the other possibilities.
Given Bonner's take on Nirenberg's research, it is perhaps surprising to learn how others at the time responded when they learned of Nirenberg's result (as initially recounted in Lawson, 2010). For example, consider this response by Francis Crick, contained in a paper published in Nature (Crick et al., 1961):
At the recent Biochemical Congress in Moscow, the audience of Symposium I was startled by the announcement of Nirenberg that he and Matthaei had produced polyphenylalanine (that is a polypeptide all the residues of which are phenylalanine) by adding polyuridic acid (that is, an RNA the bases of which are all uracil) to a cell-free system which can synthesize proteins. (p. 1232)
One has to wonder why the audience was ‘‘startled’’ by learning that a string of U's codes for phenylalanine and not for, say, valine or serine. Perhaps there is more to the story than Bonner is acknowledging. Also consider Crick's comment in a letter to Nirenberg dated 4 January 1962:
The English papers have made rather a fuss about our Nature paper, which was published on Saturday, but as far as I have stressed that it is your discovery which was the real break-through.
Crick's sentiment that Nirenberg's research represented a breakthrough was echoed in two other letters to Nirenberg (all the letters quoted here are online at http://profiles.nlm.nih.gov/). One letter, from the famous French researcher Francois Jacob, dated 20 December 1961, had this to say: ‘‘Many thanks for your two manuscripts. It is a wonderful story. All my congratulations.’’ The other letter, from H.J. Muller of Indiana University, dated 1 February 1962, stated:
Let me express the thanks and appreciation of the Committee that arranged the recent symposium on RNA coding for your kindness in having come here for the truly remarkable contribution that you have made. It was inspiring to the older and to the younger hearers alike to follow the course of the marvelous break-through that you described to us.
Nirenberg's colleagues were not the only ones startled and impressed by his ‘‘wonderful story,’’ his ‘‘marvelous break-through.’’ The newspapers were also lauding Nirenberg's achievement. Importantly, they placed it in the larger theoretical context of the day. Consider, for example, the following paragraphs written in an article titled ‘‘NIH Researchers Crack the Genetic Code,’’ published in the Medical World News (5 January 1962, p. 18; online at http://profiles.nlm.nih.gov/):
The enigma of genetic coding, considered a fundamental secret of life, may be on the verge of solution. In just-published and about-to-be published papers, several research teams are reporting experimental proof of what has been largely theory: the intricate process by which structure and function of living organisms are shaped. One group has begun to crack the DNA-RNA code –– the key to the whole mystery. Soon they expect to decipher the entire set of instructions by which genetic messengers direct the manufacture of proteins –– the basic stuff of life.
The major achievement in RNA research is the work of two young biochemists at the National Institute of Arthritis and Metabolic Diseases, Drs. Marshall W. Nirenberg and J. Heinrich Matthaei. Behind their work, however, is a whole series of investigations which has produced the basic theory and its preliminary experimental support.
Fundamentally, the theory states that the hereditary ‘‘blueprints’’ of the cell structure and function are coded within the cell nucleus as long-chain molecules of deoxyribonucleic acid (DNA). These plans are transmitted, in a series of steps, to the cytoplasmic ‘‘assembly line’’ where they direct the synthesis of each cell's characteristic products.
If we assume that this is a relatively accurate account, then we can see why Nirenberg's result caused such a fuss. He not only answered Bonner's narrow descriptive question, but he also provided a key piece of evidence in support of a much broader causal question, namely: How does DNA code for the production of proteins? Importantly, by helping answer this more fundamental theoretical question, Nirenberg had begun to ‘‘crack’’ the genetic code –– a breakthrough worthy of a Nobel Prize.
Thus, Bonner's take on Nirenberg's research as purely descriptive and exemplary of method B appears to be misleading. A more accurate interpretation is that Nirenberg was using method A –– the traditional scientific method. Nirenberg's research can be better understood in terms of theory, prediction, and test, and can be summarized something like this:
Causal question: How do the letters of DNA code for the production of proteins?
Basic theory: Specific combinations of at least three of the four letters of DNA code for the production of proteins by first serving as a template for the production of RNA. Specific combinations of at least three of the four letters of RNA then serve as a template for the production of proteins by coding for specific amino acids. Accordingly:
If the above theory is correct (basic theory),
And we conduct an experiment with RNA made only of U's (planned test),
Then a polypeptide molecule should be synthesized and it should consist of only one type of amino acid (predicted result)
And when the planned experiment was conducted by Nirenberg and Matthaei (1961), they found that a polypeptide chain consisting of only one type of amino acid (i.e., phenylalanine) was produced (observed test result).
Therefore support had been found for the basic theory (conclusion).
Of course, additional questions remained. Is the code a triplet code? Is it non-overlapping? Is it degenerate? Nevertheless, the marvelous breakthrough had been made. The basic theory had been supported and the genetic code was starting to crack.1
The Need for Guiding Theories & Hypotheses
So it would seem that in Nirenberg's case, experimentation was guided by theory. In fact, one might wonder how it could be otherwise. Without some theory or hypothesis to test, how is one supposed to know what experiment to conduct? It would seem that one cannot know, a sentiment echoed by the following words of Charles Darwin, Peter Medawar, and Carl Hempel:
How odd it is that anyone should not see that all observation must be for or against some view if it is to be of any service. (Darwin, as quoted in Schick & Vaughn, 1995: p. 191)
Inductive theory provides no formal incentive for making one observation rather than another. Any adequate account of scientific method must include a theory of incentive or special motive. We cannot browse over the field of nature like cows at pasture. (Medawar, 1969: p. 29)
In sum, the maxim that data should be gathered without guidance by antecedent hypotheses about the connections among the facts under study is self-defeating, and is certainly not followed in scientific inquiry. On the contrary, tentative hypotheses are needed to give direction to scientific investigation. (Hempel, 1966: p. 13)
John Platt also expressed this view in his now classic article ‘‘Strong Inference’’ (Platt, 1964), in which he defined strong inference in much the way that Bonner (2005) described method A. According to Platt:
Strong inference consists of applying the following steps to every problem in science, formally and explicitly and regularly:
1. Devising alternative hypotheses;
2. Devising a crucial experiment (or several of them), with alternative possible outcomes. Each of which will, as nearly as possible exclude one or more of the hypotheses;
3. Carrying out the experiment so as to get a clean result;
4. Recycling the procedure, making sub-hypotheses to refine the possibilities that remain; and so on. (p. 347)
Although Platt noted that some research fields collectively embrace these steps, he also noted that the steps were neither universally understood nor consciously applied. Again in his words:
The difference between the average scientist's informal methods and the methods of strong-inference users is somewhat like the difference between a gasoline engine that fires occasionally and one that fires in steady sequence. If our automobile engines were as erratic as our deliberate intellectual efforts, most of us would not get home for supper. (p. 347)
Implications for Biology Teaching
What does this have to do with biology teaching? First, it seems clear that we should provide students with as accurate accounts of scientific method as possible. If two scientific methods exist, students should learn about both. They should try them out and develop skill in their use. On the other hand, if only one method exists, or if one method is more powerful than others, attention should not be diverted to less effective ways of doing science.
If one agrees with Platt (1964), the key educational question becomes ‘‘How can teachers help their students develop the skills and habits of mind of strong inference?’’ Although Bonner (2005) correctly argued that it seems unreasonable to initially press all student-generated investigations into the hypothetico-predictive format of strong inference, we all nevertheless have observed or directed biology labs in which students are blindly following directions and conducting experiments in which they had no real clue as to what they were doing, why they were doing it, or how their investigations fit into some bigger theoretical picture. This means that although students can and should discuss how biologists in the past selected and conducted their investigations (incidentally, not all investigations involve experimentation), contextually relevant activities (e.g., lab and field-based inquiries) are the main vehicles for promoting intellectual development. But such activities cannot be ‘‘cookbook’’ in nature. Instead, teachers should first allow students to explore nature and encounter puzzling (to them) observations. The puzzling observations should then prompt students to generate and then test alternative explanations. But what, as Bonner noted, should teachers do when students lack the needed background knowledge to do so?
The solution to this problem does not lie in having students resort to a less effective method B. Rather, the solution lies in proper topic sequencing, in introductory remarks prior to the initiation of student explorations, and in the use of specific brainstorming and cooperative learning techniques that will help students generate several interesting hypotheses that they can then test (e.g., Lawson, 2000).
In conclusion, we need to help students develop the habits of mind of strong inference, of method A, in which the following questions become the central focus of inquiry instruction:
•• What did you observe during exploration?
•• What is puzzling about your observations?
•• What questions are raised?
•• What is the central causal question?
•• What are some possible explanations (hypotheses)?
•• How might these explanations be tested?
•• What are the expected/predicted results of each explanation and planned test?
•• Following your test, what are your observed results?
•• How do your observed and expected results compare?
•• If observed and expected results do not match, is the mismatch attributable to a faulty hypothesis, a faulty test, a faulty deduction, or some combination? Can you tell? Why or why not?
•• If observed and expected results match, what conclusion should you draw? Have your results eliminated the alternatives? If not, what additional tests are needed?
•• Can you be sure that the match or mismatch of observed and expected results is not attributable to chance? If not, what can you do to reduce the likelihood of drawing an incorrect conclusion?