Writing Analytically, 6th Edition - Rosenwasser, David & Stephen, Jill [118]
Students need to look at the actual numbers. Let’s say I do an experiment using two diff erent stat texts. Text A costs $67 and text B costs $32. I give one class text A and one class text B, and at the end of the semester I find that the class using text A did statistically significantly better than the class using text B. Most students at this point would want to switch to the more expensive text A. However, I can show them an example where the class using text A had an average test grade of 87 and the class with text B had an average test grade of 85 (which can be a statistically significant diff erence): students see the point that even though it is a statistical diff erence, practically speaking it is not worth double the money to improve the class average by only two points.
There is so much written about the advantages and limitations of empirical information that I hardly know where to begin. Briefly, if it is empirical, there is no guesswork or opinion (Skinner said “the organism is always right”—that is, the data are always right). The limitations are that the collection and/or interpretation can be fraught with biases and error. For example, if I want to know if women still feel that there is gender discrimination in the workplace, I do not have to guess or intuit this (my own experiences are highly likely to bias my guesses): I can do a survey. The survey should tell me what women think (whether I like the answer or not). The limitations occur in how I conduct the survey and how I interpret the results. You might remember the controversy over the Hite Report on sexual activities (whom did she sample, and what kind of people answer those kinds of questions, and do they do so honestly?).
Despite the controversy over the problems of relying on empirical data in Psychology, I think that it is the only way to find answers to many fascinating questions about humans. The patterns of data can tell us things that we have no other access to without empirical research. It is critically important for people to be aware of the limitations and problems, but then to go on and collect the data.
—Laura Edelman, Professor of Psychology
EXPERIMENTAL EVIDENCE
Experimental evidence is a form of empirical evidence. Empirical evidence is derived from experience, the result of observation and experiment, as opposed to theory. It is usually associated with the bodily senses; the word empirical means “capable of being observed, available to the senses,” and the word comes from the Greek word for experience.
Experimental evidence is usefully distinguished from other forms of evidence by the careful attention to procedure it requires. Evidence in the sciences is usually recorded in particular predetermined formats, both because of the importance of methodology, and because the primary test of validity in the sciences is that the experiment must be repeatable, so that another experimenter can follow the same procedure and achieve the same results (see Chapter 15, Forms and Formats Across the Curriculum).
The concern with procedure is present throughout writing in the sciences, though, not just in the Methods section of a lab report. Scientific writing constantly begins by asking the question, “How do we know what we think we know?” And, since experiments inevitably take a scientist into the unknown, it then asks, “On the basis of what we know, what else might be true, and how can we find out?” The concern with procedure in scientific writing is ultimately, then, a matter of clearly articulating the means of verifying and explaining what we think we know.
The treatment of evidence in the following example of scientific writing, a review of existing research on a given phenomenon, explores the adequacy