Ethics, part I: The ethics of research design and data reporting

Ethics is the field of philosophy that investigates the issues of right and wrong behavior. In English usage, ethics is sometimes considered to relate to principles that are universally applicable, whereas morals is considered to describe the relationship between personal behavior and the principles for right and wrong behavior that have been defined by a specific group, culture, or society. However, in common practice, most English writers use the two terms as if they mean the same thing. For the sake of simplicity, I will refer to both concepts as ethics in this article and the follow-up article.

There’s a widespread belief that science is inherently ethically neutral, since knowledge only becomes “good” or “bad” when it is applied in good and bad ways. Although that’s true, most fields of research have developed their own set of ethical principles (a “code of ethics”) to guide both how research is conducted and how researchers communicate the results of that research. That is, these codes help to clarify the principles that define good and bad behavior for researchers in that field. These principles are usually developed based on the current social and scientific contexts for a field, and are revised whenever new discoveries create new ethical contexts with implications that must be carefully considered. If you work in a specific field, you should determine whether a code of ethics exists for your field, and familiarize yourself with that code and how it constrains your behavior. If you find no code, you may be able to find a suitable code of ethics that you can follow from a related field—or perhaps you’ll feel motivated to lead an initiative to develop a code that meets the specific needs of your field.

The purpose of an ethical code is to shift the balance of the consequences of research towards behavior that is predominantly good for the group, culture, or society, and to minimize neutral and negative behavior. But an important and neglected aspect of professional ethics relates to the consequences of your actions for your colleagues. In this article, I’ll discuss two important ethical responsibilities to your colleagues. In part 2 of this article, I’ll discuss a third responsibility: to avoid wasting a reader’s time.

Your first responsibility is to design experiments that produce reliable results that are worth publishing. Many researchers, particularly early in their career, simply apply existing research methods, without carefully considering whether their research context differs enough from the context for those original methods to require a change in approach. Every experiment occurs under certain constraints that differ between contexts: different organisms, different environments, and different prior conditions can all affect the design and performance of research. Before you adopt another researcher’s method, carefully consider whether your unique situation requires a change in their methods to account for your unique constraints. This increases both the accuracy of your data and the likelihood of repaying the time and money you invest in your research. In a world with many alternative uses for money, including feeding, housing, and healing the poorest citizens, that’s an important consideration.

Your second responsibility is to rigorously check your data, no matter how boring this task seems. Any errors in recording your data or in transferring it to (for example) statistical software creates a problem in all subsequent stages of your data analysis and interpretation of the results. As an editor, I have seen (and corrected) many examples of errors in which the author reversed two columns of data in a table or the variables on two axes of a graph, leading to inaccurate and highly misleading conclusions. Asking one or more colleagues to carefully check all your data to confirm that it has been correctly transferred and analyzed is ethical, but it also provides good training for students who have not yet begun independent research careers. Similarly, double-check all calculations—compare your calculations with the same calculations performed by a coauthor, and if there are any discrepancies between the two results, sit down together and identify the source of the error so you can determine which result is correct.

This verification is time-consuming, but it greatly reduces the risk of introducing potentially serious errors into the literature. These errors may affect other researchers for many years. Once results are published, they become a permanent part of the literature, and affect all future researchers who read your paper. The most serious errors may require withdrawal of a paper from the journal (i.e., eliminating it from the journal’s Web site). Although it’s possible to publish errata and response letters, so that the correction becomes part of the literature, there’s no way to ensure that all readers of the original paper will learn about these corrections. Indeed, placing this burden on readers of the literature is unethical and impractical. Imagine, for example, that you wrote a typical journal paper that includes 20 or more literature citations. Nobody has the desire and few people have the time to do a literature search for each of those 20 papers to see whether there have been retractions and corrections or whether critical comment letters have been published. This leads us to the third ethical responsibility: to not waste the reader’s time. That’s sufficiently important that I’ll discuss this in part 2 of this article.

The problem of distributing corrections can be particularly serious for copies of papers that have already been distributed. It’s impossible to find all copies of the paper that had been read before the correction was published or distributed around the world. For example, a paper’s Abstract may appear in hundreds of online databases around the world, and there’s no way to find all of these databases and ask their owners to correct the errors. It’s also not possible to correct all the printed copies of a journal that have been purchased by subscribers and libraries around the world. Finally, many researchers archive printed copies of papers, or PDF copies, so that they will not need to find and download the paper in the future. It’s impossible to find and correct all of these copies of the original paper.

Ethics seems like something abstract that would be of interest primarily to philosophers. It’s not. You can’t eliminate 100% of the problems I described in this article, but you can try to remember that your decisions have real consequences for others, and invest a few minutes of your time to minimize the frequency and consequences of such errors.