There are not many scientific publications addressing the issue of subjectivity in geological studies and its impact. In one of them, Baddeley et al. (2004) clearly express it in a few phrases: “expert judgments may be biased by their use of heuristics to guide the formation of their opinions” or “the accepted or prior opinions of existing experts undoubtedly influence the judgments of others, including future experts in training.” Let’s discuss this matter a bit, as sedimentologists, starting from the moment we observe an outcrop.
A sedimentologist arrives at the outcrop, observes it from a distance, and then approaches to examine it a few centimeters away. They make drawings, record sedimentary data, and take notes. Are all these data objective… or not? Well, let’s assume all that collected information was objective and complete. After gathering the data, the sedimentologist decides it’s time to think about the paleoenvironment; some preliminary ideas have inevitably begun to form in their mind.
What would the environment look like when all those sedimentary rocks were just loose particles? —Was it subaerial? Underwater? Marine? Fluvial? What process would move them? Wind, water? This is just the beginning. All the information from the outcrop is recorded in a certain format, but it can be reformatted to a different one when we are back in the ‘office.’ The ‘raw data’ can also be processed through mathematical tools. At this point, the most common situation is that a sedimentologist looks for ‘evidence’ to support a hypothesis they have in mind. The sedimentologist is in the interpretation stage. The key word is: evidence. But what is evidence, really, and how do we determine it? Evidence is constructed from the sum of the following: (1) observations, (2) data recording, (3) data processing, and (4) data interpretation.
(1) Observations: these are merely descriptions we make of all those external signals that our human body receives and processes. The fact that different people may not observe the same features as others can be explained by the differences in the cellular systems that perform the reception and processing of external signals in each of us (for example, light received by the cells in our eyes, transformed into a different type of signal, and then processed by neurons). The most common difference among us is the sharpness and color of our vision. Just variability in this aspect can lead to quite different observations. But we don’t need to compare ourselves in seeking variability of observations of the same outcrop. During my years as a field geologist, I have corrected my own observations many times at certain outcrops when I got closer or simply changed my angle of view. Experience has made me increasingly cautious about recording those observations as immutable.
(2) Data Recording: here we immediately jump into the realm of heuristics. Heuristics are the strategies we follow to solve a problem that derive from past experiences when tackling similar problems. We are at an outcrop because we want to better understand nature; most likely a given aspect of it, for example, the processes and consequences of sediment transport in fluvial systems. What information will we record, and what will we ‘discard’? Yes, discard, because we cannot record all the information that an outcrop possesses, and because someone else has taught us that we should focus on certain information and not on others. We are also taught how to structure and organize our field data. Some sedimentologists choose to record data in a way or from observations that differ from the norm. How can we be sure we are recording the relevant data, and in the right format, that we need to advance in our field of science? Are you sure you do? I’m not.
(3) Data Processing: here heuristics are again fundamental. The sedimentologist might use some statistical tools or models to see if there are patterns in the data set. Those patterns could be radically different depending on the tool or statistical model used. Prediction is a highly desired capability for most sedimentologists, and patterns help define predictive rules. Again, I have found and seen others enthusiastically discussing cycles and hierarchical organization when talking about field data. Almost never about randomness. When we use statistics, we are eager to find patterns. Let’s admit it, we love to think that nature almost always shows us an organized sedimentary record. Could this mindset be a problem for sedimentological studies? I believe so.
(4) Data Interpretation: this is the final stage and perhaps the one where a sedimentologist is most prone to heuristic problems. This is because we feel the ‘pressure’ to reach a conclusion, and if it’s just one and unequivocal, all the better. There is a good variety of biases that arise from the heuristics we use; and these biases lead us to make errors. Table 1 summarizes some examples of heuristic biases with possible occurrences for outcrop studies.
Next time you arrive at an outcrop, no matter how expert you think you are, remember that you will fall into some of them.
Table 1. Examples of heuristic biases in outcrop studies.
References
Baddeley, M.C., Curtis, A., and Wood, R., 2004. An introduction to prior information derived from probabilistic judgments: elicitation of knowledge, cognitive bias, and foraging. Geological Society, London, Special Publications, 239(1), pp.15-27.
