STEMskiller: Skill Set Map for Mentors of Early Career Researchers
RESEARCH AND SCHOLARSHIP: Academic reading and writing
[Writing specifics for different academic formats] [Research articles] [Valid scientific argumentation]
Justifying inferences, predictions, conclusions
Definition:
Meticulous research design can enhance the ability of a researcher to explain any inferences and/or predictions made or any conclusions drawn from a particular study, thus assisting those evaluating the reliability, validity, generalizability, and limitations—typically, the editors, peer reviewers of scholarly journals, and scientific peers—of the study. In justifying inferences, predictions, and conclusions, the researcher must above all exhibit a thorough understanding of what kind of inferences and/or predictions are employed in a particular study; otherwise, “there will be little basis to build confidence in the [study’s] predictive ability for future realities or in new domains.”[1] Conclusions should, ideally, referentially “tie back” to inferences and/or predictions employed and include commentary about potential limitations in the ability of findings to impact disciplinary theories and/or phenomena of interest to the field. As Sanders states:
It is incumbent upon the...scientist to identify the most effective ways to provide context for and to explain why their [research] model acts the way it does, why they believe its implications, how they’ve made relevant methodological choices and assumptions, and what caveats remain in their implementation or analysis.[2]
Furthermore, scientists should be aware that statistical inference is not equivalent with scientific inference:
Unlike its statistical inference counterpart, the concept of scientific inference defies reduction to a series of allegedly neat and tidy methodological steps whose dutiful observance renders the output ‘science.’ Believing otherwise is wishful thinking. Scientific inference transcends by far the purview of statistical inference.
Scientific inferences in any given area are not made in a vacuum. Rather, they are offered in light of the totality of empirical and theoretical subject-matter knowledge that has accumulated over a long period of time. In progressive sciences this background knowledge, whose cogency rests on conditional acceptance by members of the field, plays a defining role in evaluating the plausibility of new findings. So scientific inferences are made within a dynamic context of what we believe we know, and hope to know, about our world and beyond. Importantly, these scientific conjectures and judgments do not often involve applications of formal statistical inference. They arise instead from researchers employing rules-of-thumb in their daily, and mostly exploratory…work.[3]
[1] Sanders, N. (2019). A Balanced Perspective on Prediction and Inference for Data Science in Industry. HDSR. https://doi.org/10.1162/99608f92.644ef4a4
[2] Ibid. (2019)
[3] Hubbard, R., Haig, B.D., & Parsa, R.A. (2019). The Limited Role of Formal Statistical Inference in Scientific Inference. The American Statistician, 73(sup1), 91-98. https://doi.org/10.1080/00031305.2018.1464947
Useful resources on Justifying inferences, predictions, conclusions:
Hubbard, R., Haig, B.D., & Parsa, R.A. (2019). The Limited Role of Formal Statistical Inference in Scientific Inference. The American Statistician, 73:(sup1), 91-98. https://doi.org/10.1080/00031305.2018.1464947
Discusses the limitations to formal methods of statistical inference, providing a historical contact for why form statistical inference is considered the “standard” in many fields and the difficulty in achieving formality. Provides arguments for scientific inference, describes typical scientific inference that takes place, and calls for reforms in the quantitative research educational process. Note this is a gated resource; please contact your local library if you are unable to access it and to order it via interlibrary loan or a similar service.
Sanders, N. (2019). A Balanced Perspective on Prediction and Inference for Data Science in Industry. HDSR. https://doi.org/10.1162/99608f92.644ef4a4
Defines the terms prediction and inference in relation to data science, noting how they ideally play an interconnected role, and describes why it is important for data scientists to be open with stakeholders about the limitations of models employed.
Tags: IPS IA; IPS QL; IPS PS; IAL IntL; CompQ; CompTS
Peer Review: None
Table of contents:
- 1.2. Academic reading and writing
- 1.2.1. Preparing to write
- 1.2.2. Writing specifics for different academic formats
- 1.2.2.1. Books
- 1.2.2.2. Dissertations
1.2.2.3. Research articles
- 1.2.2.3.1. Introduction to writing for publication
1.2.2.3.2. Valid scientific argumentation
- 1.2.2.3.2.1. Quantitative statistics
1.2.2.3.4. Co-authorship
- 1.2.2.4. Review articles
1.2.3. Grant proposals
1.2.4. Corresponding with others
Author: Stephanie Krueger
Peer Reviewer(s): None
Last Updated: October 22, 2021