Lecture 9 Data Analysis and Interpretation
"The goal is to turn data into information, and information into insight." - Carly Fiorina
9.5. Good data versus bad data
In research, "good data" refers to data that is accurate, reliable, and relevant to the research question being investigated. Good data is typically obtained using rigorous and well-designed methods that minimize the risk of errors or bias. Good data is also typically collected using appropriate instruments and techniques that are properly calibrated and validated.
On the other hand, "bad data" refers to data that is inaccurate, unreliable, or irrelevant to the research question. Bad data can arise from a variety of sources, including measurement errors, sampling bias, or improper instrumentation. Bad data can also be the result of errors in data entry, coding, or transcription.
It's worth noting that not all "bad data" is the result of intentional misconduct or malfeasance. In some cases, researchers may inadvertently collect or analyze data that is flawed or invalid. However, regardless of the cause, bad data can undermine the validity and reliability of research findings and can lead to inaccurate or misleading conclusions.
To minimize the risk of bad data, researchers typically employ a range of strategies, including careful planning and design of research studies, robust data collection and quality control procedures, and rigorous data analysis and interpretation techniques. By ensuring the integrity of the data, researchers can enhance the validity and reliability of their research findings and contribute to the advancement of knowledge in their field.
* * * * * *
Bad data, or data that is inaccurate, unreliable, or irrelevant, can be of limited or no value in research. In fact, bad data can actually be harmful to research by leading to inaccurate conclusions, wasted resources, and a loss of confidence in the research findings.
In some cases, it may be possible to correct or salvage bad data through various methods such as statistical techniques or recalibration of instruments. However, it's important to note that this should only be done if the bad data can be reasonably attributed to a known error or mistake, and not simply because the data does not fit with the researcher's expectations or desired outcomes.
As for whether a student can throw out or ignore bad data, the answer is generally no. Good research practices dictate that all data, both good and bad, should be properly documented and reported. This ensures transparency and accountability in the research process, and helps to prevent errors or bias from going unnoticed.
In some cases, a researcher may be able to exclude certain data points or observations from analysis if they can be shown to be outliers or due to factors beyond the scope of the study. However, this decision should be based on sound scientific reasoning and clearly justified in the research methodology and reporting. In general, it is best practice to acknowledge and address bad data rather than ignoring or discarding it.
* * * * * *
In principle there is no such thing as inherently "good" or "bad" data, and that all data can potentially provide insights into the research question being investigated. In fact, unexpected or anomalous data can sometimes be particularly valuable in helping researchers identify and correct errors or biases in their experimental design or data analysis techniques.
However, it's important to distinguish between data that is simply unexpected or anomalous, and data that is truly flawed or invalid due to errors or bias in the data collection or analysis process. In the latter case, such data can be detrimental to the research process and may lead to incorrect conclusions or wasted resources.
Therefore, while all data should be carefully documented and analyzed, it's important for researchers to critically evaluate the quality and validity of the data they collect, and to take appropriate steps to minimize the risk of flawed or invalid data. This includes implementing rigorous quality control procedures, carefully validating measurement instruments, and thoroughly analyzing and interpreting the data using appropriate statistical and analytical techniques.