Considering that a researcher can test at the <p = .05 (95%), p = .01 (99%), or p = .001 (99.9%) level for statistical significance, how would you use these three levels in relationship to risk to patients in implementing the decisions of your research?
It is asking for your decision. You might consider the following questions:
What is the topic of your research? What would be the effect on patients if there were an alpha error?
To understand how to use the different p-values in relationship to the risk to patients in implementing research decisions, let's first clarify what p-values represent.
A p-value is a measure of evidence against the null hypothesis, which is the assumption that there is no effect or relationship in the population being studied. In research, a p-value less than a predefined threshold (usually 0.05, 0.01, or 0.001) is often considered statistically significant.
The choice of the p-value threshold depends on the level of risk and uncertainty associated with implementing the research decisions. Here's how you could use the three levels of significance in relationship to patient risk:
1. p = .05 (95% level): If you choose this level, you are willing to accept a 5% chance of making a type I error, also called a false positive. A type I error occurs when you conclude that there is a significant effect or relationship when, in reality, there isn't. In terms of patient risk, accepting a higher level like p = .05 means that you have a higher tolerance for false positives, so there is a relatively lower risk of mistakenly implementing decisions that may not actually be beneficial or safe for patients.
2. p = .01 (99% level): By choosing this level, you are lowering the chance of making a type I error to 1%. It means you are less tolerant of false positives and require stronger evidence before accepting that there is a significant effect or relationship in the population. From a patient risk perspective, selecting p = .01 reduces the probability of implementing decisions that might have adverse effects or lead to ineffective treatments.
3. p = .001 (99.9% level): This level is even more stringent in terms of requiring evidence for statistical significance. By choosing p = .001, you are setting a very low threshold for concluding that there is a significant effect or relationship, reducing the chance of type I errors to 0.1%. From a patient risk standpoint, this level indicates that you are very cautious and prioritize minimizing false positives. Implementing decisions backed by p = .001 would ideally come with a lower risk of harm or ineffectiveness to patients.
In summary, the choice of the p-value threshold depends on how much risk you are willing to accept in terms of false positives and the potential impact on patients. Lower p-values (e.g., p = .01 or p = .001) may reduce the risk of implementing decisions that could harm patients, but they also require stronger evidence to support the conclusions.