p-vale Flashcards

5 papers, viewpoints, history, etc. (5 cards)

1
Q

Key takeaways Fisher, 1926

A
  • Wrote about the arrangement of field experiments
  • the concept of randomization as a critical method for controlling bias in field experiments
  • discussed the importance of random assignment of treatments to experimental units to ensure validity
  • the use of replication and blocking to reduce variability and increase the precision of estimates
  • the 0.05 rule came out of this paper, although, slightly misinterpreted and not his intention
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Key takeaways Peter Hall and Ben Selinger, 1986

A
  • social scientists have been suspicious of the 0.05 rule
  • discussed how Fisher had an unclear line of when the 5% principle was appropriate vs. inappropriate
  • discussion: (1) scientists are generally far more stringent than their fellow citizens, (2) the legal system is far more flexible than the traditional scientific one
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Key takeaways Ronald L. Wasserstein and Nicole A. Lazar, 2016

A
  • ASA statement on p-values
  • circle of how it is being used because it is being taught AND it is being taught because it is being used
  • statistic community has been concerned about issues of reproducibility and replicability
  • ASA agreed that the p-value is useful, bit is commonly misused and misinterpreted
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Key takeaways Ronald L Wasserstein, Allen L. Schirm, Nicole A. Lazar, 2019

A
  • the use of the p-value threshold is useful in some contexts, but led to misinterpretations and a lack of scientific conclusions
  • argues we can move beyond this binary approach of significant or not significant
  • alternative strategies such as effect sizes, confidence intervals, context
  • ATOM (Accept uncertainty, be Thoughtful, Open, and Modest
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Key takeaways Elliot Inman, 2019

A
  • p < 0.05 can lead to binary results
  • advocating for alternatives
  • the need to think critically about the broader context of data
How well did you know this?
1
Not at all
2
3
4
5
Perfectly