I am currently using the conn toolbox in Matlab. In the 2nd level result, we set the connection threshold at uncorrected p-value < 0.001 and the cluster threshold at uncorrected p-value < 0.05 to confirm the following results.
If you look at the result table as shown below, the values of p-uncorrected and p-FDR are shown.
I have two questions.
1. What does the pFDR value shown in the results table below calculate and mean?
2. Both the connection threshold and the cluster threshold are specified as uncorrected, but the value of pFDR was calculated in the table below; why? What does the pFDR from this mean?
Not my area, but I just want to make sure that you know that if you change the sample size, you change the p-value. With a large enough sample size (where 'large enough' depends on the application and standard deviations) you can 'disprove' anything. With a small enough sample size, nothing is disproven, based on the p-value.
This is why an isolated p-value is virtually useless.
Standard errors (unlike standard deviations) also are reduced with sample size, but they are more practically interpretable. Graphics can often be very helpful. But an isolated p-value with no mention of effect size is not helpful.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA