And is there a possible way to do this in excel?

I'm looking for outliers in many small (n=4) data sets. I have run

In one example:

Data: 33.700 30.485 29.867 29.921

G= 1.48231, Gcrit= 1.125, t-crit=1.603567

Therefore the 33.7 point is a higher outlier - but that's as far as I've been able to understand. I've confused myself as to whether to use a z-table or t-table, (or something else). I understand P-values conceptually - I just don't know how to get there mathematically.

This online calculator gives the P-value as 0.024 for this data set (higher outlier, one-sided test). I want to understand how they get there.

(https://contchart.com/outliers.aspx)

I've set up my table and formulas for the Grubb's test in Excel - so if there is a way to do the same with the P-value, that would be awesome. But I can enter the data by hand if that's what is required - so long as I can figure out how to generate the data.

Thanks in advance.

I'm looking for outliers in many small (n=4) data sets. I have run

*Grubb's tests to determine whether one value is an outlier (i.e. yes if G>Gcrit), but I am confused about how to generate a P-value from this information.***one-sided**In one example:

Data: 33.700 30.485 29.867 29.921

G= 1.48231, Gcrit= 1.125, t-crit=1.603567

Therefore the 33.7 point is a higher outlier - but that's as far as I've been able to understand. I've confused myself as to whether to use a z-table or t-table, (or something else). I understand P-values conceptually - I just don't know how to get there mathematically.

This online calculator gives the P-value as 0.024 for this data set (higher outlier, one-sided test). I want to understand how they get there.

(https://contchart.com/outliers.aspx)

I've set up my table and formulas for the Grubb's test in Excel - so if there is a way to do the same with the P-value, that would be awesome. But I can enter the data by hand if that's what is required - so long as I can figure out how to generate the data.

Thanks in advance.

Last edited: