Can;t get confidence.t to work based on column variables.

[RichXM655]

I;m trying to calculate a confidence interval for my data.

The Data is in a generated table that's already calculated the mean, sample size, and standard deviation for all the samples (there are several thousand rows in this table).

The actual data is all on a 0-1 scale, displayed as a percentage.

Mean	SD	N
83.33%	28.87%	3

So as I understand, the formula, to add a new column for CIs, I just need to run:

CI = CONFIDENCE.T(0.05,'table'[SD],'table'[N])

This doesn;t work however, and throws an error:

An argument of function 'CONFIDENCE.T' has the wrong data type or the result is too large or too small. If the argument is expected to be a date, that date must be between March 1, 1900 and December 31, 9999.

However, I've checked, and re-checked the data types and both SD and N are numbers.

If I replace the column references with static numbers, so 

CI = CONFIDENCE.T(0.05,0.2887,3)

That calculates.

Can anyone point me to what am I missing?

[RichXM655 · Accepted solution]

In the end, I gave up on using confidence.t and instead coded the equation in from scratch, using a lookup table to get the value of T based onthe value for the degrees of freedom.

[mahoneypat]

You are referncing two columns in your formula (even though there is a single row), when a scalar values is expected.  Try this instead

CI = CONFIDENCE.T(0.05,MIN('table'[SD]), MIN('table'[N]))

Pat

[RichXM655]

Sadly, that doesn;t make any difference.

Surley when running a calculation on a column, it shou;d only reference the celss in that row when they're put into a formula- this is how the other column caclulations Ihave work.

I also tried using a CALCULATE fuinction to filtering the column down so it was a single row, then running the calculation, but that didn't work either - so I'm totally stumped at the moment.