SBS output query (Differentiate between 1 and 2 dominant)

[ppdewolf]

Please specify
SDC tool used: tau-argus Version used: 4.2.0 b5 Operating system used: Windows

---

(issue copied from email)

---

We are using TAU-Argus Version 4.2.0 Build 5 to create suppression details for Business output. For primary suppression due to dominance, the area need to differentiate between 1 dominant (B) and 2 dominant (C) enterprises in the cell. We are applying dominance rules (1,80) and (2,90) and also a freq. rule of less than 3.

Because of this requirement, we use SBS output. However, when we create this output, cells we would have expected to display as B's are instead displaying as C's (even if only 1 contributor in the cell). No cells are displaying as B's. (see file OP_DomCheck)

We could use P% rule and work out whether it should be set to B or C from the contribution of the most dominant enterprise but for the moment would prefer to use the (1,80) and (2,90) dominance rules.

I'll attach sample input and the resulting sample output.

Input : (See attached file: Sample_Input.csv) (See attached file: Sample_Input.rda)(See attached file: Geo_GRPB.hrc)

Output : (See attached file: OP_DomCheck.sbs)

Any assistance would be greatly appreciated,

Regards, Karina Dineen issue.zip

[ppdewolf]

Firstly, note that your freq.rule is redundant: the (2,90) rule implies that there should be at least 3 contributors to a cell. Less than three is obviously unsafe (2 contributors contribute 100% to a cell if they are the only contributors, same for a sigleton cell).

Secondly, according to SBS regulation (I think), whenever a cell conflicts with both the frequency rule and the dominance rule, the flag for dominance is to be used. Hence the C-flag for the 1-contributor cell.

Thirdly, I think that whenever both (1,80) and (2,90) rule are violated, the flag for (2,90) is to be used. In case the (1,80) rule is the only rule that is violated, you will see the B-flag.

So, as far as I know:

if ((2,90) rule is violated) { 
    C flag
}
else {
    if ((1,80) rule is violated) {
        B flag
    }
    else {
        if (frequency rule is violated) {
            A flag
        }
    }
}

In the example instance you provided, all unsafe cells violate the (2,90) rule and hence all get the same C-flag.

However, if I change one of the cells to (81 + 5 + 5 + 5 + 4 = 100) then it violates (1,80) but not (2,90) and it still gets the C-flag in the tau-argus output. I'll have to check that!

[KarinaKelleher]

Thank you @ppdewolf for your response. Yes, currently the only time there appears to be a frequency fail now is when the total value is 0 and the number of contributors is less than 3. We are happy to keep these suppressed as just provides consistency for the areas with the rules they have been applying.

In the example given I would also have expected to see a B fail and not a C fail for 'AD' as there is only 1 contributor - as there is no second contributor and therefore cannot fail (2,90) rule - This would have been the logic applied by areas to date - As we are currently introducing TAU-Argus as a potential SDC solution to a number of areas (Business Statistics mainly) we are providing a comparison of the output produced using their current (non TAU-Argus) solution and the TAU-Argus output. We can of course change the TAU-Argus output to Bs where only 1 contributor again to keep consistent feel to output produced but ideally would prefer not having to run too many additional steps after applying TAU-Argus.

Really appreciate the time taken to look at this,

Regards, Karina.