**Edited by the Administrator - This thread is a follow on/split topic from: http://kschildsupportforum.com/kansas-child-support-calculators-and-forms-%28licensedpaid%29/bradley-software-redoux/ **


I agree with both of you on this topic.  I believe the guidelines lack sufficient clarity and guidance on the topic of income levels between the tabled values.  I think this particular item should be brought to the committee to request further clarification in the next issue of the guidelines.  If I recall correctly, there were some comments to this effect in the public surveys and possibly even some discussion amongst the committee members.  Maybe (Guru) this issue should be noted here on the website as a requested change during the next guidelines review.

As the author of the MS Excel spreadsheet/tool found here, I'll say there are definitely a few different ways to interpret the guidelines in this regard, as Dennis has pointed out.  Guru, you are also correct that my method has applied a method most consistent with the way tax liability is referenced in the 1040i document.  But, based on the comments I have read here, I may want to consider updating my method.

Dennis, I agree, the logarithmic method is the most fool proof and accurate way to calculate child support.  I strongly disagree with any method using linear interpolation of a logarithmic function though.  I haven't plotted it myself, but I'm fairly certain this would lead to a calculated child support amount which is higher than the logarithmic function.  The reason I chose to use the method I have is because, as Guru has pointed out, the language in the guidelines do not require rounding at all.  And as a Pro Se myself, I like the ability to refer directly to the tabled values.

This is a great thread.  I can appreciate the opinions and technical level of discussion.

Hi Dennis,

Thanks for bringing that to my attention.  I'll admit I didn't to a lot of research around the state to see how collection fees are applied.  I know how they are applied in district 18, but I'm not sure how other districts do it.  I intentionally added the "flat fee" feature to try to account for scenarios where a percentage wasn't applicable.  I hadn't considered the use of a cap, but I believe I'll do that.

If my tool is to be useful to more people, its quite possible I need to consider additional provisions for enforcement fees.  Maybe I should make a full featured section on enforcement fee to make the tool useful for people like yourself?

I will look into the change you've requested.  I'll be adding more options with regards to indexing the table lookups, or using the  method.  My sheet does not support income past the tabled values, so in order for it apply to more cases, I'll have to include the logarithmic equations anyway.  Some people love options - myself included.  But, others like to keep things simple, so I've tried to keep the tool as simple as possible.

I really appreciate your comments on the spreadsheet.  I created the tool and used it for many years until other parents and even attorneys asked me if I could send it to them.  After that, features were requested to make the tool more versitile.  I'm always up for a challenge, so I've enjoyed the various requests.  If you can think of anything else that might be useful, I'll definitely look into it.

Brian

Dear Brian:
Be very careful if you decide to use the Pelkowski algorithm rather than the Guidelines 12 pages of tables for five reasons.
This is her presentation of the algorithm in page 6 of her Appendix C:
SUPPORT FUNCTIONS FOR A CHILD AGED 12-18
C = Support in dollars per month per child.
I = Combined gross monthly income
^ = Exponentiation
# of   Poverty   Income up to      Poverty Level          Income Above
Child.   Level      Poverty Level      to $15,500          $15,500
1   $1,550      0.2108750156(I)   0.7756344686(I)^0.822704337   2.795182393(I)^0.689838232
2   $1,850      0.1552451982(I)   0.6676036944(I)^0.806101237   2.049742412(I)^0.689838232
3   $2,150     0.1346616608(I)   0.6060937962(I)^0.803958609   1.822812799(I)^0.689838232
4   $2,500     0.111153846(I)   0.5193605794(I)^0.803958609   1.561964695(I)^0.689838232
5   $2,800     0.09753123053(I)   0.46266702(I)    ^0.803958609   1.391460125(I)^0.689838232
6   $3,100     0.0862563277(I)   0.4208984696(I)^0.803958609   1.265842223(I)^0.689838232
For younger children equations multiply these functions by .80 for ages 0-5 and by .92 for ages 6-11.

First, there is a significant digit problem.  Note that (Child = 4, Column 3) and (Child = 5, Column 3) have 9 significant digits while the rest of Column 3 has 10 significant digits.  Similarly, (Child = 5, Column 4) has 8 significant digits while the rest of Column 4 has 10 significant digits.  It is possible that this last entry could have trailing “00” but if that is the case, then you put in the trailing “00.”

Similarly, there could be a trailing “0” for the 2 examples in Column 3but you are supposed to show the trailing “0”s just as she has the low age multiplier as 0.80.

Second, the actual equations for Columns 4 and 5 are technically correct but potentially very misleading.  Exponentials are not communitative except for some rare examples.  Thus 8^2 is 16 but 2^8 is 256.  Her algorithm for Columns 4 and 5 is (Factor A)*(Combined gross monthly income)^(Exponential Factor).  Since ^ has a higher priority than *, the ^ is performed first and then the * so you go right to let rather than  left to right.

For example, for a 3 child family with combined gross monthly income of $6,600.00, the Base Child Support is 0.6060937962*($6,600.00)^0.803858609 or 0.6060937962*$1,175.908 or $712.711.  This value is rounded to $713 which is the table entry for the age 12-18 child for a combined income of $6,600.00 in the 3 child table.

However, if the user starts on the left hand side, the results are 0.6060937962*($6,600.00)^0.803858609 or $4,000.219^ 0.803858609 or $786.260 which would be rounded to $786 but not match the table value of $713.

The good Doctor has done nothing wrong in a technical sense and everything wrong in a practical sense.  Everyone starts the calculations on the left hand side if only to read the equation.  Columns 4 and 5 should have the format of (I)^0.803958609*.6060937962 and Column 3 of (I)^1.0*0.1346616608.  (I use the notation of an underline to show that the digits extend indefinitely.  Thus 3.3 * 3 is  10 rather than 9.9.)

Third, the tables are misleading.  Go back to $6,600 combined gross income example where the Base Child Support is $712.711 directly from the Pelkowski algorithm.  The low age table entry should be 80% of $712.711 or $570.169 or $570.  But the table has the value of $571 and note that 80% of $713 is $570.40 which rounds to $570 rather than the table value of $571.  Doctor Pelkowski’s spreadsheet must be rounding in some strange manner (I guess).  This means that using the Pelkowski algorithm with the 80%/92% rule does NOT generate the Guidelines Tables because of this occasional $1 difference.

Fourth, remember the first point that her table doesn’t have the same significant digits (8 in one case and 9 in two others rather than the normal 10 digits).  Do we really care?  No in general but note that the low and middle age categories multiple the high age category by 80. and 92%.  Going back to my college days, I don’t think that my professors would be overwhelmed by my determination of the mass of carbon in an experiment was 0.80*1.23456789 = .987654312 grams when the experiment used up 20.% of the carbon based on other tests.  The professor might point out that a better answer is .99 grams and call it a day as one of my multipliers was accurate to only 2 digits.  (The difference between accuracy and precision-see Wikipedia:  Accuracy and precision.)

Fifth, maybe I was wrong.  Maybe Doctor Pelkowski had actually meant 80.00000000% so I read her report very carefully and this is the table in her report on page 5 that addresses these two multipliers (I added the last 3 lines):

               Age Group 0-5             Age Group 6-11   
Income Level      2005      2009      2005      2009
Low         76.28%      80.60%      86.05%      92.08%
Middle         80.65%      82.80%      89.05%      93.23%
High         85.09%      87.86%      91.17%      94.84%
Average      80.67%      83.75%      88.76%      93.38%
2008 Guidelines         76.%                86.%
2012 Guidelines Value         80.%                92.%.

The age multiplier is the rounded down three year’s old age difference value.  I think that the correct multiplier should be the averages which are 83.75% for the 0-5 and 93.38% for the 6-11.  However, this shows that these two multipliers could be accurate to 4 significant digits.  Of course what it really shows is that the 2016 Guidelines will increase these 2 multipliers so there is another child support increase built into the system or else declare that there is really no age difference which reduces the table size by 2/3rds and just keep the highest age bracket.

Good luck Brian

I hope everyone's links to the split thread works.  Please let me know if you have any trouble.

Dennis, have you contacted Dr. Pelkowski about the concerns you have with significant digits and rounding?