National Catastrophe Restoration, Inc. (NCRI) appeals from an order of the Kansas Court of Tax Appeals (COTA) affirming an assessment issued for unpaid retailers’ sales taxes due for the period May 1,2003, to December 31,2004. NCRI contends that COTA erred in presuming the assessment valid, in disregarding undisputed evidence that the sampling technique employed in the audit was flawed, and in rejecting three stipulations after the hearing was closed. NCRI also contends the assessment amounted to an unconstitutional assessment of sales taxes on out-of-state sales. We reject the constitutional challenge and affirm COTA’s affirmation of the assessment except to the extent that stipulations requiring adjustments to the error rate must be honored in part and reconsidered in part.
Factual and Procedural Background
NCRI is a retailer of goods and services sold to repair and restore Kansas and out-of-state property damaged by natural and man-made disasters. NCRI is based in Wichita and registered with the Kansas Department of Revenue (KDOR) as a retailer. NCRI’s Kansas sales and purchases are subject to tax assessment under the Kansas retailers’ sales tax act and the Kansas compensating use tax. The record reflects that NCRI filed sales tax returns from May 2003 to December 2004 certifying that no tax was owed.
In November 2004, KDOR initiated an audit of NCRI’s financial records for sales and use taxes for the period Januaiy 1, 2002, to December 31,2004. NCRI, through attorney Jerry Capps of Allen, Gibbs & Houlik (an accounting firm), responded to a pre-audit questionnaire; Capps was designated as NCRI’s representative for purposes of the audit. NCRI had no explanation for its failure to remit taxes. Prior to KDOR’s field audit work, however, NCRI voluntarily reviewed its books and records and admitted there were numerous untaxed or undertaxed transactions.
*192In Februaxy 2005, KDOR’s auditor met with Capps and another NCRI representative who provided the auditor with access to NCRI’s records. Because of the voluminous records provided, the auditor and NCRI agreed that using a sampling method would be the best process for an audit. Upon reviewing the records, the auditor determined that a 3-month block sample would work best. Capps advised the auditor that NCRI did not have seasonal variations, therefore any 3 months chosen would be representative of NCRFs business. The auditor then randomly chose the months of November 2002, May 2003, and February 2004 to serve as the sample months for the audit.
In March 2005, Capps executed a formal notice from KDOR that a sample audit would be performed to determine whether NCRI had remitted all sales taxes due on taxable Kansas sales transactions. The audit was to cover the period of January 1, 2002, through December 31, 2004. NCRI was notified of the sampling methodology that would entail a sample base and population of all NCRFs gross sales and all tax-free sales in the sample months. Based on dais population, the results were to be determined as follows:
“The total dollar value of the error revealed by the sample will be divided by the total dollar value of the sales in the three months transactions in the sample to obtain a percent of error. This percent of error will be multiplied by audited sales in each reporting period within the audit period that is not included in the sample to determine the additional amounts subject to sales tax. For the sample months, actual error amounts will be assessed.” (Emphasis added.)
NCRI never objected to the sampling methodology announced by KDOR during the audit procedure or at any time prior to the COTA hearing.
Between November 2004 and July 2006, KDOR performed tire sales tax audit. The business records provided by NCRI for the audit period were voluminous. NCRI admitted the business records were not complete because it did not have a good bookkeeping system. Job files contained reports in which sales figures did not match invoice amounts for the materials sold and worked performed. In some instances, sale proceeds received by NCRI were overreported and some were underreported to KDOR. NCRI *193treated some invoice amounts as estimates and later adjusted the figures. Moreover, sales report entries did not actually reflect the final amount of the invoice, and there was no documentation tracing adjustments made. In addition, NCRFs income statements compiled from sales reports did not match the income reported on income tax returns. Some of NCRFs job files and sales reports were missing, and its sales reports and cash receipts did not match.
In January 2007, KDOR sent notices of final assessment to NCRI reflecting assessments of $236,669 in retailers’ sales tax (plus interest and penalties) and $4,567 in compensating use tax (plus interest and penalties). NCRI requested an informal conference with KDOR regarding the assessments. The use tax portion of the assessments has not been disputed and is not before us in this appeal.
On October 31, 2007, after failing to receive a written determination from KDOR, NCRI filed an application for review of the final sales tax assessment with the Board of Tax Appeals (BOTA). NCRI asserted the assessment was invalid, because: (1) KDOR used November 2002—which was outside the statute of limitations—as one of the “sample” months for calculating the assessment; (2) KDOR did not credit sales taxes actually paid during the sample months; (3) KDOR did not hold a personal or telephone conference with NCRI; and (4) KDOR did not identify the KDOR employee, the employee’s identification number, and the employee’s contact phone number as required by statute.
BOTA successor COTA conducted pretrial hearings, entered a final pretrial order, and then conducted a final evidentiary hearing August 25, 2009. In a July 2011 order, COTA “reluctantly” granted NCRI’s request to amend the pretrial order to raise a new theory suggesting that out-of-state transactions had been taxed. COTA also found that it was NCRFs burden to rebut the presumption that the assessment and/or Sampling method was valid. COTA indicated that it was unclear whether NCRI’s sole witness, Sarah Guzman, was a. fact witness or an expert witness; in any event, COTA expressly found “there is no indication that she provided any specialized knowledge or expertise that could assist [COTA] in determining the validity and reliability of the sampling technique *194used in the subject audit.” Although Guzman based her testimony on review of the audit documents, COTA concluded she had “contrived her own method of determining [NCRI’s] sales tax liability” despite NCRI’s failure to introduce factual evidence to support an alternative method of assessment. COTA specifically found Guzman’s testimony lacldng “foundation and credibility.” Accordingly, COTA found that NCRI failed to rebut the presumption of validity of the assessment and held that a taxpayer could not rebut the presumption “merely by proffering, in the eleventh hour, a new hypothesis based on self-serving interpretations of its own inadequate business records.”
COTA also rejected NCRI’s constitutional challenges. COTA expressed uncertainty regarding whether NCRI was challenging KDOR’s authority to tax transactions involving materials inventoried in Kansas but sold to out-of-state customers or was asserting KDOR taxed out-of-state transactions. Both arguments were rejected. Finally, COTA found that three stipulations pertaining to a need to adjust tire error rate used to calculate the tax deficiency were ineffectual “because they were entered into based on mistaken assumptions and would result in unintended consequences.”
Both NCRI and KDOR filed timely petitions for reconsideration. NCRI argued that it was entitled to directed judgment, that any presumption of validity of the assessment was rebutted, that KDOR was obligated to abate any sales tax assessment taxing out-of-state sales, and COTA erred in ignoring tire parties’ stipulations. KDOR requested reconsideration of COTA’s decision to allow NCRI to amend the pretrial order at tire beginning of the hearing and requested clarification of the order to cite tire current version of an applicable regulation. On August 16, 2011, COTA issued an order amending the applicable citation and otherwise denying the motions for reconsideration.
NCRI filed a timely petition for judicial review with this court.
Did COTA Err in Failing to Rule on NCRI’s Motion for Directed Judgment and for Presuming KDOR’s Sales Tax Assessment Valid?
On appeal, NCRI first contends that COTA erred in denying its *195motion for directed judgment after KDOR presented the audit without witnesses and rested its case solely on doe audit documents. NCRI argues that COTA never ruled on its motion and then erred in its final order by presuming KDOR’s assessment valid when no such presumption is authorized by Kansas law. We conclude the arguments to be procedurally unusual and legally erroneous.
The Motion for Directed Judgment Was Procedurally Unusual and Need Not Have Been Ruled Upon Immediately
For reasons unclear from the record, the parties agreed that KDOR would proceed first with presentation of evidence to COTA, even though NCRI agreed that it bore the burden of proof as to the validity of the assessment. KDOR then proceeded to introduce its Exhibit 1, the notice of final assessment, complete with details of unreported sales/purchases subject to tax by jurisdiction, tax by issue, calculation of error rate for the sample months, and all auditors’ comments and logs. COTA admitted the exhibit, and then KDOR rested. NCRI’s motion for “directed judgment” followed.
We find NCRI’s motion to be procedurally unusual and legally erroneous for several reasons. First, given NCRI’s express admission that it bore the burden of proof “as to the validity of the assessment issue,” we do not understand how an “agreement” for KDOR to proceed out of order could justify a ruling for NCRI as a matter of law prior to evidence from the party with the burden of proof. See K.S.A. 60-250(a)(l) (motion for judgment as a matter of law not ripe until party has been fully heard on an issue). Second, on the record at the opening of COTA’s hearing, KDOR’s notice of final assessment was expressly stipulated to as admissible by NCRI. Third, the exhibit reflects “detail[s] of unreported sales/ purchases subject to Retailers’ Sales Tax for the period May 01, 2003 to December 31, 2004, inclusive.” Thus, to tire extent KDOR had any initial burden in the matter, the admission of the exhibit satisfied any such initial burden. Fourth, COTA was within its legal prerogative to postpone a ruling on the issue. See K.S.A. 60-250(b) (when motion is not granted, court is deemed to have submitted the action subject to a later determination of tire legal questions *196raised). Fifth, we do not find in the record any renewal of the motion at any time prior to the motion for reconsideration by NCRI, but COTA essentially ruled against the motion in its post-hearing order, where it held that KDOR was entitled to a presumption of validity. And finally, the principal evidentiary issue now argued by NCRI, specifically whether tire transactions in question were retail sales to a final user or consumer (citing incorrectly to K.A.R. 92-16-61a, but presumably intending K.A.R. 92-19-61a), was never preserved in COTA’s pretrial order and therefore not subject to adjudication. See K.S.A. 60-216(e) (implicitly controlling COTA procedure under K.A.R. 94-5-1 [2010 Supp.]); McCain Foods USA, Inc. v. Central Processors, Inc., 275 Kan. 1, Syl. ¶ 8, 61 P.3d 68 (2002) (issue not contained in pretrial order should not be considered).
For any of these reasons, and particularly in light of their cumulative effect, we conclude there was no error by COTA in failing to rule on NCRI’s motion for directed judgment prior to NCRI’s presentation of its case.
KDOR Was Entitled to a Presumption of Assessment Validity
NCRI contends that COTA erred in holding that KDOR’s assessment was entitled to a presumption of validity because there is no statutory basis for any such presumption. We disagree. COTA’s rationale and ruling on this issue were as follows:
“The Department’s sampling technique, as well as its assessment, are presumed valid, and it is Taxpayer’s burden to rebut those presumptions. See K.S.A. 79-3610 (placing burden of proof on Taxpayer to demonstrate invalidity of sampling method); see also K.A.R. 92-19-61 (placing burden on Taxpayer to prove the invalidity of the assessment).
[[Image here]]
“As noted above, a presumption of validity attaches to die Department’s audit assessment, and it is Taxpayer’s burden to rebut that presumption. The outcome of any fact-driven proceeding such as the instant case is determined by die ability of die party bearing the burden of proof to persuasively carry that burden. See In re Marriage of Grippin, 39 Kan. App. 2d 1029, 186 P.3d 852 (2008). ‘Burden of proof means preponderance of the evidence, that is, die greater weight of evidence, in view of all facts and circumstances of the case. Matter of Estate of Robinson, 236 Kan. 431, 690 P.2d 1383 (1984). The burden of proof on any point *197is upon tire party asserting it, and it is incumbent upon that party to prove its allegations by a preponderance of the evidence. [Citation omitted.]”
NCRI concedes that a KDOR income tax assessment is entitled to such a presumption, citing In re Tax Appeal of Broce Constr. Co., 27 Kan. App. 2d 967, 980, 9 P.3d 1281, rev. denied 270 Kan. 898 (2000), where the court held:
“Our Supreme Court has long held that ‘the tax found by the tax commission to be due is presumed to be valid [and] the taxpayer has the burden of showing its invalidity.’ Nutrena Mills, Inc. v. State Tax Comm., 150 Kan. 68, 75, 91 P.2d (1939) (citing United States v. Reitmeyer, 11 F.2d 648 [E.D. La. 1926]). Similarly, Crawford noted:
“ ‘The order of the [Department] carries with it a presumption of validity which continues until [the taxpayer] establishes that the method of allocation adopted is inherently arbitrary, or that its application to [the taxpayer] produces an unreasonable result .... The burden was upon [the taxpayer] to show . . . that the factor formula method ordered by tire [Department] to be applied, does not ‘clearly reflect’ a just and equitable allocation of net income to Kansas. 180 Kan. at 264.’ ”
NCRI argues that this clearly established presumption from the income tax context cannot be transported to sales and use tax context. We disagree. The auditing, assessment, and review process under K.S.A. 79-3610 specifically incorporates by reference the auditing, assessment, and review statute under the Kansas Income Tax Code. See K.S.A. 2011 Supp. 79-3226. Neither statute specifically states that the final determination of the Director of Taxation is entitled to a presumption of validity, but—as conceded by NCRI—Kansas has long recognized such a presumption. See Becker v. City of Wichita, 231 Kan. 322, 324-25, 644 P.2d 436 (1982) (presumption of validity which attaches to the determination of tire governing body regarding special assessments for sidewalks and public improvements); In re Burrell, 22 Kan. App. 2d 109, 112, 912 P.2d 187 (1996) (rebuttable presumption of validity attaches to KDOR’s determination Burrell owed drug stamp tax), rev. denied 260 Kan. 993.
Moreover, the Kansas Legislature has explicitly removed the presumption of validity as to some initial tax decisions. See K.S.A. 2011 Supp. 79-1609 (no presumption of validity of county appraiser’s determination of residential or commercial real estate val*198uations); K.S.A. 2011 Supp. 79-2005 (no presumption of validity for county appraisals of real estate in tax protests filed with COTA). Presumably, if the legislature had intended to remove the long-recognized presumptions regarding sales or income taxes, it simply could have used the same language used elsewhere.
Finally, other jurisdictions recognize the presumption of validity of an agency’s assessment of sales taxes. See, e.g., LZM, Inc. v. Virginia Dept. of Taxation, 269 Va. 105, 109, 606 S.E.2d 797 (2005) (applying presumption of validity to an assessment of sales tax and stating that “the burden is on the taxpayer to show that such assessment was the result of ‘manifest error’ or in ‘total disregard of controlling evidence’ ”); see also HKD Lo, Inc. v. Commissioner of Revenue, No. 8028 R, 2011 WL 1049196, at *4 (Minn. Tax Div. 2011) (unpublished opinion) (orders of Commissioner under Minnesota Sales and Use Tax presumed valid and correct).
Thus, we hold that KDOR’s sales tax assessment against NCRI, as reflected in the final notice of assessment, was entitled to a presumption of validity under K.S.A. 79-3610 and COTA neither erred in failing to immediately rule on NCRI’s motion for judgment nor in presuming the validity of the assessment.
Did COTA Err in Concluding NCRI Failed to Sustain its Burden to Rebut Presumption of Assessment Validity?
NCRI next argues on appeal that even if KDOR was entitled to a presumption of validity, that presumption was fully rebutted (1) by the KDOR auditor’s admission that the assessment does not represent actual tax liability; (2) because the sampling technique employed was flawed; and (3) because KDOR knowingly imposed sales tax on out-of-state transactions. We will examine each of these challenges, but we do so understanding that COTA made a negative finding that cannot be disturbed on appeal absent an arbitrary disregard of undisputed evidence or a showing of bias, passion, or prejudice. See Hall v. Dillon Companies, Inc., 286 Kan. 777, 781, 189 P.3d 580 (2008).
The approach utilized by KDOR in its audit was sampling; i.e., 3 months were selected from a 2-year period, and the error rate determined in those months was then spread over the entire pe*199riod. This approach is authorized by K.S.A. 79-3610, which provides:
“Any determination [of tax liability] may be made on the basis of a generally recognized valid and reliable sampling technique, whether or not the person being audited has complete records of transactions and whether or not such person consents. In any such case, the director shall notify the taxpayer in writing of the sampling technique to be utilized, including the design and population of such sample. If the taxpayer demonstrates that any such technique use was not in accordance with generally recognized sampling techniques, the portion of the audit shall be dismissed with respect to that portion of the audit based upon such technique, and a new audit shall be performed. Within 60 days after the mailing of notice of the director’s determination any taxpayer may request an informal conference ... relating to the taxpayer’s tax liability, including the issue of whether the use of a generally recognized sampling technique achieved a result that was reflective of die taxpayer’s actual tax liability.” (Emphasis added.)
COTA concluded that NCRI failed to sustain its burden to show the invalidity of the assessment, reasoning as follows:
“Based upon the evidence presented, we find Taxpayer has failed to rebut die presumption of validity in favor of the Department. It is not enough for a taxpayer to merely raise concerns about possible flaws in the audit or to cast aspersions, without providing any countervailing proof, on the factual basis of certain aspects of the Department’s analysis. Cf. Lewis v. City of South Hutchinson, 162 Kan. 104, 174 P.2d 51 (1946) (holding presumption favoring official actions of public agents cannot be rebutted by suspicion, surmise, insinuation, and innuendo). A taxpayer also cannot rebut the presumption merely by proffering, in the elevendi hour, a new hypothesis based on self-serving interpretations of its own inadequate business records.
“In order to rebut the presumption of validity, it is incumbent upon a taxpayer to show either that'the sampling technique used by the Department was invalid or that certain factual assumptions upon which the audit assessment was predicated were materially incorrect. The presumption of validity is overcome not by conjecture but by substantial competent evidence, such as business records and fact testimony, which proves both die manner and extent of the alleged assessment error.”
We agree with COTA. When sampling is the basis for a sales tax assessment, it is incumbent upon tire taxpayer to show that die technique was not in accordance with generally recognized sampling techniques. Perhaps because COTA found NCRI’s singular witness to lack foundation and credibility, no such challenge has been mounted on appeal. Instead, NCRI challenges the assess*200ment on appeal with purported undisputed evidence consisting principally of hearing testimony by KDOR’s auditor witness.
KDOR’s Admission of. “No Actual Tax Liability” Is Taken Out of Context
NCRI initially suggests that the KDOR auditor admitted in the auditor log that she agreed “failure to give credit [for belated payments in sample months] doesn’t represent the actual tax liability per K.S.A. 79-3610.” This so-called admission does not impeach the sampling procedure or the error rate employed, but it is a disingenuous attempt to gain advantage through distortion of the sample and the resulting discussion with the auditor.
The events that gave rise to the discussion were NCRI’s attempt to file supplemental returns and make payments of tax in the months already selected by KDOR for sampling. NCRI argued to the auditor that such payments should be accepted and credited to the sample months, thus reducing the error rate within the sample. Of course, KDOR refused to accept and credit these payments as requested because this would skew the sample. Instead, KDOR agreed to accept the payments and apply them to the overall tax liability that would be determined from completion of the audit. When NCRI’s representative Capps learned of this, he argued that refusal to credit the payments within the sample months would result in a distortion of “the actual tax liability” in the sample months as required by K.S.A. 79-3610. The auditor’s log states that the auditor “agreed” but that due credit would be given for the tax paid.
First and foremost, we agree with KDOR that belated tax payments need not be credited to months that have been selected for sampling because this would distort the error rate. The auditor’s log reflects the balance of the discussion and KDOR’s rationale for its credit procedure:
“Jeny C. called and talked to me about the audit.... We discussed the documents that he had given me that had changed the sample months. He said he gave credit for the invoices that they had remitted tax on. I told him that the sample was set up to determine the total tax due and then we gave credit for every dollar paid in. He said that this way would not be correct because we are projecting every dollar in the sample month to the other periods and the amount paid is only good *201for the one month. He said he had given Christina the sheet that determines which invoices they had remitted tax on in 2005. He emailed me the sheet and I told him I would review it and see what it looks like.
“I reviewed the file. It appears that around 75% of the invoices listed and paid on were from the sample month. The additional returns were filed a month after tire 3 month sample selection was made. It appeal's that NCRI tried to pay tax on the sample months to lower the projections. Jerry had told me that they should have credit for the tax paid in the sample months. It appears Aat Aey are trying to cheat the system. I talked to Jim about Ais and he agreed not to accept any of Ae tax paid on Ae projections and give them credit after the projections on Ae tax paid.”
We agree with KDOR. Attempting to reduce tax due within the sample months selected after the audit had begun would distort the error rate, invalidate the conclusions, and “cheat the system.” And the ensuing conversation between the auditor and Capps now relied on by NCRI to show some admission against KDOR’s interest is taken wholly out of context and is of no legal significance to the validity of the resulting assessment.
The Sampling Technique Was Not Flawed
Next, NCRI suggests that KDOR’s auditor witness “impeached the validity of its calculated error rate.” Testimony cited by NCRI to support this impeachment is from a portion of the auditor’s direct testimony intended to demonstrate the error inherent in NCRI’s contention that out-of-state transactions should be deducted from the general population before application of the error rate. The testimony was as follows:
“Q. [KDOR Attorney:] Okay. Now, Ae Taxpayer—assume Ae Taxpayer says our audit error rate is correct. Okay?
“A. [KDOR Auditor:] Okay.
“Q. And Aen assume that Ae Taxpayer wants to use that error rate and multiply Aat error rate times all of Ae transactions in the non-sample months. Okay?
“A. Mm-hmm.
“Q. Do you understand me so far?
“A. Got you.
“Q. Okay. Rut assume before it multiplies the error rate against all Aose transactions, it wants to throw out of the general population [out-of-state] transactions. Do you understand Ae—■
“A. Yes.
*202“Q. —assumption? Do you?
“A. Yes, I do.
“Q. Okay. If the error rate is not changed, does the quotient—does the result of the multiplication of the population that’s left in the assessment period months accurately reflect die tax owed by die Taxpayer?
“A. No.
“Q. And why is that? If you change one set of circumstances, do you not have to change die other? If you change the population—
“CHIEF JUDGE CROTTY: Let her answer.
“Q. (Ry Ms. Rawlings)—do you have to change die error rate?
“A. You’d have to do the same tiling in die same as what you’re projecting. So if you removed the non-taxable from the population,... [out-of-state] sales from the population, you would also have to do that for the tiiree sample montiis.
“Q. And if you removed the [out-of-state] transactions from die tiiree sample montiis, die error rate would have actually been higher. Correct?
“A. Correct.”
This testimony does not impeach KDOR’s error rate or sampling technique. The testimony merely serves to show that so long as out-of-state transactions are not indicated as taxable within tire sample montiis, the error rate so derived must be applied to a general population where the out-of-state transactions are not removed. In fact, the testimony would indicate that if out-of-state transactions are removed from the general population, the error rate would have to be similarly adjusted, and this adjustment would tend to drive the rate higher rather than lower.
Of particular note is that NCRI’s counsel objected to the testimony of KDOR’s auditor on the calculation of the error rate because “we specifically stipulated that the error rate was true and correct.” Thus, it appears NCRI merely objected to the application of an undisputed error rate to an adjusted general population. Again, this would be “cheating the system.” Having examined the cited testimony thoroughly, we reject NCRI’s contention that it impeaches the sampling technique. -
KDOR Did Not Intentionally Impose Sales Tax on Out-of-State Transactions
Finally, NCRI argues that the KDOR auditor’s testimony demonstrates that tax was intentionally imposed on out-of-state trans*203actions. We conclude that tax may have been imposed on some out-of-state transactions as an inherent outcome of the sampling approach to tax liability, but that consequence does not invalidate an assessment based on the statutory approach allowed under K.S.A. 79-3610.
KDOR’s auditor witness clarified precisely how out-of-state transactions were treated during the audit in her testimony on redirect. We find the following testimony revealing as to the approach and its treatment of out-of-state transactions:
“Q. And you agree that we don’t tax [out-of-state] transactions. Correct?
“A. Correct.
“Q. We do not directly tax [out-of-state] transactions.
“A. Correct.
“Q. But when you’re doing a sample, the whole purpose of the sample is you’re not going to go over every transaction during tire entire audit assessment period. Correct?
“A. Correct.
[[Image here]]
“Q. Okay. Now in those records it was clear that some were out-of-state transactions. Correct?
“A. You could assume they were, yes.
“Q. Okay. And you—and we—the audit department made that assumption—
“A. Right.
“Q. —if you could, from the document. Correct?
“A. Correct.
"Q. Those were never designated as taxable transactions. Correct?
“A. Correct.
“Q. They were simply used in the entire sample period to set up how many dollars were in those sample months resulting from sales, whether they be in Kansas or out of Kansas. Correct?
“A. Correct.
“Q. Now, you have to decide how many dollars are in all the sales transactions during tiróse months which should have been taxed but weren’t. Correct?
“A. Correct.
“Q. And those dollars go into the numerator. Correct?
“A. Correct.
“Q. Okay. You didn’t identify any of the . . . out-of-state sales transactions as being tax errors that were included in the numerator. Correct?
“A. Except for that one we discussed.
“Q. Except for tire one. And I appreciate your—your paying attention to the question and the details. Okay. So you divide the denominator—all the sales during the sample months, regardless of whether they’re taxable or *204not, they all go in there—and you divide that number into the numerator— which are just tax errors, and there’s no [out-of-state] transactions that were designated as tax errors. Correct?
“A. Correct.
“Q. And so you get a number, a percentage?
“A. Correct.
“Q. Then you go back to die whole audit period, and you say, Tm going to multiply all the transactions in this audit period by tills tax error rate and determine what the assessment should be.’
“A. Correct.”
Thus, we perceive NCRI’s position on appeal to be a challenge not to specific errors made in employing a valid sampling technique, but rather a facial challenge to sampling as an approach to die auditing. Sampling itself is not intended to result in precise identification of sales tax liability on each and every transaction within the audit period, but rather to calculate a precise error rate for a small period of time and then apply that rate to the entire period of tire audit. The statutory goal is to “achieve a result that is reflective of the taxpayer’s actual tax liability.” (Emphasis added). K.S.A. 79-3610. Unless we decide that sampling as authorized by this statute is unconstitutional, there has been no showing that KDOR’s audit of NCRI was in any way violative of the statutoiy methodology. This is especially confirmed by the complete absence of any competent evidence that the methodology utilized here resulted in any specific and calculable inclusion of out-of-state transactions. We recognize that NCRI attempted to establish this fact, but its singular witness was deemed to lack foundation and credibility by COTA, and we are bound to consider this finding of veracity in our review of record as a whole. See K.S.A. 2011 Supp. 77-621(d).
For these reasons, we reject NCRI’s challenge to tire methodology employed by KDOR, and we deem any inclusion of out-of-state transactions to be unproven; but if any were incidentally taxed, this was an inherent potential result of employing the statutory sampling approach of auditing sales tax returns under K.S.A. 79-3610.
In summary, we conclude that NCRI has not demonstrated that COTA arbitrarily disregarded undisputed evidence or manifested *205bias, passion, or prejudice. Thus, COTA must be affirmed in its conclusion that NCRI failed to sustain its burden to rebut the presumption of assessment validity.
K.S.A. 79-3610 Is Not Unconstitutional as Applied in KDOR’s Audit
NCRI next argues that “the Commerce Clause and the Due Process Clause forbid Kansas from imposing its sales tax on out of state transactions.” KDOR failed to brief this issue on appeal. A statute is presumed constitutional and all doubts must be resolved in favor of its validity. If there is any reasonable way to construe a statute as constitutionally valid, the court has the authority and the duty to do so. Rural Water District No. 2 v. City of Louisburg, 288 Kan. 811, 817, 207 P.3d 1055 (2009).
NCRI correctly observes that COTA has no authority to decide constitutional questions. See Zarda v. State, 250 Kan. 364, Syl. ¶3, 826 P.2d 1365, cert. denied 504 U.S. 973 (1992). COTA did, however, weigh in on these questions, stating:
“It is unclear whether Taxpayer now objects to the Department’s authority to tax transactions involving inventoried material sold to out-of-state customers pursuant to K.A.R. 92-19-66(d), or whether it contends that in this case the Department improperly taxed certain out-of-state transactions not covered by K.A.R. 92-19-66(d). Neither argument is availing. If the argument is tire former, we decline to pass on the issue because the issue was neither included in the pre-hearing order nor addressed at tire hearing. If tire argument is the latter, we find insufficient evidence to establish that the Department’s audit approach resulted in tire taxation of non-taxable transactions.”
In the absence of a record establishing that out-of-state transactions were indeed taxed (and both COTA and this court have concluded otherwise), there is an insufficient factual predicate for us to address the constitutional issue. The rule that a constitutional issue cannot be decided by an administrative agency does not preclude making a factual record adequate for the reviewing court to apply constitutional principles. Obviously, this court is not equipped to be a fact-finder; thus, constitutional issues must be raised at the administrative agency and factual predicates for constitutional questions must be established in that forum. See Martin v. Kansas Dept. of Revenue, 285 Kan. 625, 633-34, 176 P.3d 938 *206(2008); see also Bruch v. Kansas Dept. of Revenue, 282 Kan. 764, 773-74, 148 P.3d 538 (2006) (failure to raise the constitutional issue at the administrative proceeding fatal to jurisdiction of reviewing court on that issue); In re Tax Appeal of City of Wichita, 274 Kan. 915, 920, 59 P.3d 336 (2002) (factual predicate for equal protection claim made at BOTA but addressed on review). This rule is consistent with K.S.A. 2011 Supp. 77-618 and K.S.A. 77-619, which require that judicial review of disputed issues of fact shall be confined to the agency record except in certain limited situations.
Here, we agree with COTA in its conclusion that NCRI has not precisely articulated its constitutional challenge. We note at the outset that KDOR’s regulations allow for the taxation on the retail sale of materials from inventory even though the material may be used in a construction project outside of Kansas. K.A.R. 92-19-66(d) provides in material part:
“[B]ulk purchases of all mateiial by persons who are contractors only, and all material removed from inventoiy by a retailer to perform a construction project shall be subject to sales tax at the time of purchase or at the time the material is removed from inventoiy, even though the material may be used in a construction project outside of Kansas. No deduction, exclusion, refund or credit for sales tax shall be allowed when a contractor purchases material in Kansas, or when a retailer who is also a contractor removes material from inventory as a sale in interstate commerce, even though the material may be used in a construction project outside of Kansas.”
It does not appear that NCRI has mounted a facial constitutional challenge to this regulation, nor does the record on appeal indicate whether the auditor applied this regulation to the determination of taxable sales in the sample months. Moreover, our search of the record does not establish any details of a typical sales transaction by NCRI; e.g., we do not know if the projects were originated within Kansas, we do not know if the projects were serviced by trucks or equipment from Kansas, and we do not know if the projects were performed using materials or labor from Kansas.
Thus, we are unable to determine the precise nexus with Kansas for a Due Process or Commerce Clause analysis. See In re Tax Appeal of Family of Eagles, LTD, 275 Kan. 479, 485-86, 66 P.3d 858 (2003). Finally, as we have already noted, NCRI has not es*207tablished with competent evidence that the sampling technique employed for the audit resulted in any specific taxation of one or more out-of-state transactions and, if so, whether that transaction was bifurcated in accordance with K.A.R. 92-19-66(d) so as to tax only materials from Kansas inventoiy. In the absence of clarity of the challenge and factual predicate, we are unable to apply traditional Due Process or Commerce Clause analyses. See Complete Auto Transit, Inc. v. Brady, 430 U.S. 274, 279, 97 S. Ct. 1076, 51 L. Ed. 2d 326 (1977).
At best, it would appear that NCRI complains on appeal that the sampling technique employed in the audit under K.S.A. 79-3610 may have resulted in the taxation of some aspect of transactions that were performed outside the state. Yet, the only witness who testified to this was deemed by COTA to lack foundation and credibility, and our review of her testimony still leaves us without the clarity or factual predicate to perform any constitutional analysis. Perhaps—at best—the argument is that a sampling technique lacks precision to guarantee that Kansas has not overstepped its constitutional bounds in assessing sales tax to an entity that has performed some of its work outside the state. In constitutional terms, we consider this an apportionment issue.
We find litde guidance for an analysis of this sort. There does not appear to be helpful caselaw from our appellate courts or others across the nation that assess proper apportionment of state sales taxation when sampling techniques have been utilized to determine tax liability for an entity having some out-of-state transactions. We examine cases from the property tax context for some determining principles.
For example, our Supreme Court has acknowledged the unit valuation methodology and apportionment to Kansas using original cost as meeting the requirements set forth by tire United States Supreme Court in Commonwealth Edison Co., v. Montana, 453 U.S. 609, 101 S. Ct. 2946, 69 L. Ed. 2d 884 (1981). See In re Tax Appeal of ANR Pipeline Co., 254 Kan. 534, 547, 866 P.2d 1060 (1994) (The court indicated diat the unit valuation methodology, and apportionment to Kansas according to original cost results in *208a pipeline company’s property being assessed in Kansas in proportion to the pipeline’s presence in Kansas.).
In Norfolk & W. R. Co. v. Tax Comm'n, 390 U.S. 317, 88 S. Ct. 995, 19 L. Ed. 2d 1201 (1968), the United States Supreme Court considered Due Process and Commerce Clause challenges to state taxation of an interstate enterprise. The Court recognized that States could tax the fair share of an interstate transportation enterprise which is permanently or habitually employed in the taxing State, “including a portion of the intangible, or ‘going-concern,’ value of the enterprise.” 390 U.S. at 323-24. Norfolk recognized that the process of evaluating the going-concern value is “an elusive concept not susceptible of exact measurement.” 390 U.S. at 324.
“As a consequence, the States have been permitted considerable latitude in devising formulas to measure the value of tangible property located within their borders. [Citation omitted.] Such formulas usually involve a determination of the percentage of the taxpayer’s tangible assets situated in the taxing State and the application of this percentage to a figure representing the total going-concern value of the enterprise. [Citations omitted.]” 390 U.S. at 324.
Moreover, we also find persuasive decisions from other states that have concluded state sales tax auditors properly resorted to alternative auditing methodology on the ground of inadequate record keeping when the taxpayer was able to provide only limited documentation. For example, in Matter of Licata v. Chu, 64 N.Y.2d 873, 874, 487 N.Y.S.2d 552, 476 N.E.2d 997 (1985), the New York Court of Appeals concluded that the “auditor’s use of a test period and markup audit to estimate the tax due from [taxpayers] was neither arbitrary nor without rational basis” when the “[taxpayers’] sales tax records have as their source cash register tapes which show only total sales and sales tax collected by categories” since the “auditor could not determine from the tapes available whether tax had been charged on all taxable items and whether the proper tax had been charged in each instance.” See also McDonald’s v. Kosydar, 43 Ohio St. 2d 5, 8, 330 N.E.2d 699 (1975) (A “‘test-check method’ ” is acceptable where “no delineation showing what items were considered taxable or nontaxable, and no indication of which sales were for consumption on the premises as opposed to consumption off the premises.”).
*209Here, it was NCRI’s inability to produce reliable records that forced KDOR to employ this audit methodology, and any complaints have to be viewed in that context. See Matter of Center Moriches Monument Company v. Commissioner of Taxation & Finance, 211 App. Div. 2d 947, 948, 621 N.Y.S.2d 720 (1995). Simply because “ 1 “a different audit methodology might provide a more precise estimate of tax liability” ’ does not mean that the method used was unreasonable.” Matter of Rodriguez v. Tax Appeals Tribunal of State, 82 App. Div. 3d 1302, 1306, 918 N.Y.S.2d 625 (2011).
We deduce from all these cases a general rule that for purposes of compliance with Due Process and Commerce Clause dictates, apportionment of state sales tax liability auditing technique of an entity conducting business both within and outside a state can be achieved by a sampling if reasonably designed after consultation widr the taxpayer to determine tax deficiency within the best judgment and information available to the taxing authority.
Similarly, K.S.A. 79-3610 is a reasonable approach to measuring sales tax liability when a portion of the sáles occur beyond the state. Particularly where a taxpayer has failed to carefully document its business transactions, a sampling technique such as that contemplated by our statute is a necessary remedy. To require precise measurement and exclusion of each out-of-state transaction would benefit the taxpayer’s poor record-keeping. After the taxpayer has been given notice of the sampling technique to be employed, consulted about selection of the sample, and then' silent during a long and complicated audit, we have little sympathy for a belated constitutional challenge.
For all these reasons, we reject NCRI’s constitutional, challenge to K.S.A. 79-3610 as applied here.
Did COTA Err by Refusing to Honor Stipulations of the Parties?
Finally, NCRI challenges the portion of COTA’s order in which it found three stipulations made by the parties during hearing to be ineffectual because they were “entered into based on mistaken assumptions and would result in unintended consequences.”
*210All three of these stipulations were initiated and announced during the evidentiary hearing. The most significant, Stipulation No. 1, was introduced by KDOR as follows:
“MS. RAWLINGS: . . . [W]e have agreed that the Department will remove from the assessment documents for the month of February 2004,
[[Image here]]
“MS. RAWLINGS: And we are removing from the assessment the Hilton Week 1 and Hilton Week 2—
“CHIEF JUDGE CROTTY: Okay.
“MS. RAWLINGS: —as part of the error rate.
“MR. CAPPS: Correct.
“MS. RAWLINGS: And we will calculate-what the new error rate is and what the new tax would be prior to the close of this case. So we are stipulating that those two entries are actually already contained . . . under an entry identified as Hilton Garden Inn. So it’s a duplicate.”
But in KDOR’s posthearing prepared findings -and conclusions, it argued that the stipulation was a mistake and sought to be relieved of it, finding that—although the invoice was a duplicate—-it had already been disregarded in calculating the error rate.
With respect to Stipulation No. 1, NCRI relies on In re Tax Application of Emporia Motors, Inc., 30 Kan. App. 2d 621, 44 P.3d 1280, rev. denied 274 Kan. 1112 (2002). This case involved ad valorem tax exemptions sought by the taxpayers and for which the three counties involved agreed the exemption should be granted and that no hearing was needed. The taxpayers participated in the BOTA hearing “in a rather perfunctory manner” and assumed “the matter was essentially settled.” Emporia Motors, 30 Kan. App. 2d at 622. BOTA, however, found the taxpayers had failed to prove they were entitled to the tax exemptions claimed. Needlessly surprised, the taxpayers asked BOTA for another opportunity to appear and document their entitlement to the exemption. BOTA denied the application. A panel of this court reversed BOTA’s order, concluding that it was arbitrary and capricious under the circumstances, reasoning:
“We are not saying that BOTA must accept every stipulation submitted to it by the parties to BOTA proceedings. However, when BOTA fails to question the adequacy of the documentation submitted or to give the parties any notice that it is electing to reject their stipulations, this, in our opinion, denies due process. *211Where a party has been led to believe that it need not provide exhaustive evidence because everyone agrees it is entitled to the relief requested, BOTA must notify the parties should it decide to reject the stipulations of the parties and conduct a full and complete hearing on the application.” Emporia Motors, 30 Kan. App. 2d at 624-25.
See In re Tax Exempt Application of World Impact, No. 93,781, 2005 WL 3434015, at * 5-7 (Kan. App. 2005) (unpublished opinion).
Therefore, applying the rationale of this court’s opinion in Em-poria Motors, we conclude COTA’s rejection of Stipulation No. 1 must be set aside and the matter remanded for a full hearing before COTA to address the circumstances surrounding the disputed invoice, whether there was a mutual mistake in offering and agreeing to the stipulation, and other factors deemed by COTA to be in the interests of justice. See Connelly v. State, 271 Kan. 944, 963, 26 P.3d 1246 (2001), cert. denied 534 U.S. 1081 (2002). Ultimately, COTA must oversee recalculation of tax deficiency based on the stipulations that are not rejected.
Stipulation Nos. 2 and 3 were less significant, but they also addressed errors by KDOR in calculating the error rate in the sample months. Neither party sought to be relieved of these stipulations, however, and there is absolutely no record to support that they were “based on mistaken assumptions and would result in unintended consequences.” We must conclude that COTA’s rejection of Stipulation Nos. 2 and 3 was arbitrary and capricious and must be set aside. See Zinke & Trumbo, Ltd. v. Kansas Corporation Comm'n, 242 Kan. 470, 474-75, 749 P.2d 21 (1988) (action is ar-bitraiy and capricious if unreasonable, without foundation in fact, not supported by substantial evidence, or without adequate determining principles).
In summary, we hold KDOR was entitled to a presumption of validity of its assessment based on K.S.A. 79-3610. NCRI failed to sustain its burden to rebut this presumption of validity. NCRI’s challenge to the constitutionality of K.S.A. 79-3610 is rejected. Finally, COTA’s rejection of Stipulation No. 1 is reversed and vacated, its rejection of Stipulation Nos. 2 and 3 is reversed, and the *212matter is remanded for further proceedings not inconsistent with this opinion.
Affirmed in part, reversed and vacated in part, and remanded with directions.