"Net price" names two different numbers in the same section of the statute, and confusing them is the commonest error about this tool. The published figure is defined at 1015a(a)(3) as "the average yearly price actually charged to first-time, full-time undergraduate students receiving student aid at an institution of higher education after deducting such aid." The calculation is the institution's cost of attendance minus the total need-based and merit-based grant aid divided by the number of aided students. Three properties follow and all three matter. It is an average, so no particular family is described by it. Its population is first-time, full-time undergraduates who received aid, which excludes transfer students, part-time students and anyone who got nothing. And only grant aid is deducted, so loans and Federal Work-Study in a package are not netted out of it. The calculator's output, by contrast, is an individualized estimate for the household that typed in the numbers.
At a public college the published figure is a resident's figure. Subsection (b)(1) provides that for a public institution the cost of attendance, net price and tuition and fees are all calculated for first-time, full-time undergraduates "who are residents of the State in which such institution is located," with the grant-aid average computed over resident students as well. So an out-of-state family comparing published net prices across public universities is comparing prices none of them would be charged. Running each college's calculator with the household's real details is the only way to get a figure that reflects the non-resident rate.
The cost of attendance inside this section is not the cost of attendance the aid office uses, and the two are not reconcilable. For the purposes of section 132, 1015a(a)(2) defines cost of attendance as "the average annual cost of tuition and fees, room and board, books, supplies, and transportation for an institution of higher education for a first-time, full-time undergraduate student enrolled in the institution." That is five components, built for the Department's reporting and comparison lists. The figure that drives federal aid eligibility is the fourteen-component statutory definition at 20 U.S.C. 1087ll, which includes items such as dependent care, disability-related expenses, loan fees and licensure costs. A reader who assumes one definition while looking at the other will find the numbers do not tie, and neither definition is wrong.
The mandatory disclaimer is the most useful text on the page. Subsection (h)(4) requires "a clear and conspicuous notice" stating that the estimate "does not represent a final determination, or actual award, of financial assistance," that it "shall not be binding on the Secretary, the institution of higher education, or the State," and that it "may change." It further requires the notice to state that the student must complete the FAFSA in order to be eligible for and receive an actual federal award, and to include a link to the Department's FAFSA site. Read together, those three limbs are the statute telling the reader exactly how much weight to put on the number: enough to shorten a list of colleges, not enough to sign anything.
Where the estimate is weakest is predictable. Because (h)(2) asks only for the aid "available" to be calculated for the individual "as much as practicable," a calculator can model formula-driven aid tightly and can only approximate aid that is competitive. Institutional merit awards are decided by the awarder rather than by a formula a family can reproduce, so a college whose aid is heavily merit-based will produce a looser estimate than one whose aid is almost entirely need-based. That is not a defect in the tool; it is what the statute asks it to do.