How to Choose a Math Curriculum: 7 Questions to Ask Before You Buy

There is no best math curriculum, and nobody chooses well from a list. How to choose a math curriculum is to answer seven questions in the right order, because each honest answer deletes a large part of the market before brand names ever matter.

The order does the work: your teaching time first, because a program you cannot teach daily fails no matter how good it is; then your child’s best format; then the structure underneath; then placement, true cost, and evidence; and the exit plan last, before the receipt, not after the regret.

Every question below ends with one concrete thing you can do this week. Programs appear only as examples, with their facts checked on the publishers’ own pages in October 2026, because the point of this guide is the framework, not another crowned winner.

Quick answers: Question 1, time: how much of the teaching can you honestly do each day? A fully scripted program can ask for 30 to 40 parent-taught minutes daily, while an app-taught one asks for almost none.

Question 2, format: does your child learn new ideas best with objects in hand, on paper, or on a screen? Buy the format that survives being explained back the next morning.

Question 3, structure: mastery programs finish a topic before moving on, spiral programs keep circling back, and the right match is the one that fits how your child actually forgets. Question 4, placement: what level is your child actually at? Free placement tests answer this better than any grade label.

Question 5, true cost: what does year one cost all in, counting the set, required manipulatives, per-student fees, printing, and year two after the one-time parts are bought? Question 6, evidence: how will you know it is working, from built-in tests, a dashboard, cumulative review, or a folder you keep yourself?

Question 7, exit: what happens if it fails by October, a free trial, a refund window, or a free fallback that starts tomorrow?

The 7 Questions

An abacus, a stack of three workbooks, and a tablet with a blank screen on a table, the three formats a math curriculum can take
Hands-on, on paper, or on a screen: most of the seven questions are really asking which of these your family can live with daily.

Answer these in order because each question is cheaper to answer than the next one is to get wrong. Time and format cost nothing to assess and eliminate the most programs. Placement costs one printed test.

Cost and exit terms cost only careful reading. Brand comparison is whatever is left at the end, and by then it is a short list you can actually trial.

Question 1: How Much of the Teaching Can You Honestly Do Each Day?

This question decides more than any brand name, because the single best predictor of a curriculum surviving October is whether its daily teaching load fits your real day.

Programs split into three honest camps: parent-taught, where you deliver every lesson; video- or app-taught, where the program delivers and you supervise and check; and written-to-the-student, where the book teaches in print and you introduce and mark.

Prep and teaching time are different numbers, too: prep is what you do before the lesson, teaching time is you, present, for the lesson itself, and publishers rarely print either.

The honest spread looks like this. RightStart, the fully scripted parent-taught example in this site’s research, states on its own pages that its daily lessons take 30 to 40 minutes with the parent teaching throughout, as listed on rightstartmath.com at the time of checking, October 2026.

At the other pole, Teaching Textbooks teaches and grades every lesson inside its app, and its own FAQ notes each course carries at least 100 hours of audiovisual teaching.

Between the poles, Math-U-See’s parent watches a short video first and then coaches while the child builds with blocks, and Math Mammoth’s worktexts carry the teaching in the student pages, so the parent introduces a topic briefly and then checks work.

One trap hides inside this question: parent-taught minutes scale worst across siblings, because they cannot be parallelized.

Do this: take last Tuesday as your template day and count the uninterrupted 30-minute blocks you actually had before noon. If the honest count is zero, cross every fully parent-taught program off your list today, however much you admire it, and shop only in the app-taught, video-guided, and written-to-the-student camps.

Question 2: How Does Your Child Learn Best: Hands-On, on Paper, or on a Screen?

Format is the question the child answers, not the parent. The elementary years run on a ladder: children meet a new idea with objects first, then pictures and models, then symbols.

Programs board that ladder at different rungs, and a child who still needs the first rung will drown in a symbols-first book, while a child who has left it will be bored building every problem out of blocks.

The practical question is which format carries NEW ideas best for this child, because review can ride in any format.

Three format families cover nearly everything a K-5 parent will meet: manipulative-led, print-led, and screen-led. RightStart builds its lessons around the AL Abacus and card games, with its manipulative package required for Levels A through F.

Math-U-See’s early levels teach with an Integer Block Kit, and its fractions level swaps the kit for a Fraction Overlay Kit, a detail worth knowing because manipulatives can change mid-program.

Dimensions Math is the print-led pictorial example: objects become bar models and the area model before symbols, taught from a Home Instructor’s Guide.

Zearn is the screen-led example, with digital lessons, visual models, and paper problem solving inside every lesson, and reading positions on a number line is one of the model skills that transfers across all three formats.

Do this: teach your child one genuinely new idea three ways on three days, with counters on the table, with a drawn model, and with a short video, and watch which version they can explain back to you the next morning. Buy the format that survived the night.

Question 3: Mastery or Spiral: Finish a Topic First, or Keep Circling Back?

This is the structural choice underneath every program, and it is the one parents feel daily without knowing its name. A mastery program camps on one topic until it is secure, then moves. A spiral program touches many topics in rotation and deepens each pass. Neither is better in the abstract.

The mismatch is what hurts: a child who forgets fast starves in a pure mastery program when fractions vanish for four months, and a child who needs depth frays in a spiral that never lets a topic land.

You can read the structure off the table of contents before you read any marketing. A mastery year reads like a short list of long chapters, and a spiral year reads like a long list of short chapters that recur.

Many real programs blend the two, and when a publisher’s own description says cumulative review or gentle spiral review, that blend is telling on itself. Math-U-See is the clean mastery example, with Greek-letter levels each devoted to a topic band, one level on division and the next on fractions.

The Good and the Beautiful describes its own math as mainly spiral with constant review. Math Mammoth is a mastery worktext that prints cumulative review into the package, and RightStart blends hands-on mastery with review carried by its card games.

Do this: look at your child’s last month of math and name the failure you actually saw. “Learned it, lost it by next month” points spiral-ward, or to a mastery program with real cumulative review.

“Never got to finish anything before it changed” points mastery-ward. Then read your finalist’s table of contents before its marketing page, because chapter shape never lies.

Question 4: What Level Is Your Child Actually At?

Grade labels are marketing, and placement is measurement. Three of the most-chosen program families in this site’s research do not use grades at all: Math-U-See names its levels with Greek letters, RightStart runs Levels A through H, and Beast Academy runs Levels 1 through 5, none of which map one-to-one onto school grades.

A child placed by age into the wrong level gets a year that is either review or drowning, and the parent usually blames the program. Sequences also differ between publishers, so finishing grade 3 in one program can mean standing halfway through grade 3 in another.

A placement test is an end-of-year test used in reverse: give the test for the level you think the child has completed, read the score by topic, and let the gaps, not the total, place the child. Math Mammoth’s placement tests are free on the publisher’s own site, they are end-of-year diagnostics for grades 1-8 whose questions are grouped by topic, and the publisher prints the interpretation bands: 90% or more, take the next higher test; 80-89%, move on to the next grade level; 70-79%, move on but fill gaps; 50-69%, the child may place at that level or should take the lower test for a fuller picture; under 50%, take the next lower test. Place by the weakest core strand, not the average: the gaps surface first in topics like place value, and a comparison item a child reads as less than when it is not tells you more than a total score does.

Do this: before you price anything, print the free Math Mammoth end-of-year test for your child’s last completed grade, give it across two sittings, and score it by section. That one sheet of paper is now the most valuable document in your search, and every finalist gets judged against it.

Question 5: What Does Year One Really Cost, All In?

The sticker price is the least honest number in homeschool marketing, because the real bill is assembled from parts the headline omits: required manipulatives sold separately, per-student pricing that multiplies by children, consumable workbooks that cannot be handed down, printing for PDF programs, and the second year, when the one-time parts are already on your shelf.

Sort every cost into three buckets: one-time family costs bought once, per-child consumable costs that recur for every child every year, and hidden running costs like ink, paper, and a device per child.

The verified arithmetic, all as listed at the time of checking, October 2026: RightStart is the one-time-cost lesson, with a Level Book Bundle at $99.25 and its required RS2 Math Set at $219.75 on rightstartmath.com, so year one is about $319 and every later level is about $99 because the set is bought once per family.

Teaching Textbooks is the per-child lesson, at $48.95 per student for a 12-month enrollment on teachingtextbooks.com, with the publisher’s Large Family Discount Plan capping 4 to 8 students at $239.95, which only beats separate enrollments once a family passes four students.

Math Mammoth’s complete grade download is $42.50 on mathmammoth.com, reusable as a file, with printing as the variable cost.

The Good and the Beautiful inverts the problem: its full Course Book PDFs are free on goodandbeautiful.com, and its printed Course Sets run about $59 to $63, which is the benchmark your home printing has to beat.

Dimensions Math shows the component build on singaporemath.com: student books at $14.50 each and Home Instructor’s Guides at $22.50 each, the guide being the part families are tempted to skip and the part that carries the teaching.

Do this: for each finalist, write one line: year one all-in, year two all-in, times the number of children who will use it, plus printing. Compare lines, not stickers. If a required item’s price is not printed on the publisher’s own store, that is itself a finding, so budget a question to the publisher before you budget the money.

Question 6: How Will You Know It Is Working?

Parents discover this question in February, when a state review, a skeptical relative, or their own doubt asks for evidence and the program they bought kept none.

You need three kinds of evidence: unit-level checks that a chapter landed, a cumulative signal that last month’s topic is still there, and a record of scores, dates, and work samples that survives the year. App platforms generate the second and third automatically. Print programs generate the first well and leave the record to a folder.

Plan-based free programs may generate none, which is fine if you know it in August rather than February.

The verified spread: Math Mammoth’s downloads include answer keys, tests, and cumulative reviews alongside the worktexts, and its free end-of-year tests double as an annual external check for any program, not just its own.

Khan Academy’s mastery system and parent dashboard track what a child has mastered course by course at no cost, and Zearn reports progress through the individual teacher account. Easy Peasy is the honest opposite example: formal tests are scarce and it keeps no records or grades, so the parent is the registrar.

Fact fluency is the easiest strand to self-check, because fact families either come back instantly or they do not, and this site’s free tools help with the rest: a worksheet generator for a targeted re-check set, a multiplication chart as the facts reference, and a fraction calculator and long division calculator for verifying written work step by step.

Do this: decide your evidence kit before the first lesson: where scores will be written down, which test you will give in May regardless of program (the free Math Mammoth end-of-year test for the grade is the standing suggestion), and one work sample per month to file. Then check that your finalist actually produces the pieces your kit assumes.

Question 7: What Is Your Exit Plan If It Is Not Working by October?

Most curriculum regret is not choosing wrong. It is staying wrong because leaving felt expensive, embarrassing, or unplanned. An exit plan has three parts: a decision date, because a fair trial is long enough to pass the novelty dip, commonly the first six to eight weeks; a trigger written in advance, based on evidence from Question 6 rather than a mood; and a route out, through a refund window, an exchange, resale of non-consumables, or a free bridge program while you re-choose. Consumable workbooks and expired enrollments are sunk costs. The shelf items and your child’s confidence are what you are protecting.

The verified terms differ more than teaching quality does.

Teaching Textbooks offers the softest landing in this site’s research: the first 15 lessons of any level are a free trial with no payment details taken and trial work transferring into a paid course, and paid enrollments carry a 30-day money-back opt-out, as stated in the publisher’s own FAQ at the time of checking, October 2026.

CTCMath states a 12-month money-back guarantee in its own promotional copy, the longest window in this series.

The Good and the Beautiful removes the exit problem for its course books entirely, because the full PDF is free: the trial is the product, and money only enters if you later choose the printed set.

And the free fallbacks verified on this site (Khan Academy, Zearn, Easy Peasy, MEP, Under the Home, CK-12) are what make a mid-year exit a plan instead of a panic, because a child can be learning from a free spine within a day.

Do this: write your finalist’s exit terms on the same line as its price: trial length, refund window, and what is consumable and therefore lost.

Then circle a date six to eight weeks into your school year and write down, in advance, the two or three observations that would trigger a switch. Deciding the trigger while you are calm is the whole trick.

A Worked Example: One Family, Seven Answers

An open planner with two workbooks standing side by side and one pencil, one family's shortlist narrowed to two finalists
Seven answers turned eight recommendations into two finalists and one free trial.

This family is an invented example, built to show the framework working; every program fact inside it is real and attributed. One parent works part-time.

Their third grader is hands-on and fact-shaky, standing two topics behind in fractions after a year in a program the family is leaving, and a kindergartner starts alongside. The budget is real but not zero, and eight recommended program names are on the table.

Watch the list shrink.

Time: the parent’s honest count is one 20-minute block on school mornings, shared between two children, so every fully parent-taught scripted program is out, including the 30-to-40-minute daily commitment RightStart states for its lessons. Eight names become five.

Format: the third grader explains counters back perfectly and forgets worksheet pages by morning, so the new-idea format must be hands-on or screen-with-models, and the print-led pair drops out. Five become four. Structure: the failure pattern is “learned it, lost it”, so the pick needs cumulative review built in, and the pure one-topic mastery level drops out. Four become three.

Placement: the free Math Mammoth end-of-year test for grade 2 comes back strong in computation, where the regrouping work is secure, and under 50% in the fractions sections. By the publisher’s own bands that is not a grade-3 start in a new sequence, and one finalist’s placement test agrees. Three names are now two levels of two programs.

True cost: line one is the app-taught finalist at $48.95 for one student for 12 months, as listed on teachingtextbooks.com at the time of checking, October 2026, with no second enrollment this year because the kindergartner is below its starting level.

Line two is the worktext finalist at $42.50 for the grade download plus home printing, as listed on mathmammoth.com at the time of checking, October 2026, reusable for the kindergartner later. Both lines fit, so the decision moves to evidence and exit terms, exactly as it should.

Evidence: the app reports automatically, and the worktext brings tests, answer keys, and cumulative reviews in the download, so both satisfy the family’s May-test plan and no one is eliminated.

Exit: the app finalist gives its first 15 lessons free with no card taken and a 30-day money-back opt-out, so the family trials it for two weeks against the worktext’s free samples and the placement evidence, knowing Khan Academy or Zearn can bridge within a day if both finalists fail.

No question asked which program is best, and the family never read a single review. Time, format, and structure did the eliminating, placement set the level, and cost and exit terms picked the order to try things in.

Red Flags: Marketing Traps to Skip

A glass piggy bank with four blank price tags in front of it, for the hidden costs behind math curriculum marketing
If it needs your card, it is not free. If the required pieces are sold separately, the sticker is not the price.

“Free” that is a countdown. Mathseeds’ own site headlines a free 30-day trial that converts to a subscription. SplashLearn is free for classrooms, but a parent at home gets a 7-day trial that requires payment details. DreamBox offers a 14-day trial and no free tier. The rule: if it needs your card, it is not free, it is a deadline.

A free tier dressed as a course. IXL gives non-members 10 free practice questions per subject per day, with progress not saved, and it is a practice platform rather than a taught course.

Prodigy’s own FAQ says its educational questions are free forever, and it is still a game rather than a curriculum: wonderful practice, no sequenced teaching. The rule: free practice is a supplement, never the spine.

The manipulative that lives outside the price. RightStart’s required Math Set is a separate purchase from its level bundle, and The Good and the Beautiful states on its own pages that its early-level Math Box is also required to complete the course.

Neither publisher hides this. Both print it where a skimming shopper will miss it. The rule: before checkout, search the product page for the words “also needed” and “required”.

A grade label on a level program. Greek letters, Levels A-H, and Levels 1-5 are skill ladders, not grades, and their publishers place children by test, not by birthday.

A listing that calls a level “the 5th grade book” without mentioning placement is selling you the wrong unit of measurement. The rule: if the program has levels, the placement check is step one, not a courtesy.

Per-student pricing that multiplies quietly. A price of about $49 sounds gentle until the FAQ’s words “per student” meet a family of four children, which is why one publisher’s own family plan exists. The rule: any price with “per student” in the FAQ gets multiplied by your real headcount before it gets compared.

“Covers everything” with no tests or records. Some programs, openly and honestly, keep no records and offer scarce formal testing. Others bury the same gap. The rule: if the product page cannot show you its tests, answer keys, or reports, the evidence system is you, priced at your time.

If Your Answers Point Here, Look at This Kind of Program

Little teaching time and a screen-comfortable child: look at app-taught programs first, and price them per student before you fall in love. Daily teaching time and a hands-on child: look at scripted manipulative programs, and budget the one-time manipulative set as part of year one.

A child who forgets fast: look at spiral programs or mastery programs with printed cumulative review, and check the table of contents for the review rather than the brochure.

A tight or zero budget: look at the verified free spines, free course-book PDFs and nonprofit platforms, and spend money only where a catch demands it, on printing or a required manipulative box.

A child between levels or full of gaps: look at programs with free placement tests and level freedom, and place by the weakest strand.

This site’s grade-by-grade and free-curriculum guides apply exactly this flow level by level, for families who want the questions already answered for a specific grade or a zero budget.

Frequently Asked Questions

What is the best math curriculum for homeschool?

There is not one, and any list that crowns one is answering a different family’s questions.

The best program for your family is the one that fits your teaching time, your child’s best format for new ideas, a structure matched to how they forget, a correctly placed level, an all-in cost you have actually totaled, and terms that let you leave if it fails.

In this site’s research the same programs trade places at the top of every grade list depending on those answers, which is the point: run the seven questions first, and the short list writes itself.

How do I know what level to start my child at?

Test, do not guess, and do not trust the grade on the cover. Several strong programs use levels instead of grades, and sequences differ between publishers.

The deepest free diagnostic is Math Mammoth’s set of free placement tests on mathmammoth.com: end-of-year tests for grades 1-8, grouped by topic, with printed score bands (80-89% means ready to move on, 70-79% means move on and fill gaps, under 50% means test a level lower).

Place by the weakest core topic, not the average score.

Is mastery or spiral better for elementary math?

Neither, in general, because the match to your child decides. Mastery programs finish one topic before moving on, which suits children who need depth and completion. Spiral programs revisit topics constantly, which suits children who learn it and lose it.

Many families land on a blend: a mastery spine with printed cumulative review, or hands-on mastery with review built into games. Diagnose your child’s forgetting pattern first, then read the finalists’ tables of contents, because chapter shape shows the structure honestly.

How much does a homeschool math curriculum cost per year?

Anywhere from $0 to about $319 for the first year, using publisher prices checked in October 2026. Complete curricula exist at $0, through free course-book PDFs and nonprofit platforms. A complete download runs about $42.50 a grade, and an app-taught enrollment runs about $49 per student per year.

The top of the range in this site’s research is a scripted hands-on program at about $319 in year one, because its required manipulative set (about $220) is bought once, and later years fall to about $99. The honest number is always the all-in line: set plus required extras, times your number of children, plus printing.

Can I switch math curriculum mid-year?

Yes, and it is common enough that your exit plan should assume it is possible from the day you buy.

Switch well by placing before you purchase the replacement, because a free end-of-year test shows exactly which topics the new program must start from, by bridging with a free spine that can start the same day if there is a gap, and by treating consumable workbooks as a sunk cost.

Buying with the exit in view helps: one verified program gives its first 15 lessons free with no card taken and a 30-day money-back opt-out on enrollments, and free-PDF programs make the trial the product itself.

How much time should homeschool math take each day?

It depends on who is teaching, which is why this is Question 1. The one publisher in this site’s research that states a daily figure puts its parent-taught lessons at 30 to 40 minutes a day.

App-taught and worktext programs run at the child’s pace inside a similar daily slot, with the parent’s share shrinking to checking and helping.

The number to fix before buying is not the child’s minutes but yours: count the uninterrupted teaching blocks your real day contains, and only shop in the camps that fit them.

Choosing is a sequence, not a verdict. Answer the seven questions in order and the market shrinks to a short list you can trial with confidence, and the right program is simply the one still standing in June because it fit the answers you gave in August.

Our free math tools page holds the practice and checking helpers that support whichever program survives: a worksheet generator for targeted practice, calculators that show every step, and a multiplication chart for the facts underneath everything.

Try It With a Free Lesson

Every topic clicks faster with practice. Our Resource Library has free lesson packs with slides, printable worksheets and full answer keys, ready to download and teach.

Members unlock every lesson in the library, including the full Fractions, Multiplication, Place Value, Addition & Subtraction, Time & Money and Geometry & Shapes units. See the full library

Leave a Comment