Summary: We analyzed NWEA MAP Growth math results for 147 students in grades 7–12 who received one-on-one Passion + Purpose tutoring through Andromeda Education from 2022–23 to 2025–26. Across 266 fall-to-spring test pairs, students gained 6.7 RIT points on average versus 3.3 projected by national norms (95% CI for the difference: 1.7 to 5.1). There was no control group.
Sean Alexander, M.S. | Alexander Tutoring | October 2026 | Data: NWEA MAP Growth Mathematics, school years 2022–23 to 2025–26
Key findings
- Sample: 147 students in grades 7–12 contributed 266 matched fall-to-spring test pairs over four school years.
- Growth vs. norms: Observed growth averaged 6.7 RIT against 3.3 projected; paired t(265) = 3.90, p < 0.001.
- Effect size: Mean growth above projection was 3.4 RIT (95% CI 1.7 to 5.1), a standardized effect of dz = 0.24.
- Consistency: 57% of pairs met or exceeded their projection; the median conditional growth percentile was 59.5.
- Sensitivity: Excluding 17 pairs with score changes over 30 RIT, growth above projection falls to 1.2 RIT (p = 0.07).
- Design limit: A single-group design with no comparison group shows association with tutoring, not causation.
Background
Many students disengage from math because they cannot see what it is for. Two bodies of research bear on this problem. Randomized trials of tutoring show large average gains in achievement, with a pooled effect of 0.37 standard deviations across experimental studies (Nickow, Oreopoulos and Quan, 2020). Separately, a randomized field experiment found that asking high school students to connect course material to their own lives raised both interest and grades, with the largest benefit for students who expected to do poorly (Hulleman and Harackiewicz, 2009). Research on interest development describes how interest that is sparked and supported in a specific context can grow into a lasting individual interest (Hidi and Renninger, 2006).
Passion + Purpose tutoring combines these elements in one-on-one instruction: the tutor starts from the student’s own interests and teaches math as the tool behind them. This report describes the method and examines how students taught with it grew on a nationally normed math assessment.
The intervention: Passion + Purpose tutoring
Sean Alexander, a physicist (B.S. UC Santa Barbara, M.S. San Francisco State University) who founded Alexander Tutoring in 2009, developed the method through one-on-one teaching. Its premise appears in a 2020 Alexander Tutoring video built around the question students ask most often, “When am I ever going to use math?” (Alexander Tutoring, 2020). By 2022, the Alexander Tutoring website described the approach as rewriting the student’s math narrative by “discovering what the student IS passionate about” and “showing how math applies to their passion.”
The method has five components:
- Interest elicitation. Before content work, the tutor learns what the student cares about, such as music, art, sports, history or the natural sciences.
- Tutor matching. Students are paired with a tutor who shares those interests, considering personality and learning style.
- Derivation over memorization. Tutors build formulas with the student and connect them to physical or real-world meaning rather than drilling procedures.
- Confidence-first framing. The tutor starts at the student’s actual level and treats “I’m bad at math” as the product of past experiences rather than ability.
- Goal connection. Each topic is tied to an application in the student’s area of interest or stated goals.
| Date | Development |
|---|---|
| 2009 | Alexander Tutoring founded |
| August 2020 | “When Am I Ever Going to Use Math?” video published |
| 2022 | Website describes the interest-first approach and its principles: compassion, not shame; knowledge, not memorization; inspiration, not intimidation |
| Spring 2022 | One-on-one math tutoring at no cost begins for low-income Bay Area students (the program that became Andromeda Education) |
| Fall 2022 | MAP Growth testing begins (first school year analyzed) |
| December 2022 | Andromeda Education incorporated as a California nonprofit public benefit corporation |
| September 2024 | The method is named “Passion + Purpose” |
| June 2026 | Andromeda Education’s 501(c)(3) status takes effect |
Methods
Setting and participants
Andromeda Education is a San Francisco Bay Area nonprofit, founded by the author, that provides one-on-one math tutoring at no cost to students from low-income families. Its tutoring program was built on the interest-first, one-on-one approach that Sean Alexander developed at Alexander Tutoring and that is described above. Between fall 2022 and spring 2026, 232 Andromeda students took at least one MAP Growth Mathematics test (743 test events in total).
The analysis sample is every fall-to-spring pair within a school year for which NWEA reported both observed and projected growth: 266 pairs from 147 students in grades 7–12. Students could contribute one pair per school year. Session frequency and total tutoring hours varied by student and were not linked to test records.
Measure
NWEA MAP Growth Mathematics is a computer-adaptive assessment reported on the RIT scale, an equal-interval scale that spans grade levels. Andromeda tutors administered it in the fall (September to November) and spring (April to May). For each student, NWEA reports a projected fall-to-spring growth: the typical growth of students nationally in the same grade, with the same starting RIT score and a similar amount of instruction between tests (NWEA, 2020). NWEA also reports a conditional growth percentile, where 50 represents growth typical of comparable students nationally. Projections are those NWEA reported at the time of testing.
Analysis
The primary outcome is growth above projection: observed minus projected growth for each pair. We tested whether its mean differs from zero with a two-sided paired t-test and report a 95% confidence interval and a standardized effect (dz, the mean difference divided by its standard deviation).
Because some students contribute more than one school year, we repeated the test using one averaged value per student. We also ran a Wilcoxon signed-rank test, computed a 10% trimmed mean, and repeated the analysis excluding pairs with absolute score changes greater than 30 RIT, to check whether a few extreme results drive the average. Subgroups with fewer than 10 students are not reported. Student names and IDs were replaced with hashed codes before analysis, and no individual records are published.
Results
Primary outcome
- +3.4 RIT mean growth above projection (95% CI 1.7 to 5.1)
- 2.0× observed vs. projected growth (6.7 vs. 3.3 RIT)
- 57% of test pairs met or exceeded projection
- 59.5 median conditional growth percentile (50 = typical)
Observed growth averaged 6.70 RIT (SD 14.4; median 5) against a mean projection of 3.29 RIT. Mean growth above projection was 3.41 RIT (95% CI 1.69 to 5.13; t(265) = 3.90, p = 0.0001; dz = 0.24). Of 266 pairs, 151 (56.8%) met or exceeded their projection. Allowing for NWEA’s measurement error, 38.7% of pairs exceeded their projection by more than one standard error, 25.6% fell short by more than one standard error, and the rest were within measurement error of their projection.
Sensitivity analyses
| Analysis | Pairs (students) | Observed growth | Projected growth | Above projection [95% CI] | p |
|---|---|---|---|---|---|
| All pairs (primary) | 266 (147) | 6.70 | 3.29 | 3.41 [1.69, 5.13] | 0.0001 |
| One averaged value per student | (147) | 6.62 | 3.38 | 3.24 | 0.0007 |
| Wilcoxon signed-rank test | 266 (147) | 0.002 | |||
| Pairs with fall record in the export | 256 (145) | 6.62 | 3.29 | 3.33 [1.57, 5.09] | 0.0002 |
| 10% trimmed mean | 266 (147) | 5.53 | 2.21 | ||
| Excluding score changes over 30 RIT | 249 (144) | 4.43 | 3.25 | 1.18 [−0.10, 2.45] | 0.07 |
The result holds when each student counts once and under a rank-based test. It is smaller when extreme values are trimmed, and when the 17 pairs with score changes over 30 RIT are removed, growth above projection is 1.2 RIT and no longer statistically significant at the 0.05 level. A minority of students with very large gains therefore account for much of the difference in the mean. The median student gained 5 RIT, above the mean projection of 3.3.
Results by school year
| School year | Pairs | Observed growth | Projected growth | Above projection [95% CI] | Met projection | p |
|---|---|---|---|---|---|---|
| 2022–23 | 62 | 4.55 | 3.48 | 1.06 [−1.38, 3.51] | 56% | 0.39 |
| 2023–24 | 77 | 6.49 | 3.13 | 3.36 [0.56, 6.16] | 61% | 0.02 |
| 2024–25 | 90 | 9.46 | 3.13 | 6.32 [2.43, 10.22] | 57% | 0.002 |
| 2025–26 | 37 | 4.05 | 3.68 | 0.38 [−3.32, 4.08] | 49% | 0.84 |
Results by grade
| Grade | Pairs | Observed growth | Projected growth | Above projection [95% CI] | Met projection | p |
|---|---|---|---|---|---|---|
| 7 | 6 | Not reported (fewer than 10 students) | ||||
| 8 | 33 | 9.85 | 5.70 | 4.15 [−2.67, 10.97] | 48% | 0.22 |
| 9 | 62 | 6.50 | 3.32 | 3.18 [0.27, 6.09] | 63% | 0.03 |
| 10 | 68 | 6.50 | 3.03 | 3.47 [0.37, 6.57] | 59% | 0.03 |
| 11 | 72 | 4.96 | 2.40 | 2.56 [−1.03, 6.14] | 50% | 0.16 |
| 12 | 25 | 6.44 | 2.40 | 4.04 [−2.34, 10.42] | 60% | 0.20 |
Mean growth was above projection in every school year and every reported grade. The difference was statistically distinguishable from zero in 2023–24 and 2024–25 and in grades 9 and 10; other years and grades have intervals that include zero, partly because of smaller samples.
Attrition
Of 333 fall tests, 77 (from 73 students) had no matching spring growth result, mostly because the student did not take a spring test. Their mean fall score (236.8 RIT) was slightly lower than that of students in the analysis (240.6 RIT), a difference that was not statistically significant (Welch t-test, p = 0.21).
Student and parent survey
In Andromeda’s end-of-term student survey (May 2025; 66 responses from 64 returning students; five-point agreement scale), the share who agreed or strongly agreed with each statement was:
| Survey item | Agree or strongly agree |
|---|---|
| The program helped me improve my performance in math | 93.9% |
| The program helped me feel more positive about math | 89.4% |
| I know how to get help when I need it | 89.4% |
| I am comfortable making mistakes in math and receiving feedback | 81.8% |
| I have a better sense of my strengths and challenge areas as a learner | 80.3% |
| Learning skills from tutoring helped me in other classes | 75.8% |
| Math is a valuable tool | 68.2% |
| I am interested in taking advanced math | 62.1% |
| I enjoy the challenge of math | 59.1% |
Item wording above is paraphrased. In the parallel parent survey (75 responses), 100% agreed that their student’s math performance had improved and 96% that their student felt more positive about math.
Limitations
- No comparison group. Students were not randomly assigned, and no untutored group was tested. NWEA projections account for grade and starting score, but students who enroll in tutoring may differ from national peers in motivation or support, so growth above projection cannot be attributed to tutoring alone.
- Influence of extreme values. A small number of very large score changes raise the mean. The restricted analysis is not statistically significant, and the median and trimmed estimates are more conservative than the mean.
- Attrition. About one in four fall tests had no spring match. Students lost to follow-up started slightly lower, which could bias results in either direction.
- Testing conditions. Tutors administered the tests outside a school testing session, and test-taking effort was not independently monitored.
- Dosage. Tutoring hours were not linked to test records, so the relationship between amount of tutoring and growth could not be examined.
- Survey design. Survey responses are self-reports from returning students, were collected by tutors during sessions and were not anonymous, all of which tend to inflate positive responses.
- Implementation fidelity. Tutors’ adherence to each component of the method was not measured, and tutor training materials changed over the study period, so the results describe the program as delivered rather than a fully specified protocol.
- Author interest. The author founded Alexander Tutoring and Andromeda Education and developed the method. The analysis has not been independently reviewed.
Discussion
Students taught with Passion + Purpose tutoring grew faster on MAP Growth Mathematics than national norms project for comparable students, by about 3.4 RIT points per school year on average, and a majority met or exceeded their individual projections. The size of the difference varied by school year and shrinks when the largest score changes are excluded. The most defensible summary is that tutored students’ growth was at or above national expectations, with the average raised by a subset of students who made very large gains.
These results are consistent with experimental evidence that intensive tutoring and relevance-based instruction improve achievement (Nickow, Oreopoulos and Quan, 2020; Hulleman and Harackiewicz, 2009), but this design cannot separate the effect of the method’s components, such as interest matching or derivation-based teaching, from the effect of one-on-one tutoring itself. Two next steps would strengthen the evidence: comparing tutored students with similar untutored students in the same schools, and linking tutoring hours to test results to examine dose and response.
Conclusion
Over four school years, 147 middle and high school students who received one-on-one Passion + Purpose tutoring showed fall-to-spring math growth above national projections on average, with a confidence interval that excludes zero in the primary analysis. The evidence is descriptive and should be read alongside the limitations above.
Suggested citation
Alexander, S. (2026). Passion + Purpose tutoring and math growth: A four-year analysis of NWEA MAP Growth results, 2022–2026. Alexander Tutoring. https://alexandertutoring.com/math-physics-resources/passion-purpose-tutoring-case-study/
References
- Alexander Tutoring. (2020, August 31). When am I ever going to use math? [Video]. YouTube. https://www.youtube.com/watch?v=Ip0oAmuX-Ko
- Hidi, S., & Renninger, K. A. (2006). The four-phase model of interest development. Educational Psychologist, 41(2), 111–127. https://doi.org/10.1207/s15326985ep4102_4
- Hulleman, C. S., & Harackiewicz, J. M. (2009). Promoting interest and performance in high school science classes. Science, 326(5958), 1410–1412. https://doi.org/10.1126/science.1178712
- Nickow, A., Oreopoulos, P., & Quan, V. (2020). The impressive effects of tutoring on PreK-12 learning: A systematic review and meta-analysis of the experimental evidence (NBER Working Paper No. 27476). National Bureau of Economic Research. https://www.nber.org/papers/w27476
- NWEA. (2020). MAP Growth normative data overview. https://www.nwea.org/uploads/MAPGrowthNormativeDataOverview.pdf
Frequently asked questions
What is Passion + Purpose tutoring?
Passion + Purpose tutoring is a one-on-one math teaching method developed by Sean Alexander at Alexander Tutoring. The tutor learns what a student cares about, is matched to share those interests, and teaches math as the tool behind them, deriving formulas rather than drilling memorized procedures.
What did the study find?
Across 266 fall-to-spring MAP Growth math test pairs from 147 students in grades 7–12, students gained 6.7 RIT points on average against 3.3 projected by national norms. The 3.4-point difference had a 95% confidence interval of 1.7 to 5.1, and 57% of pairs met or exceeded their projection.
Does the study prove that Passion + Purpose tutoring causes higher math growth?
No. The study has no comparison group of untutored students, so it shows that tutored students grew faster than national norms project, not that tutoring caused the difference. Excluding a small number of very large score changes reduces the difference to 1.2 RIT points, which is not statistically significant.
What are RIT scores and projected growth on MAP Growth?
RIT is the equal-interval scale NWEA uses to report MAP Growth scores across grades. Projected growth is the typical fall-to-spring gain for students nationally in the same grade, with the same starting score and a similar amount of instruction between tests.
Who conducted the study?
Sean Alexander of Alexander Tutoring analyzed MAP Growth results from Andromeda Education, a Bay Area nonprofit he founded that provides one-on-one math tutoring at no cost to students from low-income families. The analysis has not been independently reviewed.
