Executive Summary
Key findings from the one-way ANOVA analysis of exam scores across student ethnicity groups.
Among 500 students across 5 ethnicity groups, one-way ANOVA detects statistically significant group differences for Math scores (F=7.37, p=< 0.001, η²=0.056). group E scores highest with a mean of 73.5 versus 59.6 for the lowest-performing group, a gap of 13.9 points. The η²=0.056 indicates a small practical effect, meaning group membership accounts for 5.6% of total score variance.
Descriptive Statistics by Group
Per-group mean scores, standard deviations, and student counts across math, reading, and writing.
| Ethnicity Group | N Students | Math Mean | Math SD | Reading Mean | Writing Mean |
|---|---|---|---|---|---|
| group A | 43 | 59.58 | 13.43 | 61.98 | 59.7 |
| group B | 99 | 64.12 | 14.86 | 67.59 | 65.54 |
| group C | 166 | 64.25 | 14.91 | 68.82 | 67.95 |
| group D | 127 | 66.98 | 13.62 | 69.03 | 68.96 |
| group E | 65 | 73.48 | 16.42 | 72.26 | 70.02 |
group E leads all groups in math (mean=73.5) while group A scores lowest (mean=59.6), a gap of 13.9 points. Group sizes range from 43 to 166 students, providing adequate statistical power across all groups. Reading and writing patterns broadly mirror the math rankings.
Score Distribution by Ethnicity Group
Box plot of individual math scores within each ethnicity group, showing medians, interquartile ranges, and outliers.
group E achieves the highest median math score while group A sits lowest, a 13.9-point gap between group means. Within-group variability is substantial across all groups, with a typical interquartile range of 19 points, indicating considerable score overlap between groups. Individual performance within each group spans nearly the full 0 to 100 range.
Mean Math Score by Group
Mean math scores ranked from lowest to highest across ethnicity groups.
group E leads all groups with a mean math score of 73.5, outperforming group A (mean=59.6) by 13.9 points. The overall student mean is 65.7. Groups show a clear ordinal ranking with no ties at the group-mean level.
ANOVA Results: F-Test and Effect Sizes
One-way ANOVA F-statistics, degrees of freedom, p-values, and eta-squared effect sizes for math, reading, and writing.
| Subject | F Statistic | P Value | Df Between | Df Within | Eta Squared |
|---|---|---|---|---|---|
| Math | 7.37 | 1e-05 | 4 | 495 | 0.0562 |
| Reading | 3.379 | 0.0096 | 4 | 495 | 0.0266 |
| Writing | 4.061 | 0.003 | 4 | 495 | 0.0318 |
3 of 3 exam subjects show statistically significant group differences (p < 0.05). The largest eta-squared is 0.056, indicating that ethnicity group explains up to 5.6% of score variance in that subject. Effect sizes across subjects are small by Cohen's benchmarks.
Tukey HSD: Pairwise Group Significance
Symmetric heatmap of mean Math score differences between all ethnicity group pairs after Tukey HSD multiple-comparison correction.
5 of 10 pairwise group comparisons are statistically significant (adjusted p < 0.05) for Math scores. The largest mean difference is 13.9 points. Darker cells indicate wider score gaps; the diagonal represents a group compared to itself (difference = 0).
Group Score Profiles: All Subjects
Mean scores by ethnicity group across math, reading, and writing, showing whether group performance patterns are consistent across all exam subjects.
The top-performing group for math (group E), reading (group E), and writing (group E) is the same group across all three subjects. This consistency suggests that group-level performance differences reflect a general academic advantage rather than subject-specific factors. Reading and writing scores tend to be higher than math scores across all groups, but the relative ranking of groups remains stable across subjects.
Assumption Checks: Levene's Test
Levene's test p-values for homogeneity of variance across ethnicity groups for each exam subject.
Levene's test for homogeneity of variance passes all of 3 exam subjects (p > 0.05). The equal-variance assumption holds for all subjects, validating the ANOVA results. The horizontal reference line at p=0.05 marks the significance threshold.
Your turn
Bring your own data and the question you actually need answered.
MCP AnalyticsData science by email Send us your data and your question, and we will reply. ds@mcpanalytics.ai