Feature Validation
Statistical analysis of linguistic features across synthetic personas. All confidence intervals computed with bootstrap resampling (1000 iterations, seed=42–47).
Inter-Persona Variance
Do the synthetic personas differ meaningfully? Higher variance indicates the personas span a diverse range of linguistic patterns.
avg Length
Mean:35.700
Variance:834.013
95% CI:[15.667, 61.333]
n=6, bootstrap 1000 iterations
emoji Rate
Mean:0.394
Variance:0.133
95% CI:[0.119, 0.713]
n=6, bootstrap 1000 iterations
latin Share
Mean:0.396
Variance:0.161
95% CI:[0.091, 0.709]
n=6, bootstrap 1000 iterations
bubbles Per Turn
Mean:1.547
Variance:0.003
95% CI:[1.510, 1.590]
n=6, bootstrap 1000 iterations
laughter Rate
Mean:0.471
Variance:0.069
95% CI:[0.271, 0.662]
n=6, bootstrap 1000 iterations
no Punct Rate
Mean:0.571
Variance:0.139
95% CI:[0.277, 0.862]
n=6, bootstrap 1000 iterations
Persona Distinctiveness
How unique is each persona's linguistic profile? Computed as 1 minus average cosine similarity to all other personas. Higher = more distinctive.
| Rank | Persona | Distinctiveness Score |
|---|---|---|
| 1 | Emoji Enthusiast | 0.0220 |
| 2 | Technical Researcher | 0.0085 |
| 3 | Expressive English | 0.0063 |
| 4 | Concise Mandarin | 0.0060 |
| 5 | Balanced Bilingual | 0.0056 |
| 6 | Bilingual Early Adopter | 0.0046 |
Pairwise Persona Similarity
Cosine similarity in 6-dimensional feature space. Lower values = more distinct persona pairs.
| Persona A | Persona B | Cosine Similarity |
|---|---|---|
| Emoji Enthusiast | Technical Researcher | 0.9682 |
| Expressive English | Emoji Enthusiast | 0.9752 |
| Balanced Bilingual | Emoji Enthusiast | 0.9767 |
| Concise Mandarin | Emoji Enthusiast | 0.9844 |
| Bilingual Early Adopter | Emoji Enthusiast | 0.9855 |
| Concise Mandarin | Technical Researcher | 0.9940 |
| Concise Mandarin | Expressive English | 0.9957 |
| Bilingual Early Adopter | Technical Researcher | 0.9963 |
| Concise Mandarin | Balanced Bilingual | 0.9973 |
| Bilingual Early Adopter | Expressive English | 0.9983 |
| Bilingual Early Adopter | Concise Mandarin | 0.9984 |
| Bilingual Early Adopter | Balanced Bilingual | 0.9988 |
| Balanced Bilingual | Technical Researcher | 0.9992 |
| Expressive English | Technical Researcher | 0.9995 |
| Expressive English | Balanced Bilingual | 0.9998 |
Feature Correlation Matrix
Pearson correlation coefficients. Values near ±1 indicate strong linear relationships.
| avg Length | emoji Rate | latin Share | bubbles Per Turn | laughter Rate | no Punct Rate | |
|---|---|---|---|---|---|---|
| avg Length | 1.00 | -0.41 | 0.92 | -0.49 | -0.71 | -0.60 |
| emoji Rate | -0.41 | 1.00 | -0.16 | 0.24 | 0.89 | 0.91 |
| latin Share | 0.92 | -0.16 | 1.00 | -0.44 | -0.47 | -0.32 |
| bubbles Per Turn | -0.49 | 0.24 | -0.44 | 1.00 | 0.44 | 0.10 |
| laughter Rate | -0.71 | 0.89 | -0.47 | 0.44 | 1.00 | 0.87 |
| no Punct Rate | -0.60 | 0.91 | -0.32 | 0.10 | 0.87 | 1.00 |
Cells highlighted green have |r| > 0.4 (darker = stronger correlation)