DA and MHN Relationship
DA and MHN Relationship Analysis
This section reports the initial statistical relationship between Developmental Affordances (DA) and Maslow’s Hierarchy of Needs (MHN) scores, using three analytic approaches: binary (presence/absence), weighted (category-balanced), and cosine similarity (pattern overlap). The aim is to establish whether the occurrence or richness of certain affordances is systematically associated with the activation of particular human needs, prior to more detailed modelling.
Methodology
DA: 10 Object Categories per image
HMN: 6 Need Levels
Binary data analysis (Raw): Each cell recoded as 1 (category present, i.e., any non-empty cell) or 0 (category absent).
Category balancing: Weighted analysis subsamples an equal number of images per search category to ensure fair representation
Statistical Test:
- Pearson’s r (linear correlation) for all DA–MHN pairs (60 tests per method)
- Cosine similarity (pattern overlap) for each pair, focusing on images where at least one of the two variables is present
Multiple Testing: Bonferroni correction applied (α = 0.05/60 ≈ 0.00083).
Results
| Analysis Type | Correlations | p < 0.05 | Bonferroni Sig | Max |r| Observed | Strongest Pair |
|---|---|---|---|---|---|
| Binary (filtered) | 60 | 60 | 60 | 0.66 | Safety Needs – Vehicles (r=0.66) |
| Weighted (filtered) | 60 | 59 | 58 | 0.47 | Self-Actualisation – Plants (r=0.47) |
| Cosine similarity | 60 | – | – | 0.79 | Self-Actualisation – Plants (cos=0.79) |
Table 1. Summary of relationships between Developmental Affordances (DA) and Maslow’s Hierarchy of Needs (MHN) in images, using binary (presence/absence), weighted (object count per image, category-balanced), and cosine similarity (pattern overlap) approaches.
Filtered” = only images where at least one of the two variables is present.
Conclusion
This analysis demonstrates that, although nearly all DA–MHN pairs are statistically significant (mainly due to the large sample size), effect sizes are typically modest.
In the binary (filtered) analysis, the strongest association observed was between Safety Needs and Vehicles (r = 0.66), while the highest in the weighted (object count, filtered) analysis was between Self-Actualisation and Plants (r = 0.47).
Cosine similarity further highlights Self-Actualisation and Plants as a pair with substantial co-occurrence (cos = 0.79). These results confirm that while DA–MHN relationships exist in the dataset, most are moderate rather than strong, emphasizing the need to interpret both p-values and effect sizes when assessing the practical significance of these links.