Table 1.
Global model results showing strength of effect for each variable in explaining gentrification in the presence of covariates.
| New greenspace from prior period (same as gentrification +2) |
Prior green coverage (% of area in 1990) |
New residential buildings in prior period (total number) |
Population change (city-level) | GDP change (city-level) | City center (distance to) | Tract density (2010) | City size (+/– 1 million) | Time from 1990 (until first time period) | |
|---|---|---|---|---|---|---|---|---|---|
| H1: Composite gentrification score, period 2 (2000s) | − | + + | ++ | − − | |||||
| H2: Composite gentrification score, periods 2–3 (2000 + 2010s) | + | − − | + + | − − | − − | ||||
| H3: Composite gentrification score, period 3 (2010s) | ++ | + + | − − | ++ | − − | − − | + + | − − |
The results were based on a general interpretation as follows: p > 0.50 → positive effect on gentrification; p < 0.50 → negative effect on gentrification; p ~ 0.00 → variable is relevant and negative effect; p ~ 0.50 → variable not relevant; p ~ 1.00 → variable is relevant and positive effect. Note that new transit data is not included in global results due to a lack of information on variables across all cities.
++ strong positive effect, + positive effect, − negative effect, −− strong negative effect.
Greenspace appears to be an increasing factor in gentrification over time, with relevance shifting from negative 2000s gentrification to positive for 2000-2010s gentrification and strongly positive for 2010s gentrification.