# Results

The following page summarizes the results of our project.

## Research Question

As stated in the methods section of this webpage, our goal was to find an effect of gentrification on population growth in Zurich's quarters. Our H_{0} hypothesis became: There is no significant causality between economic status growth (in gentrified quarters) and population growth.

## Plots

### Development of our gentrified quarters

Having determined five gentrified quarters (Mühlebach, Seefeld, Hottingen, Fluntern and Escher Wyss) in the city of Zurich and calculated all the necessary values for them, you can see the development of these five quarters in terms of economic status and population growth from 1999 to 2012 in the diagram below. The quarter of Escher Wyss hereby stands out with a very big population increase of 71.2% during this time period, while the other quarters pretty much stayed between -5% and 5%.

**Development of our gentrified quarters: **Plot of the economic status change and the population change compared to 1999 of our 5 gentrified quarters (Mühlebach, Seefeld, Hottingen, Fluntern and Escher Wyss) over the years, until 2012. The line allows align the data points on a timescale, since they are not more recent, the further right they lay.

To improve the visualization of these four lower change quarters, by not stretching the y-axis so much, we also added a diagram of those four without the Escher Wyss quarter.

**Development of our gentrified quarters: **Plot of the economic status change and the population change compared to 1999 of 4 gentrified quarters (Mühlebach, Seefeld, Hottingen, Fluntern) over the years, until 2012. The line allows align the data points on a timescale, since they are not more recent, the further right they lay. The quarter of Escher Wyss is excluded to allow for a better resolution of the other 4 quarters.

Just looking at those two plots already can lead to some assumptions about our research question. Especially the seemingly mostly random pattern of the four quarters that experienced a lower population increase, doesn't seem to show an effect of economic status increase on population change.

### Changes in gentrified quarters per year

To further visualize the possible effect of economic status change in gentrified quarters on the population change, there is also a scatter plot showing the yearly changes of those two variables for all the gentrified quarters combined. If there was a correlation, those points should approximately lay on a line, and while there can be argued, that they seem to lay on a horizontal line around 0% population change, this would mean that there is an increase in economic status, while there is basically no increase or decrease in population, which sows further doubt about the effect of economic status change on population change in gentrified quarters.

**Yearly population and economic status changes in gentrified quarters: **This scatter plot shows the percental change of economic status and population in our 5 gentrified quarters, over the course of one year to the next, for each year in the time period of 1999 to 2012.

### Changes in all quarters per year

Broadening the spectrum of quarters, which we look at, we can also plot a scatterplot of the yearly changes in economic status and population of all the quarters in Zurich. This, much like the previous diagram, doesn't seem to show any relevant correlation at all between our two studied variables, instead making it even more certain, that there most likely won't be any, by increasing the number of samples, while still showing about the same pattern of distribution.

**Yearly population and economic status changes in Zurich's quarters: **This scatter plot shows the percental change of economic status and population in all of Zurich's quarters, over the course of one year to the next, for each year in the time period of 1999 to 2012.

## Statistical tests

### Linear regression analysis for the combined values of the gentrified quarters

Feeding our values of the yearly change in gentrified quarters into the statistics tool IBM SPSS statistics 21, and running a linear regression analysis, results in following output. In the ANOVA output, we can see that the F-test value is very low with 0.067, while the significance value is quite high with 0.797. This is exactly opposite of what would be desired to have a correlation between those two variables and our hypothesis is falsified, respectively the H_{0} hypothesis is true, due to the significance value of 0.797 which should be below 0.05 for a significant influence of the economic status on population. Furthermore, even if the significance wasn't this low, the R square value in the Model Summary further indicates, that only 0.1% of the population change can be explained by the economic status change. This is far beyond anything that would be desired in context of our hypothesis.

#### ANOVA^{a}

Model | Sum of Squares | df | Mean Square | F | Sig. |
---|---|---|---|---|---|

Regression | 0.966 | 1 | 0.966 | 0.067 | 0.797^{b} |

Residual | 922.056 | 64 | 14.407 | ||

Total | 923.021 | 65 |

a: Dependent Variable: change_population_perYear

b: Predictors: (Constant), change_economicStatus_perYear

#### Model Summary^{b}

Model | R | R Square | Adjusted R Square | Std. Error of the Estimate |
---|---|---|---|---|

0.032^{a} |
0.001 | -0.015 | 3.79567 |

a: Predictors: (Constant), change_economicStatus_perYear

b: Dependent Variable: change_population_perYear

### Linear regression analyses for individual gentrified quarters

Quarter | F-Test value | Significance | R Square | Significant correlation |
---|---|---|---|---|

Mühlebach | 0.138 | 0.717 | 0.012 | no |

Seefeld | 0.076 | 0.788 | 0.007 | no |

Hottingen | 1.018 | 0.333 | 0.078 | no |

Fluntern | 0.485 | 0.5 | 0.042 | no |

Escher Wyss | 0.008 | 0.929 | 0.001 | no |

As for the linear regression analysis of the combined quarters, there cannot be seen a significant dependency in any of the individual quarters, of population changes compared to economic status changes. The quarter of Hottingen came closest to being relevant, but even though those values might not seem too bad compared to the rest of the quarters, they are still not good by any means.

### Linear regression analysis over all of Zurich's quarters

As a last effort, we analyze the data of all of the quarters of Zurich. As already seen in the scatterplots above, there shouldn't be expected to much of a difference, and equivalently to the data of only the gentrified quarters, we got a very low R Square value, meaning next to nothing of the population change is explained by the economic status change. The significance value however decreased from 0.797 to 0.390, putting the results closer to being significant, which was, however, to be expected, since we considerably increased the number of data points we used for our regression. All in all we have to come to the conclusion that this dataset also can't support our hypothesis of a causality between the economic change and the population change in a quarter.

#### ANOVA^{a}

Model | Sum of Squares | df | Mean Square | F | Sig. |
---|---|---|---|---|---|

Regression | 3.470 | 1 | 3.470 | 0.741 | 0.390^{b} |

Residual | 2060.209 | 440 | 4.682 | ||

Total | 2063.679 | 441 |

a: Dependent Variable: change_population_perYear

b: Predictors: (Constant), change_economicStatus_perYear

#### Model Summary^{b}

Model | R | R Square | Adjusted R Square | Std. Error of the Estimate |
---|---|---|---|---|

0.041^{a} |
0.002 | -0.001 | 2.16386 |

a: Predictors: (Constant), change_economicStatus_perYear

b: Dependent Variable: change_population_perYear

### Summary

As already expected when looking at the first few charts of our results, we were unable to find any significant effect of economic status change compared to population change in Zurich's gentrified quarters. This even when analyzing smaller portions or bigger portions of Zurich's quarters, than the whole group of gentrified quarters. However, our method of spotting gentrified quarters seems to be not too bad. We found quarters like Seefeld and Mühlebach, where Seefeld is famous for being an example for gentrification and Mühlebach being right next to it, is also very plausible to be gentrified. The quarter of Escher Wyss, with its very big population increase, due to it being formerly an industrial quarter that became trendy to live in, is also well known for being an example of gentrification. Hottingen and Fluntern, lying next to each other on the "Zürichberg", are not that well known to us as being gentrified, but it could very well be possible, another explanation could be that those quarters on the traditionally more expensive "Zürichberg", show one of the limitations of our assumption, of big economic status increase being due to gentrification, whereas it could also be an increase in economic status of the existing population. This could in future work, be tried to counterbalance, by factoring in moving data to have not only the increase in economic status part of the gentrification definition, but also the replacement part.