Lecture 4-6 Flashcards
What is a spatial weight file?
Provides neighbourhood information of your map
Spatial Weight/Contiguity File
For contiguity based method that uses 1 or 0 to denote whether areas are neighbours or not, the contiguity information is sometimes stored in a file, which is referred to as a spatial weight file
Spatial Weights Matrix
the neighbouring connections are on the map and how important these are relative to one another
Defining Neighbours: Adjacency-based
Rook-based Contiguity - only shapes that touch common boundary line
Queen-based Contiguity-
share a common boundary point
Local Moran’s I
Shows the locations of spatial clusters aka. the clustering of the indicator at specific locations
The output is a map instead of a value.
Social Determinants of Health
the complex, integrated, and overlapping social structures and economic systems that are responsible for most health inequities
Proximal factors of Health
They have a direct impact on a health outcome
Distal factors of Health
Have an affect health through intermediary factors (non-direct)
Relationship of Distal and Proximal factors
the built environment in which individual lives and works-a distal factor-may have an impact on the amount of physical activity-a proximal factor-and thus on bodyweight
Hypothesis testing
Also known as significance testing. Test the extent to which a hypothesis might be significant.
What are the two things we need to do before Hypothesis Testing?
- Specify the meanings of all the variables involved in the hypothesis.
- Specify how we’ll measure the variables –> operationalization (specifying the exact operations involved in measuring a variable).
How is hypothesis testing done (three steps)?
- Define your variables: specify the meanings of all the variables involved.
- Specify how you’ll measure the variables.
- Analyze the data. (eg. using regression analysis)
When should we apply spatial regression and not classic regression?
Independence Assumption: use a map display OR Moran’s I to test for spatial patterns of the residuals
Spatial Error Regression differs from Classic Regression because….
- it has a map (in the form of a spatial weights file or matrix) that is added and required for the analysis needed in the data input - Allows for data dependence
- Also has a autoregressive coefficient
Relational statements Deterministic vs. Probabilistic
Deterministic:
- Two variables go together all of the time.
- If the variables are not related in this way, then the statement must be modified.
E.g., as the economy shrinks by 1 per cent, the arson rate increases by 5 per cent.
Probabilistic:
- Two variables go together with some degree of regularity.
- Not always related in that way.
E.g., people who smoke are more likely to get lung cancer.
What is an index?
Index is a composite measure
Composite measure is created by combining two or more separate indicators into a single (summary) measure.
Many phenomena are too complex to be captured properly by one variable
Multiple Spatial Regression
One dependant variable with two or more covariates (independent variables)
Correlation
The Event that process A is associated with event (process B)
Causation
The Event that process A leads to or causes event (process B)
Does Correlation mean Causality?
Correlations does not always mean Causation
What are two types of Causation?
Necessary (Y cannot happen without X) and sufficient (Y always happens when X happens)
Give two examples of Necessary Causation
- Being female is a necessary condition of becoming pregnant.
- Having a grocery store nearby is a necessary condition of walking to buy groceries.
Give two examples of Sufficient Causation
- Being female is not a sufficient cause of becoming pregnant. Otherwise, all women would get pregnant.
- Having a grocery store nearby is not a sufficient cause of walking to buy groceries.
