Methodology & Statistics

Collection by The Sociological Cinema • Last updated 2 weeks ago

209 
Pins
 • 
9.59k 
Followers
The Sociological Cinema
Methodology: Epistemological Concerns
Methodology: Data Visualization
Methodology: Correlation and Causation
Methodology: Logical Fallacies
Methodology: Sampling
Other Pins

Methodology: Epistemological Concerns

A counterfactual is a statement that follows this form: "If Joseph Swan had not invented the modern incandescent light bulb, then someone else would have invented it anyway." People use counterfactuals every day; however, its analysis is not so clear. Swan, after all, did invent the modern incandescent light bulb, so how can the statement be true, if it is impossible to examine its correspondence to reality? ...

Counterfactuals

Counterfactuals - A counterfactual is a statement that follows this form: "If Joseph Swan had not invented the modern incandescent light bulb, then someone else would have invented it anyway." People...

A: Do you think the owl is a predator? B: No, of course not. He's never bothered me. A: Exactly. A: No idea what Mr. Mouse was going on about.  Artist: Nathan W. Pyle

A: Do you think the owl is a predator? B: No, of course not. He's never bothered me. A: Exactly. A: No idea what Mr. Mouse was going on about. Artist: Nathan W. Pyle

1) Metaphysics (What is real?); 2) Epistemology (How do we know?);  3) Axiology (What is right?)

The Three Areas of Philosophy Directly Influence Teaching, Theory, and Practice

1) Metaphysics (What is real?); 2) Epistemology (How do we know?); 3) Axiology (What is right?)

Philosophy of Science - Realism vs. Anti-Realism

Ontology - What can we know? (Realism - Relativism) Epistemology - How can we know? (Positivism - Interpretivism) Methodology How can we find out? (Quantitative - Qualitative) Method - What tool to use? (experiments - interviews)

Research as Continuum

Ontology - What can we know? (Realism - Relativism) Epistemology - How can we know? (Positivism - Interpretivism) Methodology How can we find out? (Quantitative - Qualitative) Method - What tool to use? (experiments - interviews)

A: "Can we ever be certain an observation is true?" B: "Yep." A: "How?" B: "Lookin'."  Source: SMBC Comics

A: "Can we ever be certain an observation is true?" B: "Yep." A: "How?" B: "Lookin'." Source: SMBC Comics

Positivism and Interpretivism in Social Research  Relationship between society and the individual - Society shapes the individual - "Society consists of 'social facts' which exercise coercive control over individuals' (Durkheim).  - Individuals have consciousness and are not just puppets who react to external social forces as Positivists believe.   Source: @UCPSociology

Positivism and Interpretivism in Social Research Relationship between society and the individual - Society shapes the individual - "Society consists of 'social facts' which exercise coercive control over individuals' (Durkheim). - Individuals have consciousness and are not just puppets who react to external social forces as Positivists believe. Source: @UCPSociology

A: "Patsy, what's wrong?" Patsy: "Positivists."

A: "Patsy, what's wrong?" Patsy: "Positivists."

The Münchhausen Trilemma, also called Agrippa's trilemma, purports that it is impossible to prove any certain truth even in fields such as logic and mathematics. According to this argument, the proof of any theory rests either on circular reasoning, infinite regress, or unproven axioms.

Epistemology - Münchhausen Trilemma

The Münchhausen Trilemma, also called Agrippa's trilemma, purports that it is impossible to prove any certain truth even in fields such as logic and mathematics. According to this argument, the proof of any theory rests either on circular reasoning, infinite regress, or unproven axioms.

Methodology: Data Visualization

What's wrong with this pie chart?

Fox News routinely provides shining examples of how NOT to present data. The values on the Y-axis are moving in the wrong direction, giving the impression that enrollment in in the ACA is somehow decreasing. Also, the unused space just below the title helps give the impression that enrollment has always been low.

Fox News routinely provides shining examples of how NOT to present data. The values on the Y-axis are moving in the wrong direction, giving the impression that enrollment in in the ACA is somehow decreasing. Also, the unused space just below the title helps give the impression that enrollment has always been low.

Fox News demonstrates how to make 4 percentage points look like 80. Don't do this.

Fox News demonstrates how to make 4 percentage points look like 80. Don't do this.

Sustainability practitioners have long relied on images to display relationships in complex adaptive systems on various scales and across different domains. These images facilitate communication, learning, collaboration and evaluation as they contribute to shared understanding of systemic processes...The research identifies, defines and illustrates 16 key features of complex systems and contributes to an evolving visual language of complexity. | Source: Boehnert, J., et al, 2018.

(3 of 3) The Visual Representation of Complexity: Definitions, Examples & Learning Points

Sustainability practitioners have long relied on images to display relationships in complex adaptive systems on various scales and across different domains. These images facilitate communication, learning, collaboration and evaluation as they contribute to shared understanding of systemic processes...The research identifies, defines and illustrates 16 key features of complex systems and contributes to an evolving visual language of complexity. | Source: Boehnert, J., et al, 2018.

Sustainability practitioners have long relied on images to display relationships in complex adaptive systems on various scales and across different domains. These images facilitate communication, learning, collaboration and evaluation as they contribute to shared understanding of systemic processes...The research identifies, defines and illustrates 16 key features of complex systems and contributes to an evolving visual language of complexity. | Source: Boehnert, J., et al, 2018.

(2 of 3) The Visual Representation of Complexity: Definitions, Examples & Learning Points

Sustainability practitioners have long relied on images to display relationships in complex adaptive systems on various scales and across different domains. These images facilitate communication, learning, collaboration and evaluation as they contribute to shared understanding of systemic processes...The research identifies, defines and illustrates 16 key features of complex systems and contributes to an evolving visual language of complexity. | Source: Boehnert, J., et al, 2018.

Sustainability practitioners have long relied on images to display relationships in complex adaptive systems on various scales and across different domains. These images facilitate communication, learning, collaboration and evaluation as they contribute to shared understanding of systemic processes...The research identifies, defines and illustrates 16 key features of complex systems and contributes to an evolving visual language of complexity. | Source: Boehnert, J., et al, 2018.

(1 of 3) The Visual Representation of Complexity: Definitions, Examples & Learning Points

Sustainability practitioners have long relied on images to display relationships in complex adaptive systems on various scales and across different domains. These images facilitate communication, learning, collaboration and evaluation as they contribute to shared understanding of systemic processes...The research identifies, defines and illustrates 16 key features of complex systems and contributes to an evolving visual language of complexity. | Source: Boehnert, J., et al, 2018.

The Anatomy of a Box and Whisker Plot

Source: www.zapbi.com

Visual Analytics: the Official ZAP Chart Type Cheat Sheet

Source: www.zapbi.com

Source: Anna Vital

How to Think Visually: Using Visual Analogies

Source: Anna Vital

Methodology: Correlation and Causation

Headline: "Women who own horses live longer"  Implied correlation: horses make you live longer.  Reality: if you own a horse, you can probably afford health insurance.  ~ @Justinkingnews

Headline: "Women who own horses live longer" Implied correlation: horses make you live longer. Reality: if you own a horse, you can probably afford health insurance. ~ @Justinkingnews

A Type of Continuous Probability  Source: Does anyone know who created this?

Normal Distribution

A Type of Continuous Probability Source: Does anyone know who created this?

Generalized Linear Models (GLM) extend the ordinary linear regression and allow the response variable y to have an error distribution other than the normal distribution.  GLMs are: a) Easy to Understand b) Simple to fit and interpret in any statistical package c) Sufficient in a lot of practical application  1. Linear Regression 2. Logistic Regression 3. Poisson Regression  Source: Marketing Distillery

Three Regression Types

Generalized Linear Models (GLM) extend the ordinary linear regression and allow the response variable y to have an error distribution other than the normal distribution. GLMs are: a) Easy to Understand b) Simple to fit and interpret in any statistical package c) Sufficient in a lot of practical application 1. Linear Regression 2. Logistic Regression 3. Poisson Regression Source: Marketing Distillery

Source: Bachman, R. D. The Practice of Research in Criminology and Criminal Justice.

Spurious Relationships

Source: Bachman, R. D. The Practice of Research in Criminology and Criminal Justice.

ASSUMPTIONS - For a successful regression analysis, it's essential to validate these assumptions. 1. Linearity: The relationship between dependent and independent variables should be linear. 2. Homoscedasticity: (constant variance) of the errors should be maintained. 3. Multivariate Normality: Multiple regression assumes that the residuals are normally distributed. 4. Lack of Multicollinearity: It is assumed that there is little or no multicollinearity in the data. Source: Avik Jain

(2 of 2) Multiple Linear Regression

ASSUMPTIONS - For a successful regression analysis, it's essential to validate these assumptions. 1. Linearity: The relationship between dependent and independent variables should be linear. 2. Homoscedasticity: (constant variance) of the errors should be maintained. 3. Multivariate Normality: Multiple regression assumes that the residuals are normally distributed. 4. Lack of Multicollinearity: It is assumed that there is little or no multicollinearity in the data. Source: Avik Jain

Multiple linear regression attempts to model the relationship between two or more features and a response by fitting a linear equation to observed data. The steps to perform multiple linear regression are almost similar to that of simple linear regression. The difference lies in the evaluation. You can use it to find out which factor has the highest impact on the predicted output and how different variables relate to each other.  Source: Avik Jain

(1 of 2) Multiple Linear Regression

Multiple linear regression attempts to model the relationship between two or more features and a response by fitting a linear equation to observed data. The steps to perform multiple linear regression are almost similar to that of simple linear regression. The difference lies in the evaluation. You can use it to find out which factor has the highest impact on the predicted output and how different variables relate to each other. Source: Avik Jain

Artist: xkcd

p-values and their interpretations

Artist: xkcd

True or false: If a scientific study finds a correlation between poverty and crime, it proves that poverty causes crime.  Source: Cards Against Humanity

Cards Against Humanity survey shows how absurdly inconsistent Americans can be with their political views

The findings highlight some of the hilarious yet troubling inconsistencies in Americans' political views, on the Democratic and the Republican sides.

Methodology: Logical Fallacies

Favoring action over inaction, often reflecting an urge to *do something*. Although it may satisfy the obligation of beneficence, it is more likely to violate the obligation of non-maleficence or *first, do no harm*.

Commission Bias

Favoring action over inaction, often reflecting an urge to *do something*. Although it may satisfy the obligation of beneficence, it is more likely to violate the obligation of non-maleficence or *first, do no harm*.

"I'm sorry, Jeannie, your answer was correct, but Kevin shouted his incorrect answer over yours, so he gets the points."  Artist: Joe Dator

"I'm sorry, Jeannie, your answer was correct, but Kevin shouted his incorrect answer over yours, so he gets the points." Artist: Joe Dator

Methodology: Sampling

Statistical tests calculate whether the difference between two sets of results is significant enough for us to be confident that any changes seen in the dependent are due to manipulation of the independent variable.  The choice of statistical tests depends on three factors: 1. What type of data is it? 2. Which experimental design is being used? 3. Are you testing for difference or correlation?  Source: Daydream

Selecting a Statistical Test

Statistical tests calculate whether the difference between two sets of results is significant enough for us to be confident that any changes seen in the dependent are due to manipulation of the independent variable. The choice of statistical tests depends on three factors: 1. What type of data is it? 2. Which experimental design is being used? 3. Are you testing for difference or correlation? Source: Daydream

Source: Pew Research Center

After brief plateau, telephone survey response rates have fallen again

Source: Pew Research Center

Other Pins

In just 15 days the total number of #COVID19 cases in Georgia is up 49%, but you wouldn't know it from looking at the state's data visualization map of cases. The first map is July 2. The second is today. Do you see a 50% case increase? Can you spot how they're hiding it?  ~ @andishehnouraee

Can you spot how Georgia is hiding the increase in COVID-19 cases? July 2020

In just 15 days the total number of #COVID19 cases in Georgia is up 49%, but you wouldn't know it from looking at the state's data visualization map of cases. The first map is July 2. The second is today. Do you see a 50% case increase? Can you spot how they're hiding it? ~ @andishehnouraee

What's wrong with this pie chart?

Scenario 1: The police officer stops 20 people pulling aside equal numbers of Black and white people. Of the 20 people stopped, the officer uses force against 8 of them. The police officer used force against stopped white people and stopped Black people at the same rate: 40%. But that's not the only scenario that can lead to that 40% number.  Scenario 2: This time, of the 100 people the officer sees, he stops 50. But this time he is biased in whom he pulls aside...

How numbers that appear equitable can obscure bias

Graphics by Jasmine Mithani Across the U.S., demonstrators have spent the past few weeks protesting against racial disparities in the country’s criminal justice…

Headline: "Women who own horses live longer"  Implied correlation: horses make you live longer.  Reality: if you own a horse, you can probably afford health insurance.  ~ @Justinkingnews

Headline: "Women who own horses live longer" Implied correlation: horses make you live longer. Reality: if you own a horse, you can probably afford health insurance. ~ @Justinkingnews

Favoring action over inaction, often reflecting an urge to *do something*. Although it may satisfy the obligation of beneficence, it is more likely to violate the obligation of non-maleficence or *first, do no harm*.

Commission Bias

Favoring action over inaction, often reflecting an urge to *do something*. Although it may satisfy the obligation of beneficence, it is more likely to violate the obligation of non-maleficence or *first, do no harm*.

1. Nominal Data 2. Ordinal Data 3. Discrete Data 4. Continuous Data  Source: intellspot.com

Types of Data

1. Nominal Data 2. Ordinal Data 3. Discrete Data 4. Continuous Data Source: intellspot.com

Fox News routinely provides shining examples of how NOT to present data. The values on the Y-axis are moving in the wrong direction, giving the impression that enrollment in in the ACA is somehow decreasing. Also, the unused space just below the title helps give the impression that enrollment has always been low.

Fox News routinely provides shining examples of how NOT to present data. The values on the Y-axis are moving in the wrong direction, giving the impression that enrollment in in the ACA is somehow decreasing. Also, the unused space just below the title helps give the impression that enrollment has always been low.

A Type of Continuous Probability  Source: Does anyone know who created this?

Normal Distribution

A Type of Continuous Probability Source: Does anyone know who created this?

Fox News demonstrates how to make 4 percentage points look like 80. Don't do this.

Fox News demonstrates how to make 4 percentage points look like 80. Don't do this.

Source: BarCharts

Methods of Assigning Probability

Source: BarCharts

Source: BarCharts

Probability Rules

Source: BarCharts

Source: BarCharts

Probability Definitions

Source: BarCharts

Introductory Concepts in Probability

These are the resources I’ve found on the internet while I was learning. A few of the mind maps I’ve made for my own understanding while a lot of other infographics I just uploaded from my hard disk…

Source: Thalassa Studios

Probability and Statistics Symbols

These are the resources I’ve found on the internet while I was learning. A few of the mind maps I’ve made for my own understanding while a lot of other infographics I just uploaded from my hard disk…

Statistical tests calculate whether the difference between two sets of results is significant enough for us to be confident that any changes seen in the dependent are due to manipulation of the independent variable.  The choice of statistical tests depends on three factors: 1. What type of data is it? 2. Which experimental design is being used? 3. Are you testing for difference or correlation?  Source: Daydream

Selecting a Statistical Test

Statistical tests calculate whether the difference between two sets of results is significant enough for us to be confident that any changes seen in the dependent are due to manipulation of the independent variable. The choice of statistical tests depends on three factors: 1. What type of data is it? 2. Which experimental design is being used? 3. Are you testing for difference or correlation? Source: Daydream

Generalized Linear Models (GLM) extend the ordinary linear regression and allow the response variable y to have an error distribution other than the normal distribution.  GLMs are: a) Easy to Understand b) Simple to fit and interpret in any statistical package c) Sufficient in a lot of practical application  1. Linear Regression 2. Logistic Regression 3. Poisson Regression  Source: Marketing Distillery

Three Regression Types

Generalized Linear Models (GLM) extend the ordinary linear regression and allow the response variable y to have an error distribution other than the normal distribution. GLMs are: a) Easy to Understand b) Simple to fit and interpret in any statistical package c) Sufficient in a lot of practical application 1. Linear Regression 2. Logistic Regression 3. Poisson Regression Source: Marketing Distillery

Pinterest