Difference between revisions of "An initial path towards statistical analysis"
Line 122: | Line 122: | ||
== Only continuous variables == | == Only continuous variables == | ||
+ | Your data is only continuous. Are there dependencies between the variables? | ||
<imagemap>Image:Statistics Flowchart - Continuous - Dependencies.png|650px|center| | <imagemap>Image:Statistics Flowchart - Continuous - Dependencies.png|650px|center| | ||
poly 328 376 12 696 332 1008 640 688 [[An_initial_path_towards_statistical_analysis#No_dependencies|No dependencies]] | poly 328 376 12 696 332 1008 640 688 [[An_initial_path_towards_statistical_analysis#No_dependencies|No dependencies]] | ||
Line 132: | Line 133: | ||
=== No dependencies: Correlations === | === No dependencies: Correlations === | ||
− | If there are no dependencies between your variables, you should do a Correlation. | + | '''If there are no dependencies between your variables, you should do a Correlation.''' |
A correlation ... ADD. | A correlation ... ADD. | ||
Check the entry on [[Correlations]] to learn more. | Check the entry on [[Correlations]] to learn more. | ||
Line 151: | Line 152: | ||
=== Clear dependencies === | === Clear dependencies === | ||
+ | You have one (or more) dependent variable(s) in your dataset EXPLAIN BETTER. Is the dependent variable normally distributed? | ||
<imagemap>Image:Statistics Flowchart - Dependent - Normal Distribution.png|650px|center| | <imagemap>Image:Statistics Flowchart - Dependent - Normal Distribution.png|650px|center| | ||
poly 328 376 12 696 332 1008 640 688 [[An_initial_path_towards_statistical_analysis#Linear_Regression|Linear Regression]] | poly 328 376 12 696 332 1008 640 688 [[An_initial_path_towards_statistical_analysis#Linear_Regression|Linear Regression]] | ||
Line 162: | Line 164: | ||
==== Normally distributed dependent variable: Linear Regression ==== | ==== Normally distributed dependent variable: Linear Regression ==== | ||
+ | If your dependent variable(s) is/are normally distributed, you should do a Linear Regression. A linear regression .. ADD | ||
+ | Check the entry ADD ?? | ||
+ | |||
+ | There may be one exception to a plain linear regression: if you have several predictors, you should go one step further... LINK TO NEXT LEVEL (SEE FLOWCHART) | ||
==== Not normally distributed dependent variable ==== | ==== Not normally distributed dependent variable ==== | ||
+ | The dependent variable(s) is/are not normally distributed. Which kind of distribution does it show, then? For both Binomial and Poisson distributions, your next step is the Generalised Linear Model. However, it is important that you select the proper distribution type in the GLM ADD MORE INFO | ||
<imagemap>Image:Statistics Flowchart - Dependent - Distribution type.png|650px|center| | <imagemap>Image:Statistics Flowchart - Dependent - Distribution type.png|650px|center| | ||
poly 328 376 12 696 332 1008 640 688 [[An_initial_path_towards_statistical_analysis#Generalised_Linear_Model|Generalised Linear Model]] | poly 328 376 12 696 332 1008 640 688 [[An_initial_path_towards_statistical_analysis#Generalised_Linear_Model|Generalised Linear Model]] | ||
Line 173: | Line 180: | ||
* ADD INFO - HOW DO I KNOW THE DISTRIBUTION TYPE? | * ADD INFO - HOW DO I KNOW THE DISTRIBUTION TYPE? | ||
* Check the entry on [[Data_distribution#Non-normal_distributions|Non-normal distributions]] to learn more. | * Check the entry on [[Data_distribution#Non-normal_distributions|Non-normal distributions]] to learn more. | ||
− | |||
==== Generalised Linear Models ==== | ==== Generalised Linear Models ==== | ||
− | With non-normally distributed data, you arrive at a Generalised Linear Model (GLM). GLMs are... ADD | + | '''With non-normally distributed data, you arrive at a Generalised Linear Model (GLM).''' GLMs are... ADD |
Depending on the existence of random variables, there is a distinction between Mixed Effect Models and Generalised Linear Models, which are based on regressions. | Depending on the existence of random variables, there is a distinction between Mixed Effect Models and Generalised Linear Models, which are based on regressions. | ||
Line 196: | Line 202: | ||
[Binomial GLM|Dependent variable is 0/1 or proportions]] | [Binomial GLM|Dependent variable is 0/1 or proportions]] | ||
− | |||
= Multivariate statistics = | = Multivariate statistics = |
Revision as of 10:02, 23 March 2021
This interactive page allows you to find the best statistical method to analyze your given dataset.
Go through the images step by step, click on the answers that apply to your data, and let the page guide you.
Start here with your data! This is your first question.
How do I know?
- Inspect your data with
str
orsummary
. Are there several variables? - What does the data show? Does the underlying logic of the data suggest dependencies between the variables?
Example: Inspecting the swiss dataset
Contents
Univariate statistics
So either have only one variable, or the variables are dependent on each other. But what kind of variables do you have?
How do I know?
- Check the entry on Data formats to understand the difference between categorical and numeric variables.
- Investigate your data using
str
orsummary
. integer and numeric data is not categorical, while factorial and character data is.
At least one categorical variable
Your dataset does not only contain continuous data. Is it only categorical, though?
How do I know?
- Investigate your data using
str
orsummary
. integer and numeric data is not categorical, while factorial and character data is.
Only categorical data: Chi Square Test
If you have only categorical variables, you should do a Chi Square Test. A Chi Square test... LINK TO CHI SQUARE TEST R EXAMPLE
Categorical and continuous data
Your dataset is a mix of categorical and continuous variables. How many factor levels ADD BETTER EXPLANATION
How do I know?
- R commands: quantile(), str, summary
- Investigate your categorical dependent variables using...
- ADD MORE
One or two factor levels: t-test
With one or two factor levels, you should do a t-test. A t-test ... ADD. Check the entry on the T-Test to learn more.
Depending on the variances in your data, the type of t-test differs.
ADAPT THE T-TEST ENTRY SO THAT DIFFERENCE BETWEEN STUDENT AND WELCH IS CLEAR
How do I know?
- Use an F-Test to check whether the variances of the datasets are equal. LINK LEFT BOX TO F-TEST
More than two factor levels
MISSING - COMPLICATED FIGURE
Analysis of Variance
R commands: aov, Anova, ezAnova, var.test(), lm
Relevant figures: boxplot()
Is your dependent variable normally distributed?
R commands: ks.test, shapiro.test, hist
Yes, my dependent variable is normally distributed!
No, my dependent variable is binomial distributed!
No, my dependent variable is Poisson distributed!
Gaussian Anova
R commands: aov, lm
Relevant figures: boxplot
Is your dependent variable binomial or Poisson
Poisson GLM|Dependent variable is count data
R commands: glm,
Relevant figures: plot
Binomial GLM|Dependent variable is 0/1 or proportion
R commands:
Relevant figures:
Type III Anova
R commands: Anova(car)
Relevant figures: boxplot
Data_distribution#Non-normal_distributions|Dependent variable not normally distributed]
Poisson GLM|Dependent variable is count data
R commands: glm
Relevant figures: plot
Binomial GLM|Dependent variable is 0/1 or proportions]
R commands: glm
Relevant figures:
Only continuous variables
Your data is only continuous. Are there dependencies between the variables?
How do I know?
- ADD INFO - HOW DO I KNOW IF THEY ARE DEPENDENT?
No dependencies: Correlations
If there are no dependencies between your variables, you should do a Correlation. A correlation ... ADD. Check the entry on Correlations to learn more. The type of correlation depends on your data distribution.
- ADD INFO ON PEARSON AND SPEARMAN CORRELATIONS; WITH R CODE
LINK TO CORRELATION R EXAMPLES (pearson, spearman)? How do I know?
- ADD INFO - HOW DO I KNOW IF THE DATA IS NORMALLY DISTRIBUTED?
- Check the entry on Normal distributions to learn more.
Clear dependencies
You have one (or more) dependent variable(s) in your dataset EXPLAIN BETTER. Is the dependent variable normally distributed?
How do I know?
- ADD INFO - HOW DO I KNOW IF THE DATA IS NORMALLY DISTRIBUTED?
- Check the entry on Normal distributions to learn more.
Normally distributed dependent variable: Linear Regression
If your dependent variable(s) is/are normally distributed, you should do a Linear Regression. A linear regression .. ADD Check the entry ADD ??
There may be one exception to a plain linear regression: if you have several predictors, you should go one step further... LINK TO NEXT LEVEL (SEE FLOWCHART)
Not normally distributed dependent variable
The dependent variable(s) is/are not normally distributed. Which kind of distribution does it show, then? For both Binomial and Poisson distributions, your next step is the Generalised Linear Model. However, it is important that you select the proper distribution type in the GLM ADD MORE INFO
How do I know?
- ADD INFO - HOW DO I KNOW THE DISTRIBUTION TYPE?
- Check the entry on Non-normal distributions to learn more.
Generalised Linear Models
With non-normally distributed data, you arrive at a Generalised Linear Model (GLM). GLMs are... ADD
Depending on the existence of random variables, there is a distinction between Mixed Effect Models and Generalised Linear Models, which are based on regressions.
How do I know?
- HOW DO I KNOW IF I HAVE RANDOM VARIABLES???
- R commands: glmer, glmmPQL
Relevant figures:
WHAT IS THIS ABOUT?
Dependent variable is count data
[Binomial GLM|Dependent variable is 0/1 or proportions]]
Multivariate statistics
How do I know?
- In an Ordination, you arrange your data alongside underlying gradients in the variables to see which variables most strongly define the data points.
- In a Cluster Analysis, you group your data points according to how similar they are, resulting in a tree structure.
- In a Network Analysis, you arrange your data in a network structure to understand their connections and the distance between individual data points.
Ordinations
You are doing an ordination. An ordination ... ADD Check the entry on Ordinations (to be added) to learn more.
There is a difference between ordinations for different data types - for abundance data, you use Euclidean distances, and for continuous data, you use Jaccard distances.
How do I know?
- Check the entry on Data formats to learn more about the different data formats.
- Investigate your data using
str
orsummary
. Abundance data is marked as FORMATNAME, and continuous data is marked as FORMATNAME.
MAKE THE STUFF BELOW CLEARER - OR LINK SOMEWHERE ELSE?
Linear-based ordinations
Linear-based ordinations are... It uses Euclidean distances, which is...
R commands:
Relevant figures:
Correspondance analysis
A correspondence analysis is... It uses Jaccard distances, which is...
R commands:
Relevant figures:
Cluster Analysis
So you decided for a Cluster Analysis. A Cluster Analysis .. ADD Check the entry on Clustering Methods to learn more.
There is a difference to be made here, dependent on whether you want to classify the data based on prior knowledge (supervised) or not (unsupervised). DIFFERENCE BETWEEN SUPERVISED AND UNSUPERVISED?? DISTINCTION NOT MADE IN THE CLUSTERING ENTRY
WHERE DO WE LINK?
How do I know?
- HOW DO I KNOW IF ITS SUPERVISED OR NOT?
Network Analysis
You have decided to do a Network Analysis. In a Network Analysis... ADD Check the entry on Social Network Analysis to learn more.
There is a distinction here between bipartite and tripartite networks, with two or three kinds of nodes, respectively.
How do I know?
- Check your data using the R code ADD CODE
ADD MORE BELOW - OR DO WE LINK SOMEWHERE ELSE? Bipartite If your data has two different kinds of nodes, your network is called a "bipartite" network.
R commands:
- is_bipartite(graph)
- make_bipartite_graph(types, edges, directed = FALSE)
Tripartite
R commands:
Relevant figures:
Resterampe
Analysis of Variance
INSERT TYPE II
INSERT RANDOM FACTOR
INSERT LMM
Dependent variable is count data
Dependent variable is 0/1 or proportions