Week 1- Spine of statistics pt 1 Flashcards
What’s the spine of statistics
- Standard error
- Parameters
- Interval estimates
- Null hypothesis significance testing (NHST)
- Estimation
The General Linear Model
GLM
Outcome = (modeli) + errori Outcome = b0 + b1 predictori + … + errori
With no predictors how do you predict an outcome
From only the intercept b0
What are confidence intervals
Intervals that contain the ‘true’ population value of the parameter in 95% of samples
What are confidence intervals not
- They do not reflect confidence in the value of the population parameter
- There is not a 95% probability that a given interval contains the population value
Parameters define the
Model and represent hypotheses of interest
Parameters are estimated based on
Sample data
All parameters have
Interval estimates
Confidence intervals might tell us
About the true value of the parameter
Standard error informs …
The variability in parameter estimates from sample to sample
Significance tests and confidence intervals rely on
Standard error
Each predictor has its own
Parameter
If all the data is categorical which test should you use
Chi squared
A spearman correlation is
A non-parametric test
A Pearson correlation is
A parametric test
A Pearson correlation uses which type of data
Continuous
What does a t test do
Compare two means
What does a one way anova test do
Compare 3 means
Sum of squared error graphs look like
U shaped
What is used as an estimate of a parameter
A sample value
The estimate of a parameter will differ across
Samples
Standard error can be used to
Get an idea of the extent to which estimates differ across samples
What does 95% confidence interval mean
For 5% of the samples the confidence interval does not contain the mean in the population
In Central Limit theorem, the average of the sample means and standard deviations will equal
The population mean and standard deviation
In CLT, 95% of all values are within
+-1.96 SD from the central mean
If predictors are standardised the intercept ends up being
0
What do Greek b symbols mean
Predictors are standardised
What does the broom::glance output include
- R squared
- Adjusted r squared
We want the f ratio to be as
Large as possible