The type of analysis that simply describes the data you have, and includes examples of these kinds of measures.
Descriptive Statistics
-Frequency distribution (think histogram/bar chart),
-measure of central tendencies (mean, median, mode),
-measures of dispersion - (standard deviation) are examples
This statistical measure is often quoted to make an argument that two things are related, but in reality it can be spurious - I could use it to argue connect snake bites and applesauce consumption.
Correlation coefficient
This single value will tell you whether or not ANY independent variable in the model demonstrates a statistically significant relationship with the dependent.
It is the first thing you should look at when reading regression results
What is F-Score and associated p-value - read it just as a t-score/p-value. - if over .05 (using the standard break point of 95% confidence), something is statistically significant. The question is, what?
that's when you go to the individual variables and look at each one
This term describes when you have included everything in your regression model and excluded everything that doesn't belong there.
What is a fully specified model
It is used when the research study focuses on testing hypotheses; or making inferences about something (such as the relationship between a mentoring program and academic achievement or voting turnout and location of polling places) based on a sample of data.
Inferential Statistics
This test is used if you want to know if there is a difference between two groups where one variable is nominal (from a two parent household or not) or ordinal (high income/low income) and the other is interval (average test scores), what test must be used?
T-test
(in options on the computer, use independent samples - it is the most rigorous, so if there is stat significance found with it, there will be stat significance with all the other t-test types.)
This value tells you what % of variation in your Dependent variable (the y) that is explained by the whole model (all the independent variables together).
It is the 2nd thing to look at when looking at results.
What is R -squared? It runs from 0 to 100%.
What is no measurement error?
It refers to analysis of one variable.
Univariate analysis
One-sample t-test.
These values tell you, variable by variable, which independent variables are statistically significant in terms of their relationship to the dependent variable.
What are the t-scores and associated p-values?
The best song by the B-52s
What is Rock Lobster
(although Love Shack is also acceptable)
It refers to analysis of two variables such as independent and dependent variables
Bivariate analysis
This statistical test is used to understand if there is a statstically significant difference when the data are expressed in terms of frequencies or percentages (that is, when you are using only nominal variables for both variables).
Chi-Square Test for Independence
In the hit show Breaking Bad, Jesse makes this mistake in disposing of a body early in the series.
What is Doesn't use the correct container to try to dissolve the body, making a huge mess?
These two things describe the important qualities of the data for a good regression analysis
It involves analysis of the multiple relations between multiple variables.
Multivariate analysis
The two bands with hits in the top 10 of all songs between 2010-2020 that included the word "feeling" in the title
Justin Timberlake – Can't Stop the Feeling!
Black Eyed Peas – I Gotta Feeling. ..
This represents the exact relationship between the independent and dependent variables. You describe it by saying "when X increased by 1, Y goes up (or down) by this amount."
the SLOPE!
If the X variable is not stat sig, then you ignore this -
unstandardized beta coeffecient - the number in front of x in a regression equation.
This is the only kind of error you want in your model
What is Random error -
error should only be error
error should be error should be error