lecture 5 summary

Question 1

Hypothesis test

Answer 1

checks whether the observed difference occurred randomly because of sampling error or whether it indicates a difference between the two samples

Question 2

Two sided test

Answer 2

states that the difference value must fall within one of the two end intervals

Question 3

A one sided test states that

Answer 3

the difference value must only fall within one of the two end intervals

Question 4

Type 1 error

Answer 4

a hypothesis test thats been declared positive while in reality its not true, a false positive

Question 5

Type 2 error

Answer 5

is a hypothesis test that has been rejected while in reality it is true, a false negative

Question 6

Chi-quare statistic

Answer 6

Is a fuction of the squares of the deviations of the observed counts n from their expected values (under some null hypothesis) E(n1) weighted by the reciprocals of their expected values:

with K-1 degrees of freedo

This test is mostly used y managers as a test for independence

Question 7

Regression analysis

Answer 7

is the quantification of the slope of a regression line. Furthermore, it is designed to estimate the influence of one variable (X) on another variable (Y)

R^2 expresses the proportion of the explained variance in the dependent variable (Y) that is explained by the regression line (value range: 0 to 1)

Question 8

We are interested in knowing how good our predictions are. FOr this

Answer 8

we use R quared. This is hte measure of the regression models ability to predict, also called the coefficient of determination of a model. it indicates how well a model explains the variance of the dependent variable

Question 9

R Formula

Answer 9

R^2 = (regression coefficient)^2 * Variance of X/Variance of Y)

Question 10

Regression analysis can be used for various purposes

Answer 10

explanation of relationships

stimulation of effects

Prediction

Identification of driving factors

Question 11

Linear regression takes sevel keky assumptions

Answer 11

Multiple linear regression requires at least two independent variables

There should be a linear rrelationship between the dependent and independent variables

The error term is normally distributed

No multicollinearity: multiple regression assumes that the independent variables are not highly correlated with each other

Homoscedasticity: this indicates the variance of error terms are similar across the values of the independent vairables

Sample size (a rule of thumb): regression analysis requires at least 20 cases per independent variable