4.4 Stats at Erasmus University Rotterdam | Flashcards & Summaries

Lernmaterialien für 4.4 stats an der Erasmus University Rotterdam

Greife auf kostenlose Karteikarten, Zusammenfassungen, Übungsaufgaben und Altklausuren für deinen 4.4 stats Kurs an der Erasmus University Rotterdam zu.

TESTE DEIN WISSEN
Regression formula
Lösung anzeigen
TESTE DEIN WISSEN
Y(pred.) = intercept(called b0) + coefficient of x1(called b1) * x1 + coefficient of x2(called b2) * x2 ... etc.
Lösung ausblenden
TESTE DEIN WISSEN
Regression - Creating a straight line.

Lösung anzeigen
TESTE DEIN WISSEN
y = b0 + b1xi +ei 
b1 = Rxy * Sy/Sx
b1 - Regression coefficient, slope of the line 
b0 - Intercept where Regression line crosses the y axis where x = 0
Lösung ausblenden
TESTE DEIN WISSEN

Quantitative Variables

Lösung anzeigen
TESTE DEIN WISSEN

Variables whose values result from counting or measuring something.

Examples: height, weight, time in the 100 yard dash, number of items sold to a shopper


- continuous

Lösung ausblenden
TESTE DEIN WISSEN

 Qualitative Variables

Lösung anzeigen
TESTE DEIN WISSEN

 Variables that are not measurement variables. Their values do not result from measuring or counting.

Examples: hair color, religion, political party, profession


- categorical 

Lösung ausblenden
TESTE DEIN WISSEN
Regression - what is it?

Lösung anzeigen
TESTE DEIN WISSEN
Hypothetical, linear model that predicts the value of one variable from another. The relationship is described using the equation of a straight line.
Simple - one predictor 
Multiple - more than one predictor 
Lösung ausblenden
TESTE DEIN WISSEN

regression with dummies 

Lösung anzeigen
TESTE DEIN WISSEN

Y = b0 + b1dummy1 + b2dummy2
Y = b0+ (b1xD1)+( b2xD2 )

lecture example:

b0 = mean of the no training group
b1 = difference between the mean of training A and the no training group
b2 = difference between the mean of training B and the no training group

Lösung ausblenden
TESTE DEIN WISSEN

GLM hypothesis 

null hypothesis

Lösung anzeigen
TESTE DEIN WISSEN

GLM hypotheses:

•General hypothesis:
•H0: R2= 0 (in the population) OR:
•H0: β1= β2= 0 (H0: μnotraining= μtrainingA= μtrainingB) 

•Specific hypotheses:
•H0: β1= 0 (μnotraining= μtrainingA)
•H0: β2= 0 (μnotraining= μtrainingB)

Lösung ausblenden
TESTE DEIN WISSEN

dummie scores 

Lösung anzeigen
TESTE DEIN WISSEN

Dummies have only two scores:
0 and 1

✓E.g. Gender →males= 0; females= 1

Youhave to choose one groupas a reference group (e.g. males), this group has a score of 0 on each dummy
•The remaining group sare each defined by one dummy: females 1 or if more groups then :

Dummy1: people who got training A have a score of 1, all others a 0.
Dummy2: people who got training B have a score of 1, all others a 0. 

0 -> no traning - as reference group in this case 

Lösung ausblenden
TESTE DEIN WISSEN

Moderation 

Lösung anzeigen
TESTE DEIN WISSEN

Definition moderation: the effect of predictor X on dependent variable Y depends on the value of third variable M, where M is the moderator.
Moderation analysis: test for external validity: how universal (generalizable) is the relationship between Xand Y?

Lösung ausblenden
TESTE DEIN WISSEN

Regression notation of a moderation-effect

Lösung anzeigen
TESTE DEIN WISSEN

Yi=b0+b1Xi+b2Mi+b3XMi+εi

The interaction XM, and thus coefficient b3, indicates the moderation-effect.
Coeffient b1 indicates the simple effect of X on Y. This is the effect of X on Y when M is 0.
Coeffient b2 indicates the simple effect of M on Y. This is the effect of M on Y when Xis 0.

Lösung ausblenden
TESTE DEIN WISSEN

Mediation

Lösung anzeigen
TESTE DEIN WISSEN

Definition mediation:  effect of a predictor variable X on outcome variable Y can be (partly-> partial mediation) explained by a third variable M. Variable M is the mediator.
a= direct effect of X on M
b= direct effect of M on Y
c= total effect of X on Y
c’=direct effect of X on Y


The mediation model is a causal model. Thus it is assumed that the mediator causes Y and not vice versa. If this assumption is false, than the interpretation is disputable.

Lösung ausblenden
TESTE DEIN WISSEN

How does bootstrapping work?

Lösung anzeigen
TESTE DEIN WISSEN

1.We draw a large number (here: 1000) of bootstrap samples of Ncases, from the original sample. We sample with replacement (this means cases can be drawn more than once or not once) In each sample we compute indirect effect ab

2.We make a histogram of all values of ab. Thisis the bootstrap sampling distribution of ab. It will be positively skewed.
3.We determine the 95% CI by reading of the 2.5th and 97.5th percentile in the distribution. NB. Het CI of ab is also skewed (asymmetrical).
4.This interval is corrected towards the right tail and the upper limit is streched out towards the right tail: Bias corrected and accelerated.

Lösung ausblenden
  • 3015 Karteikarten
  • 106 Studierende
  • 0 Lernmaterialien

Beispielhafte Karteikarten für deinen 4.4 stats Kurs an der Erasmus University Rotterdam - von Kommilitonen auf StudySmarter erstellt!

Q:
Regression formula
A:
Y(pred.) = intercept(called b0) + coefficient of x1(called b1) * x1 + coefficient of x2(called b2) * x2 ... etc.
Q:
Regression - Creating a straight line.

A:
y = b0 + b1xi +ei 
b1 = Rxy * Sy/Sx
b1 - Regression coefficient, slope of the line 
b0 - Intercept where Regression line crosses the y axis where x = 0
Q:

Quantitative Variables

A:

Variables whose values result from counting or measuring something.

Examples: height, weight, time in the 100 yard dash, number of items sold to a shopper


- continuous

Q:

 Qualitative Variables

A:

 Variables that are not measurement variables. Their values do not result from measuring or counting.

Examples: hair color, religion, political party, profession


- categorical 

Q:
Regression - what is it?

A:
Hypothetical, linear model that predicts the value of one variable from another. The relationship is described using the equation of a straight line.
Simple - one predictor 
Multiple - more than one predictor 
Mehr Karteikarten anzeigen
Q:

regression with dummies 

A:

Y = b0 + b1dummy1 + b2dummy2
Y = b0+ (b1xD1)+( b2xD2 )

lecture example:

b0 = mean of the no training group
b1 = difference between the mean of training A and the no training group
b2 = difference between the mean of training B and the no training group

Q:

GLM hypothesis 

null hypothesis

A:

GLM hypotheses:

•General hypothesis:
•H0: R2= 0 (in the population) OR:
•H0: β1= β2= 0 (H0: μnotraining= μtrainingA= μtrainingB) 

•Specific hypotheses:
•H0: β1= 0 (μnotraining= μtrainingA)
•H0: β2= 0 (μnotraining= μtrainingB)

Q:

dummie scores 

A:

Dummies have only two scores:
0 and 1

✓E.g. Gender →males= 0; females= 1

Youhave to choose one groupas a reference group (e.g. males), this group has a score of 0 on each dummy
•The remaining group sare each defined by one dummy: females 1 or if more groups then :

Dummy1: people who got training A have a score of 1, all others a 0.
Dummy2: people who got training B have a score of 1, all others a 0. 

0 -> no traning - as reference group in this case 

Q:

Moderation 

A:

Definition moderation: the effect of predictor X on dependent variable Y depends on the value of third variable M, where M is the moderator.
Moderation analysis: test for external validity: how universal (generalizable) is the relationship between Xand Y?

Q:

Regression notation of a moderation-effect

A:

Yi=b0+b1Xi+b2Mi+b3XMi+εi

The interaction XM, and thus coefficient b3, indicates the moderation-effect.
Coeffient b1 indicates the simple effect of X on Y. This is the effect of X on Y when M is 0.
Coeffient b2 indicates the simple effect of M on Y. This is the effect of M on Y when Xis 0.

Q:

Mediation

A:

Definition mediation:  effect of a predictor variable X on outcome variable Y can be (partly-> partial mediation) explained by a third variable M. Variable M is the mediator.
a= direct effect of X on M
b= direct effect of M on Y
c= total effect of X on Y
c’=direct effect of X on Y


The mediation model is a causal model. Thus it is assumed that the mediator causes Y and not vice versa. If this assumption is false, than the interpretation is disputable.

Q:

How does bootstrapping work?

A:

1.We draw a large number (here: 1000) of bootstrap samples of Ncases, from the original sample. We sample with replacement (this means cases can be drawn more than once or not once) In each sample we compute indirect effect ab

2.We make a histogram of all values of ab. Thisis the bootstrap sampling distribution of ab. It will be positively skewed.
3.We determine the 95% CI by reading of the 2.5th and 97.5th percentile in the distribution. NB. Het CI of ab is also skewed (asymmetrical).
4.This interval is corrected towards the right tail and the upper limit is streched out towards the right tail: Bias corrected and accelerated.

4.4 stats

Erstelle und finde Lernmaterialien auf StudySmarter.

Greife kostenlos auf tausende geteilte Karteikarten, Zusammenfassungen, Altklausuren und mehr zu.

Jetzt loslegen

Das sind die beliebtesten 4.4 stats Kurse im gesamten StudySmarter Universum

IM/Stats

TU München

Zum Kurs
STS

University of the East, Ramon Magsaysay Memorial Medical Center

Zum Kurs
STS

Lyceum of the Philippines University

Zum Kurs

Die all-in-one Lernapp für Studierende

Greife auf Millionen geteilter Lernmaterialien der StudySmarter Community zu
Kostenlos anmelden 4.4 stats
Erstelle Karteikarten und Zusammenfassungen mit den StudySmarter Tools
Kostenlos loslegen 4.4 stats