Multiple Linear Regression

?
What does multivariate mean?
More than one variable (constructs do not have just a single predictor)
1 of 18
Why do we use an mlr?
To see the correlation between two or more variables - we can take into account other variables
2 of 18
Why else would we use an mlr?
It can tell us about the performance of a theoretical model overall - and it can tell us about the variability that one predictor has on an outcompetes while the other predictors in the model are controlled
3 of 18
How can we check if the model significantly predicts change in my outcome variable?
We check this by comparing the variance explains and the variance unexplained (residuals) in the outcome variable.
4 of 18
What can we tell if the variance explained is significantly different to the residuals?
If the variance explained is significantly different to the residuals (the variance unexplained) then we can say that the theoretical model significantly predicts change in the outcome variable. We check this with A
5 of 18
How do you report the significance as an ANOVA?
Report as ANOVA(Fdf1,df2=F-statistic,p=p-value) e.g. F(3,223)=29.00, P<0.001
6 of 18
How much of the variability in the outcome variable is predicted by my theoretical model?
Check this with the R Square (R2) or R square adjusted
7 of 18
When do we quote the adjusted R square?
For smaller samples
8 of 18
How do we calculate whether to use R square or R square adjusted?
Calculate 50+8*(number of predictors) - this is critical value - then check your sample size- if sample size is greater than or equal to critical - use R2 but if less R2 adjusted
9 of 18
What would (for example) R square .281 mean?
that means the 28.1% of the variability in the exam score is explained by the theoretical model (the predictor variables together)
10 of 18
If i increase one of the predictors will my outcome variable increase or decrease (or does it have no effect)?
If the predictor variable is significant - means that for a one unit increase in the predictor variable there is a <unstandardized beta value>> increase/decrease in the outcome variable
11 of 18
What do we look at for the significance?
the coefficients table - B and Sig.
12 of 18
What does the unstandardised beta show?
Predicting values of the criterion variable using the Betas associated with each of the significant predictors. e.g. with a 1.298 b value - for a 1 unit increase in attendance there is an increase of 1.30 in exam score (from 54% to 55.3% e.g.)
13 of 18
What else does the unstandarized beta value refer to?
The values in the regression equation
14 of 18
What is the regression equation?
Y(predicted)=b0 + b1(x1)+ b2(x2) + b3(x3) + error
15 of 18
Which of the significant predictors is the strongest predictor of the variability in my outcome variability?
To compare the effects of different predictor variables and to know the strength of the relationship- look at standardised beta values- for one standardised unit increase in predictor variable there is a increase/decrease in outcome variable
16 of 18
What is standardisation?
one standard deviation rise in the predictor variable leads to a <<standardised beta value>> standard deviation increase/decrease in the outcome variable
17 of 18
What does the standardized coefficients Beta show?
e.g. attendance is strongest predictor of exam scores with highest standardised value of 0.27- if below 0.3 - strength is seen as weak
18 of 18

Other cards in this set

Card 2

Front

Why do we use an mlr?

Back

To see the correlation between two or more variables - we can take into account other variables

Card 3

Front

Why else would we use an mlr?

Back

Preview of the front of card 3

Card 4

Front

How can we check if the model significantly predicts change in my outcome variable?

Back

Preview of the front of card 4

Card 5

Front

What can we tell if the variance explained is significantly different to the residuals?

Back

Preview of the front of card 5
View more cards

Comments

No comments have yet been made

Similar Psychology resources:

See all Psychology resources »See all Visual System resources »