WILD 7150 Midterm Exam Study Guide
What is the difference between truth and fact - ANSWER Truth: as it is found in
nature
Fact: an accumulation of data in search of truth
define p-value - ANSWER probability of getting a more extreme result than what
is observed if null hypothesis is true
Two possible causes of non-significant p-value - ANSWER small sample size and
small effect size
Two attributes are attained when calculating the betas for a general linear
model - ANSWER average residual=zero and lowest sum of squares possible
5 assumptions of general linear models - ANSWER 1. linearity
2. homoscedasticity
3. no auto-correlation
4. normally distributed residuals
5. continuous y
Difference between using an x variable as continuous vs categorical - ANSWER
continuous: linear
categorical: no linear assumption
define r^2 - ANSWER proportion of y explained by x
Purpose of post hoc tests - ANSWER an attempt to balance error and power
while maintaining experiment wide type one error
6 reasons to use multivariate model - ANSWER 1. elegance
2. collinearity
3. pseudoreplication
4. interactions
5. swamping
6. blocking
Two symptoms of collinearity among your x variable - ANSWER betas change
dramatically and model p-value is significant but individual p values are not
If you have two confounding x-variables and you leave one out of your model,
what is the negative consequence? - ANSWER bias in your estimates
What is the difference between truth and fact - ANSWER Truth: as it is found in
nature
Fact: an accumulation of data in search of truth
define p-value - ANSWER probability of getting a more extreme result than what
is observed if null hypothesis is true
Two possible causes of non-significant p-value - ANSWER small sample size and
small effect size
Two attributes are attained when calculating the betas for a general linear
model - ANSWER average residual=zero and lowest sum of squares possible
5 assumptions of general linear models - ANSWER 1. linearity
2. homoscedasticity
3. no auto-correlation
4. normally distributed residuals
5. continuous y
Difference between using an x variable as continuous vs categorical - ANSWER
continuous: linear
categorical: no linear assumption
define r^2 - ANSWER proportion of y explained by x
Purpose of post hoc tests - ANSWER an attempt to balance error and power
while maintaining experiment wide type one error
6 reasons to use multivariate model - ANSWER 1. elegance
2. collinearity
3. pseudoreplication
4. interactions
5. swamping
6. blocking
Two symptoms of collinearity among your x variable - ANSWER betas change
dramatically and model p-value is significant but individual p values are not
If you have two confounding x-variables and you leave one out of your model,
what is the negative consequence? - ANSWER bias in your estimates