QUESTION 1

For the binomial GLM in R:

		The default link function is the identity
		The default link function is the logistic
		The default link function tends to its extreme values faster than the cauchit link function
		Link functions need to restrict the dependent variable to being non-negative
		The default link function is the probit
		Link functions need to restrict the dependent variable to being between zero and one
		A possible link function is the logistic
		A possible link function is the identity
		A possible link function is the probit

1 points

QUESTION 2

What would be the best name for the experimental design for this experiment on the effects of feral predator management on native mammal populations?

The aim of the experiment is to investigate whether baiting for feral predators will increase native mammal populations. Twenty locations are chosen at random within a large national park. Two separate sites are chosen at each location. At each location, one of the two sites is randomly chosen as a baiting site, where baits targeting feral predators are employed. At each of the forty sites a single camera trap is deployed. At the end of a certain period, the number of observations ofnative mammals is recorded for each trap.

		Fully randomised design
		Graphical design
		Block design with nesting
		Block design
		Split-plot design
		Nested design

1 points

QUESTION 3

Over-dispersion in the binomial GLM:

		Is indicated when the AIC is much greater than one
		Is indicated when the null degrees of freedom is much greater than the null deviance
		Is indicated when the residual deviance is much greater than the residual degrees of freedom
		Is indicated when the dispersion parameter in a quasibinomial GLM is much greater than one
		Is indicated when the residual degrees of freedom is much greater than the residual deviance
		Is indicated when the dispersion parameter in a quasibinomial GLM is much less than one
		Is indicated when the null deviance is much greater than the null degrees of freedom
		Should always be checked for and accounted for if present

1 points

QUESTION 4

Match the following types of model with the kind of over-dispersion they deal with.

	Binomial GLM
+	Poisson GLM
	Quasibinomial GLM
	Quasipoisson GLM
	beta-binomial GLM
	negative-binomial GLM

A.	over-dispersion in count data
B.	over-dispersion in binomial data
C.	doesnt deal with over-dispersion

1 points

QUESTION 5

For the Poisson GLM in R:

		The default link function can cause problems by predicting negative counts for some values of the explanatory variable(s)
		The default link function gives an exponential relationship between the explanatory variable and the dependent variable
		The default link function gives a linear relationship between the explanatory variable and the dependent variable
		The default link function is the log
		The default link function is the identity
		The default link function can cause problems by predicting very large counts for some values of the explanatory variable(s)

1 points

QUESTION 6

What would be the best name for the experimental design for this experiment on the effects of feral predator management on native mammal populations?

The aim of the experiment is to investigate whether baiting for feral predators will increase native mammal populations. Twenty locations are chosen at random within a large national park. Ten of these locations are randomly chosen as baiting sites, where baits targeting feral predators are employed. At each location a single camera trap is deployed. At the end of a certain period, the number of observations ofnative mammals is recorded for each trap.

		Fully randomised design
		Nested design
		Block design
		Split-plot design
		Block design with nesting
		Graphical design

1 points

QUESTION 7

What would be the best name for the experimental design for this experiment on the effects of feral predator management on native mammal populations?

The aim of the experiment is to investigate whether baiting for feral predators and olfactory deterrents will increase native mammal populations. Twenty locations are chosen at random within a large national park.At each location, four separate sites were allocated. One of the four sites at each location was randomly selected as a baiting site, where baits targeting feral predators were employed. Another of the four sites at each location was randomly selected as a deterrent site, where a olfactory deterrent was used. At a third randomly selected site at each location, both baiting and olfactory deterrents were employed, and at the fourth site no management was employed. At each site a single camera trap was deployed at different positions. At the end of a certain period, the number of observations ofnative mammals is recorded for each trap.

		Split-plot design
		Graphical design
		Nested design
		Block design with nesting
		Block design
		Fully randomised design

1 points

QUESTION 8

Which of the following models are more prone to over-fitting than the others?

		Binomial generalised linear models
		Linear models
		Binomial generalised additive models
		Poisson generalised additive models
		Poisson generalised linear models

1 points

QUESTION 9

Match the following type of data with an approach well-suited to deal with it:

	Nested data eg. species counts within multiple quadrats along multiple transects at Multiple sites.
	Data from a survey-type study where I have many possible explanatory variables for each data point, and these possible explanatory variables have not been fully controlled for.
	Data where I have many dependent variables for each data point.
	Data where I have many variables measured for each data point, a relatively large number of data points, and some clear hypotheses about which variables are likely to affect or depend on other variables.
	Data that show non-linear responses to the explanatory variable, when I have no particular hypothesis about the functional form of the relationship.
	Data that show non-linear responses to the explanatory variable, when I have a particular hypothesis about the functional form of the relationship.
	Data on the time to germination of a large number of seeds of two different species that both take a relatively long time to germinate
	Data on the time for which different individuals of two species were able to survive

A.	multiple regression or regression trees
B.	Non-linear regression
C.	GLM with Gamma distribution error
D.	GAMs
E.	multivariate techniques
F.	mixed effects models
G.	structural equation modelling

1 points

QUESTION 10

What would be the best name for the experimental design for this experiment on the effects of feral predator management on native mammal populations?

The aim of the experiment is to investigate whether baiting for feral predators will increase native mammal populations. Twenty locations are chosen at random within a large national park. Ten of these are randomly chosen as baiting sites, where baits targeting feral predators are employed. At each site a single camera trap was deployed for a week at the start of each month, for five months. At the end of each week of deployment, the number of observations ofnative mammals is recorded for each trap.

		Block design
		Nested design
		Fully randomised design
		Split-plot design
		Block design with nesting
		Graphical design

1 points

QUESTION 11

Treating all observations as independent replicates would be a case of pseudo-replication in which of the following experimental designs?

		Block design with nesting
		Graphical design
		Nested design
		Block design
		Split-plot design
		Fully randomised design

1 points

QUESTION 12

What would be the best name for the experimental design for this experiment on the effects of feral predator management on native mammal populations?

The aim of the experiment is to investigate whether baiting for feral predators and olfactory deterrents will increase native mammal populations. Twenty locations are chosen at random within a large national park. At each location, two separate sites were allocated, and one of the sites randomly chosen as a baiting site, where baits targeting feral predators were employed. Within each of the forty sites, whether baited or not, two separate areas were defined. One of the two areas was randomly selected at each site, for deployment of an olfactory deterrent. Within each of 80 defined areas, a single camera trap was deployed. At the end of a certain period, the number of observations ofnative mammals is recorded for each trap.

		Block design with nesting
		Split-plot design
		Block design
		Graphical design
		Nested design
		Fully randomised design