final

Question 1

What is a number that summarizes sample data values and is often used to estimate a characteristic of a population?

Answer 1

Statistic

Question 2

What is a number that describes some characteristic of an entire population?

Answer 2

Parameter

Question 3

Values from repeated measurements of the same sample are evenly distributed around the true population value, but are highly variable. What does this indicate?

Answer 3

Low precision, unbiased

Question 4

Select ALL statistics below that are considered ‘robust statistics’.

Answer 4

• Median
• IQR

Question 5

What does it mean when two events are mutually exclusive?

Answer 5

Both events cannot occur simultaneously

Question 6

According to the law of total probability, what must the sum of all probabilities equal?

Answer 6

1

Question 7

What is theoretical probability?

Answer 7

The frequency that some event will happen that is based on all known outcomes

Question 8

What is empirical probability?

Answer 8

The frequency that some event will happen in the long run, based on observations

Question 9

What is the difference between simple random sampling and stratified random sampling?

Answer 9

Simple: Randomly selecting people without a strategy; Stratified: Population divided into subgroups and random samples taken from each

Question 10

Why do we test the null hypothesis?

Answer 10

Because the exact expected values of a test statistic are known under the assumption there is no difference

Question 11

What are the two reasons why the sample statistic may deviate from the null hypothesis?

Answer 11

• Difference due only to random variation
• If there really is a difference

Question 12

A p-value is defined as the probability of obtaining what?

Answer 12

A test statistic as extreme (or more extreme) if the null hypothesis were true

Question 13

Under the null hypothesis statistical testing framework, what is the probability of declaring there is an effect when there is no effect known as?

Answer 13

Type one Error rate

Question 14

What is the null hypothesis value for the population parameter?

Answer 14

The expected parameter value under an assumption that there is no difference or effect

Question 15

What is the effect of increasing sample size on a p-value, all other things being equal?

Answer 15

Decreases the p-value

Question 16

Which of the following would decrease the p-value of a t-test, assuming all other factors were held the same?

Answer 16

Increase sample size

Question 17

What would increase the width of a confidence interval?

Answer 17

Increasing the confidence level

Question 18

If you know the population standard deviation and population mean, what test would you use?

Answer 18

Z

Question 19

If your data are positively skewed, what transformation would you use to attempt to ‘normalize’ it?

Answer 19

Logarithm

Question 20

If your data are negatively skewed, what transformation would you use to attempt to ‘normalize’ it?

Answer 20

Square

Question 21

Statistical power is defined as what?

Answer 21

The probability that a study would detect an effect, if the effect is of a specified size

Question 22

Which of the following would increase statistical power, assuming all other factors were held the same?

Answer 22

Increase sample size

Question 23

To determine if resistance training can prevent or reverse loss of bone density in post-menopausal women, what test should be used?

Answer 23

2-sample t-test

Question 24

What test should be used if the control group is negatively skewed with six outliers?

Answer 24

Mann-Whitney U-Test

Question 25

According to the owner's manual for an analytical balance, what type of test is used for repeated measurements of a calibration weight?

Answer 25

One sample t-test

Question 26

What test should be used to determine if adding nitrogen fertilizer increases the abundance?

Answer 26

Paired t-test

Question 27

If you increase the number of tests for a single null hypothesis, what is the effect on the experiment wise type one error?

Answer 27

Experiment wise type-1 error will decrease

Question 28

How can you control (or account for) the experiment wise type-1 error rate for multiple tests?

Answer 28

* Bonferroni method * ANOVA

Question 29

What does the Bonferroni method do?

Answer 29

Reduces individual comparison-wise type-one error to control for an overall experiment-wise type-1 error

Question 30

Select ALL statements below that are assumptions of ANOVA.

Answer 30

* Each group is a random sample from population of interest * The error variances within each group are equal * The variable is approximately normally distributed within each group * Sample sizes across groups must be equal

Question 31

ANOVA stands for what?

Answer 31

Analysis of Variance

Question 32

What does 'among-group variance' include?

Answer 32

* Random variation * The treatment effect

Question 33

A random block design is used when what occurs?

Answer 33

There is increased variability due to the group the samples are from

Question 34

A factorial design is used when what is present?

Answer 34

There is an interaction between two variables

Question 35

What would be appropriate if you collect data from three lakes and N=10, data not distributed (Shapiro Wilkes)?

Answer 35

Levene’s Test

Question 36

If previous tests determined variances were equal, what test should be used?

Answer 36

ANOVA

Question 37

If you collect 10 water samples from two lakes and compare 5 variables, what adjustment should be used?

Answer 37

Bonferroni Adjustment

Question 38

What type of ANOVA should be used when testing the effect of three diets on peristaltic blood pressure with variability across individuals?

Answer 38

Random Block ANOVA

Question 39

If you select 60 people to test a diet pill, each group gets 20 people, and the data is not normally distributed with unequal variances, what test should be used?

Answer 39

Welch’s ANOVA

Question 40

To determine if a food supplement affects activity in mice, but one mouse escaped and the data is not normally distributed with unequal variance, what test should be used?

Answer 40

Kruskal Wallis

Question 41

If you are interested in the effect of acetaminophen and there might be an interaction between two variables, with data normally distributed and variances equal, what test should be used?

Answer 41

Factorial ANOVA

Question 42

Two variables have an association when what occurs?

Answer 42

Values of one variable coincide with a pattern of change in the other variable

Question 43

What is an observation in statistical analysis?

Answer 43

Finding a statistical pattern of co-variation among two variables

Question 44

What is an inference in statistical analysis?

Answer 44

Concluding one variable causes an effect on the other

Question 45

Identify the lurking variable.

Answer 45

A

Question 46

Assumptions for linear regression include what?

Answer 46

* Linearity * Statistical independence * Homoscedasticity * Normality

Question 47

What is the null hypothesis of correlation analysis?

Answer 47

P=0

Question 48

How do you assess the linearity assumption of linear regression?

Answer 48

Scatterplot

Question 49

Applying Bayesian thinking to decision making involves incorporating prior information represented as what?

Answer 49

P(D|H1,X)

Question 50

What does the law of inverse probability (Bayes' Theorem) calculate?

Answer 50

P(H|D)

Question 51

If tree farms represent a major long term investment, what method do foresters often use to predict the productivity of a piece of land?

Answer 51

Pearson’s Correlation

Question 52

What does a forester want to do after identifying the soil/site variable most strongly associated with forest plantation productivity?

Answer 52

Use this association in Regression

Question 53

An environmental health scientist wants to determine the association between exposure to ozone air pollution and lung capacity. What method should be used?

Answer 53

Spearman’s Correlation

Question 54

What type of statistical method is used when bacteria from lake sediments are cultured in the presence of different concentrations of methyl mercury?

Answer 54

Regression

Question 55

Describe the difference between a 95% confidence interval and a 95% credible interval.

Answer 55

* Confidence: Range of values where the true population mean lies with 95% certainty * Credible: Represents probability about the true value

Question 56

What is the difference between Bayesian and Frequentist approaches?

Answer 56

* Bayesian: Uses prior knowledge, parameters are random * Frequentist: Parameters are fixed, probability is a frequency of occurrence