Me Myself and I

Question 1

Why is biology considered a quantitative

Answer 1

Biology involves collecting and analyzing quantitative data to test hypotheses and make prediction. Biological research relies on quantitative measurements of different parameters.

Question 2

Why is understanding data so important?

Answer 2

There is a lot of variation within biology, by understanding data, we can establish causes and make predictions. We can establish cause-effect relationships by manipulating variables and understanding the resultant data.

Question 3

Why aren’t excel files used to collect data?

Answer 3

Excel files cant be easily opened and used in other software, CVS files are easier to work with.

Question 4

How can graphs and statistics help you understand data?

Answer 4

Graphs allow you to visualize data, statistics allow you to summarize data.

Question 5

What is the difference between samples and populations?

Answer 5

A sample is usually only a subset of the total population. A population is all the members of a defined group.

Question 6

How can we make sure our sample is representative of the population?

Answer 6

Sampling bias is the result of poor experimental design and biases the results. In order to avoid this, the sample should be as representative of the entire population as possible.

Question 7

What is sampling error?

Answer 7

The random variation introduced into a data set as a result of only sampling a subset of the population. The results from your sample may not be applicable to the entire population.

Question 8

What type of data collection may create bias?

Answer 8

Self-reported data can create bias due to inaccurate reports.

Question 9

Why is statistical testing important?

Answer 9

Its not enough to look at the data and make assumptions, it is important to statistically test the reliability and significance of our findings.

Question 10

What is a statistical hypothesis?

Answer 10

A statistical hypothesis is a statement or assumption about the characteristics of a population or the relationship between variables that is subject to statistical testing

Question 11

What is the null hypothesis?

Answer 11

The default expectation that categorical outcomes are equally likely, and there is no relation between two measured phenomena or that there is no association between groups.

Question 12

What is an alternative hypothesis?

Answer 12

The expectation that categorical outcomes are not equally likely, that there is a relation between two measured phenomena or an association between groups

Question 13

When should you used a chi-squared test?

Answer 13

When comparing two categorical variables e.g. diabetes prevalence amongst males and females

Question 14

When should you use a t-test?

Answer 14

When comparing one categorical variable with one continuous variable- compares the means of both groups e.g. blood pressure amongst males and females

Question 15

When should you use a general linear model?

Answer 15

When trying to establish a relationship between two continuous variables e.g. height and finger length.

Question 16

What is statistical significance?

Answer 16

Statistical significance is the claim that the produced results would be very unlikely under the null hypotheses, so there is a relationship or association between variables.

Question 17

What is the p-value?

Answer 17

P-value is a measure of statistical significance, it is the probability of the shown results occurring under the null hypothesis

Question 18

What is the p-value threshold?

Answer 18

The p-value threshold is 0.05. If a p-value is lower than this, results are generally considered statistically significant.

Question 19

Why does a very low p-value indicate more significance?

Answer 19

The lower the p-value, the less likely the results would occur under the null hypothesis. A p-value of 0.05 suggests there is a 5% chance the data would show randomly under the null hypothesis.

Question 20

Why may a p-value very close to the threshold not be reliable?

Answer 20

A p-value very close the the threshold may be due to a sampling error. If this is the case, bigger more representative samples may help reduce sampling bias

Question 21

What is a type one error?

Answer 21

False positive, this is when results provide evidence against the null when it is true.

Question 22

What is a type two error?

Answer 22

False negative, this is when results provide evidence for the null when it is not true.

Question 23

What is effect size?

Answer 23

Effect size is the magnitude of the effect seen in the results. The p-value may be very low but the effect may be negligible e.g. drug trial shows very low change in variable is not useful. Effect size with continuous variables measures the strength of the association and the gradient of the line.

Question 24

Why is biological context important for interpreting data?

Answer 24

Biological context helps interpret data correctly, can give a explanation to why a result has occurred e.g. spikes in names such as harper when Beckhams daughter was born, or drops in names such as alexa.

Question 25

What data does a boxplot show?

Answer 25

A boxplot is a efficient way to present continuous data against categorical data. It shows the median, max/min, interquartile range and outliers.

Question 26

What does a statistically significant result indicate?

Answer 26

A statistically significant result indicates likelihood, we can use a statistically significant result to make predictions but it does not have 100% confidence.

Question 27

What is the difference between correlation and causation?

Answer 27

Correlation is when two variables are related to each other, causation is when one variable causes a change in another.

Question 28

What can correlation indicate?

Answer 28

1. Variable x causes variable y
2. Variable y causes variable x
3. A third independent variable is causing both.

Question 29

What does a t-test tell you?

Answer 29

Whether the means of two groups are statistically different.

Question 30

What does a regression model produce?

Answer 30

A regression model produces a p-value testing how significant the results are, a y-intercept and gradient of the line of best fit, and an r2 value to measure the strength of correlation.

Question 31

What is a line of best fit?

Answer 31

The straight line that best represents the relationship between the dependent and independent variable. The sum of residuals will be close to zero on a strong line of best fit.

Question 32

What are residuals?

Answer 32

Residuals are the differences between the observe and those predicted by the regression line.

Question 33

How can we use the line of best fit?

Answer 33

The line of best fit can be used to predict the value of one variable at the value of the other. The straight line equation can be used to do this.

Question 34

What is the workflow for data analysis?

Answer 34

1. Plot data
2. Initial visual analysis
3. Statistical tests
4. Interpret test output
5. Interpret results in biological context

Question 35

Why is a 95% confidence interval used?

Answer 35

When using a sample, it is likely that the results will not be exactly equivalent to the population. The 95% confidence interval is the range in which the population will lie 95% of the time.

Question 36

Why is using large samples important?

Answer 36

The larger the sample the more likely it is to be representative of the population. Larger samples = smaller 95% confidence intervals.

Question 37

What does overlapping confidence intervals mean?

Answer 37

When the intervals overlap, the results are not likely significant.

Question 38

How does football relate to height?

Answer 38

Football is a male dominated sport so football players are usually taller than non football players, it is more likely a measure of gender against height. Example of third independent variable.

Question 39

When is a fishers exact test used?

Answer 39

When you have a small sample size.

Question 40

What is a multi-variate model used for?

Answer 40

A multi-variate model can be used to test multiple independent variables against a dependent variable. They can tell you how much each variable contributes to your results. This allows you to control for confounding variables.

Question 41

What is the difference between cross sectional data and longitudinal data?

Answer 41

Cross sectional data is taken all at one time point by collecting data from different individuals and generations (cohorts). Longitudinal data follows an individual over a long period of time.

Question 41

What is multiple testing?

Answer 41

Testing many variables at once and hoping one is significant. Research question should be decided before experiment takes place. This often results in real co-variables but also chance associations.

Question 42

What is cherry picking?

Answer 42

Cherry picking is only presenting positive results and ignoring other findings. This is a misrepresentation of the data. All positive and negative data should be published.