biostats exam part 4

1. why do we need experiments? 

1. we want to understand the world, but the world is complicated

1. what is an experimental unit? 

1. subject or group of subjects that have been assigned an experimental treatment independently of other subjects or groups

1. the goal of experimental design is to eliminate _________ and reduce _________ when estimating and testing the effects of one variable on another. 

1. eliminate bias

2. reduce sampling error

1. what are three ways to reduce bias? 

1. control groups

2. randomization

3. blinding

1. what is a control group? 

1. unit that does not receive the treatment of interest

2. baseline to compare the effect of the treatment against procedural effects

1. how many groups should treatments include? 

1. at least two → one with the treatment of interest and one control group

1. what is true between treatment and control groups? 

1. all other conditions (minus treatment) must be the same

2. example: 

   1. fertilizer treatment

   2. the plants must have the same sunlight, water, soil, etc.

why do you need two controls in this experiment?

different rivers that contain different species and different resources

having two controls helps assess effects across multiple habitats

1. what is temporal bias? provide an example

1. bias that results from when data is collected

1. example of temporal bias

1. you have two groups, you give the treatment and control to the groups at the same time, then wait and swtich

2. things can happen between the treatments (ex. election, war, etc.) → can skew results

1. what is randomization? what’s the purpose? 

1. subjects are randomly assigned to treatments 

2. reduce bias introduced by confounding variables

1. what is random sampling? 

1. randomly selecting members of a population to participate in a study

1. what is random assignment? 

1. randomly sorting the sample participants into control and experimental groups

1. how does random sampling help with reducing bias? 

1. ensures that the sample is representative of the whole population

1. how does random assignment help with reducing bias? 

1. ensures there is no systematic differences between the members in each group

1. what is the modern approach to assigning treatments to subjects randomly? 

1. computer to randomly assign units

1. what is the classical approach to assigning treatments to subjects randomly? 

1. assign random numbers to the subjects (ex. 1-80)

2. low numbers get A treatment (1-40)

3. high numbers get B treatment (41-80)

1. what is blinding? 

1. concealing information from participatns about which experimental unit receives which treatment

1. what is a single-blind experiment? 

1. only subjects are unaware of treatment

1. what is a double-blind experiment? 

1. both subjects and researchers are unaware of which treatment has been assigned to which subject

1. how do blinded experiments reduce bias? 

1. removes the biases introduced by subject or researcher’s unconsciously acting or feeling different depending on knowing their treatment

2. helps reduce the placebo effect

1. which of these is not necessarily a good method to reduce bias in experimental results? 

1. a design in which the recipients don’t know which treatment they are in

2. randomization of which subjects are included in each treatment

3. the use of control treatments

4. the use of large sample sizes

1. what does sampling error do to power and precision? 

1. decreases power and precision

1. what are 5 ways to reduce the influence of sampling error? 

1. replication

2. balanced design

3. blocking

4. matched pair design

5. extreme treatments

1. what is replication? 

1. each treatment should include multiple independent experimental units (replicates)