3.2 Experimentation Flashcards
How do you achieve valid results?
control all confounding variables in order to prove that results (changes in dependent variable) are because of the independent variable
What are reliable results?
results that are consistent throughout repeats and independent replicates
What does it mean when results are accurate?
when data points or means of data sets are close to the true value
(bullseye)
What does it mean when data points are precise?
data points are close together, but are not necessarily accurate
(grouping)
Pilot studies are used to help plan ___, assess ___ and check ___.
procedures, validity, techniques
The results of a pilot study allow for the main experiments design to be…
evaluated and modified
Pilot studies ensure that the main study has an appropriate range of values for the ___ ___.
independent variable
Summary:
Why would you conduct a pilot study?
allows the main study’s design to be evaluated and modified, and
ensures it has an appropriate range of values for the independent variable.
Pilot studies are used to help plan procedures, assess validity and check proficiency of techniques.
Pilot studies are used to determine the number of repeat measurements required to give a representative value for each datum point
A high degree of variability in the results of a pilot study indicates that…
the number of repeats should be increased for the main study
What is the independent variable?
the variable that is manipulated by investigator
What is the dependent variable?
the one you measure
What is the difference between a continuous and a discrete variable?
What are some examples of each?
A continuous variable is a range of values (eg age, time, concentration of solution ect)
A discrete variable is distinct and countable (eg different species, different solutions)
What is the difference between a simple and a multifactorial experiment?
simple experiments have one independent variable, multifactorial ones have many
Conditions are more easily controlled in the ___ than in the ___.
lab, field
What is a drawback to conducting experiments in the lab?
results may not be applicable in a real life setting
What is an observational study?
a study where the independent variable is not directly controlled by the investigator, for ethical or logistical reasons
(eg experiment on smoking - you can’t ask people to smoke and then investigate them, so you have to find people that have already been smoking - but there is little control of the confounding variable)
What are observational studies good at detecting?
correlation, but not causation
What are confounding variables?
any variables which affect the relationship between the independent and dependent variables
What is used in cases where confounding variables cannot be easily controlled?
randomised block design
What does randomised block design do?
it ensures the effect of any confounding variables is the same across all treatment groups
What do you compare the results of control groups to?
the treatment groups
What is a negative control?
What does it seek to prove?
a group without the treatment/independent variable, used to prove that changes in the treatment groups are due to the treatment/independent variable
What is a positive control?
What does it seek to prove?
a control with a treatment that is known to provide positive results, used to prove that the experiment is capable of detecting positive results, should they occur in other groups
What is a placebo group?
a group that is anonymously not given the treatment.
It is identical to a negative control, however it is subject to the placebo effect
What is the placebo effect?
a measurable change in the dependent variable as a result of the patient’s expectations rather than the treatment
‘In vitro’ refers to performing a procedure ___ an organism.
Give an example.
outside.
cells growing in a culture medium
‘In vivo’ refers to experimentation ___ an organism.
Give an example.
inside.
a new drug
Name a benefit and drawback to performing experiments in vitro and in vivo.
in vitro: confounding variables easy to control, but the results are less relevant
in vivo: allows researchers to investigate the overall effect on a treatment on an organism, but confounding variables are harder to control
When it is impractical to measure every individual in a population, a ___ ___ is taken.
representative sample
When should the sample size be increased?
when there is more natural variation
Samples should have the same ___ as the population as a whole.
mean
Name the three types of sampling.
stratified, systematic, random
Describe stratified sampling.
members are divided into groups and sampled proportionally
(eg 7 blue monkeys and 14 red monkeys in a population - you’d sample 2 blue and 4 red or something along the 1:2 ratio)
Describe systematic sampling.
members lined up and selected at regular intervals
Describe random sampling.
members selected randomly - everyone has an equal chance of being selected
How can you determine the reliability of measuring instruments?
repeated measurements of an individual datum point
Discrete and continuous variables give rise to ___, ___ and ___ data.
quantitative, qualitative, ranked
What is quantitative data?
data that is measured objectively, usually with a number
What is qualitative data?
subjective and descriptive
What is ranked data?
data that is lined up based on numerical (quantitative) values and assigned a rank
What is the mean average?
What is the median average?
What is the mode average?
sum of numbers / number of numbers.
the middle number when all numbers are lined up.
most common value
When does correlation exist?
Correlation does not imply ___.
when there is a relationship between two values.
causation
Causation exists if changes in the ___ ___ affect the ___ ___.
independent variable, dependent variable
Correlation is strong if there is a ___ spread of values from the line of ___ ___ on a graph
narrow, best fit
Positive correlation is represented by (an upwards/a downwards) line of best fit.
Negative correlation is represented by (an upwards/a downwards) line of best fit.
an upwards.
a downwards
Which graph has the strongest positive correlation?
1. A gently sloping upwards line of best fit, with a small spread of points on the graph.
2. A more steeply sloping upwards line of best fit, with a larger spread of points on the graph.
- (closeness is a better indicator than steepness)
What is an error bar?
a line through a point on a graph that shows variation
Data from an experiment without an independent replicate is not ___.
reliable
Other than independent replication, suggest an improvement that can be made to experiments that may increase its validity.
increased sample size
