Research Skills 2 : Experimental Design

Question 1

What is t first step of designing experiments?

Answer 1

Review Existing Work

• Being able to find and evaluate previous work
• Look for what others have done before starting your experiment or research project
• It helps define and tune your ideas
• Avoids reinventing the wheel and wasting time

Question 2

What do you do next?

Answer 2

Must design an experiment that will test your hypothesis.

This experiment will allow you to change some conditions or variables to test your hypothesis

Question 3

What consists for a Good Experiment ?

Answer 3

• Tests ONE variable at a time
• Is FAIR and UNBIASED. As an experimenter you must not allow your opinion to influence the experiment.
• Is VALID- the experiment must test your hypothesis – if it does not the experiment is invalid and results will make no sense!
• Has REPEATS. Repeating the experiments will reduce the effect of experimental errors and give a more accurate conclusion

Question 4

What are Variables?

Answer 4

A variable is
- anything in an experiment that can change or vary

• any factor that can have an effect on the outcome of the experiment

Question 5

How many type of variables are there and what are they?

Answer 5

3 MAIN TYPES:

• Independent
• Dependent
• Controlled

Question 6

What are Independent Variables?

Answer 6

• Something that is intentionally changed by the experimenter
• What is tested/manipulated
• You can only change ONE variable in an experiment!
• To determine the independent variable ask yourself ‘what is being changed?’

Question 7

What are Dependent Variables?

Answer 7

• Something that might be affected by the change in the independent variable
• What is observed and measured – the data collected
• To determine the dependent variable ask yourself – ‘what will I measure and observe?
  – Be specific in when answering this question and include all units
  – i.e I will measure weight in g etc

Question 8

What are controlled Variables?

Answer 8

• A variable that is not changed and kept the same
• NOT the same as a ‘control’
• Any given experiment could have many controlled variables
• To determine the controlled variable/s ask yourself – ‘what should not be allowed to change?
• Examples could be pH or temperature

Question 9

Experimental Controls - NEGATIVE CONTROL

Answer 9

• Negative control
• A sample or experiment to determine what happens without the experimental intervention

A control should ideally be identical to the experiment, except for the one intervention being tested

Question 10

Experimental Controls - POSITIVE CONTROL

Answer 10

• Positive control
• Our hypothesis predicts an effect
• The experiment shows no effect
• Therefore the hypothesis is refuted?
• Or is it just a technical failure of the experimental system
• A positive control is a sample or experiment to show that the experimental system will show an effect, if the intervention really does have an effect

Question 11

What are other types of controls?

Answer 11

• Control to test altrnative hypotheses
• Additional controls
• A control should differ from the experiment in one factor only
• design of controls is crucial

Question 12

Explain Biases an examples

Answer 12

Biases – if you haven’t eliminated all biases from your experiments, then any statistics you calculate are going to be meaningless.

Examples of biases include

• Using inappropriate controls
• Failing to take a truly random sample from a population
• Human bias in observation
• Selecting which data to include after you’ve seen the results of the experiment

Question 13

What are Systematic Errors?

Answer 13

Systematic errors are errors that bias the data in a particular direction.

e.g. a miscalibrated pipette always adds too low a volume of sample

Important to validate your equipment, your reagents and your assays before starting the experiment

Question 14

What are Random Errors?

Answer 14

Random errors are errors that may randomly increase or decrease your readings

Each time you perform an experiment you will get a slightly different answer, because of errors and of natural variation

Question 15

What do you do to reuduce the effects of errors?

Answer 15

• If you increase the number of observations, the random errors tend to cancel each other out.
• Therefore you decrease the effect of random errors, but not the effect of systematic errors.
• We take the mean of multiple observations to lessen uncertainty about how close our experimental answer is to the “real” answer
• Where the real answer is the answer with no random error
• Or the mean of a whole population

Question 16

What is the standard error of the mean?

Answer 16

• combine the overall variability with the number of times the experiment has been repeated to get an indication of uncertainty
• [Mathematically it is the standard deviation divided by the square root of the number of repeats]
• S.e.m. = s.d. / √(n)