Medical Sciences 3900F/G Lecture 1: Lecture 1 - Experimental Design
3 May 2018
School
Department
Course
Professor
Experimental Design
STEPS OF SCIENTIFIC METHOD
1) Formulate research question
2) Do background research on your topic
3) Identify variables and generate hypothesis
4) Design and test an experiment
5) Analyze, interpret and communicate results
Note: some things have to be in a certain order, but other steps can be moved around
1) RESEARCH QUESTION
- Important
- Always be at the top of the list
2) BACKGROUND RESEARCH
- Necessary for many reasons because you need to know:
o Have to be aware of what work has already been done in the field
▪ Even though it is a good idea to repeat findings, there should be a novel aspect
of your own work
o Help you to design your experiment
▪ Number of subjects
▪ Appropriate tests to use
o How to analyze the data and interpret the results appropriately and draw conclusions
from it
3) VARIABLES
- Identify variables that are significant and choose the ones you want to look at in terms of cause
and effect
- Dependent variable
- Independent variable
- A relationship exists between an exposure (independent variable) and an outcome (dependent
variable)
o Investigation is trying to determine what relationship exists between the variables if we
know that there is one
find more resources at oneclass.com
find more resources at oneclass.com
Dependent variable
- It is what are you measuring/what is affected within the experiment
- Responds to the independent variable
- Question: What is the dependent variable in our experiment (what is being measured)?
Independent variable
- Also called factor
- Unlike the dependent variable, the independent variable is manipulated by the researcher
- Note: there could be multiple independent variables
- Each independent variable may have different levels or conditions (i.e. dose)
o Example: Different dosages of a drug
- Question: what is the independent variable we want to manipulate?
Other variables?
- Are there other variables we should consider that can impact the outcome?
- Control variables
o Kept constant during the experiment or monitored and recorded so that their potential
influence can be assessed
o Kept constant IF possible, if not they are monitored
- Question: can you think of some examples of control variables in our experiment?
4) HYPOTHESIS
- Hypothesis = proposed explanation or phenomenon
- Main feature of a hypothesis
o Conceptual
o Declarative statement
o States a relationship between two or more variables
o Written as a forward or future reference
- Directional hypothesis
o E.g. Female mice treated with drug X will have decreased tumour cell proliferation
compared to controls
- Non-directional hypothesis
o E.g. there will be a difference in motivation between male and female university
students
- Type of hypothesis developed will form the way you present your data
- DIRECTIONAL VS NON DIRECTIONAL:
o Directional predicting the side
o Non directional there is a difference but do not know if it is positive or negative
▪ Explorative research will usually have a non directional hypothesis
Writing a hypothesis
- Based on observations and it should be testable
- Should be able to accept or reject the null hypothesis
o Null hypothesis: the hypothesis that there is no significant difference
- Hypothesis can be:
o Directional (one-tailed) or non directional (two-tailed)
o Advantages and disadvantages to both
find more resources at oneclass.com
find more resources at oneclass.com
- Questions:
o What is the null hypothesis and alternative hypothesis in our experiment?
o Are we working with a directional or non-directional hypothesis? Briefly explain your
response
5) EXPERIMENTAL DESIGN
- What information do you need to test your hypothesis?
o What do you need in order to accept or reject hypothesis?
- What techniques can you use to obtain the information?
- Ensure that the information obtained allows you to draw conclusions from the data
Techniques
- How do we determine if healing time has decreased?
o 1. Simple measurement (i.e. area of wound)
▪ Includes a human study
o 2. In animal model: look histologically (e.g. at number of fibroblasts)
▪ Look at fibroblasts to infer about healing
o 3. In animal model: look for markers of healing (e.g. Amount of FGF mRNA)
- Based on our example, it would be best if we followed through with option 1
Optimization
- Occurs before you begin your experiment
- Some of the variables may need to be optimized
o Age, time, dose, location, severity
- This is a set of experiments on its OWN
o Time consuming BUT essential process of design process and SHOULD NOT be forgotten
- How to optimize?
o Need to isolate each variable so that you can test them individually
▪ Can determine which responds best in the conditions you set up
o Want to know which one responds the best for your experiment?
o Example: optimize dose and time
▪ 3 time points
▪ 2 doses
▪ Once it is determined which time point and doses you are going to include, you
can perform the technique
▪ Get which time and dose combination works best – then proceed to actual
experiment
- Question: Think of two variables or conditions that you could optimize in our example
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
Steps of scientific method: formulate research question, do background research on your topic, identify variables and generate hypothesis, design and test an experiment, analyze, interpret and communicate results. Note: some things have to be in a certain order, but other steps can be moved around: research question. Always be at the top of the list: background research. Identify variables that are significant and choose the ones you want to look at in terms of cause and effect. A relationship exists between an exposure (independent variable) and an outcome (dependent variable) Investigation is trying to determine what relationship exists between the variables if we know that there is one. It is what are you measuring/what is affected within the experiment. Unlike the dependent variable, the independent variable is manipulated by the researcher. Note: there could be multiple independent variables. Each independent variable may have different levels or conditions (i. e. dose: example: different dosages of a drug.
