topic 1 Flashcards
What are the 8 steps required for successful data analysis?
1 - decide the topic of interest for research
2- formulate a hypothesis
3- design the experiment
4- carry out dummy data
5- use key to direct to appropriate tests
6- carry out the tests
7- no problems, collect real data, in case of problems revert back to step 3
8- perform tests on real data
Why is Hypothesis testing essential for modern science?
Modern science is surrounded on the principle of proving or disproving falsifiable ideas.
The hypothesis can then be rejected or accepted on the grounds of statistical data analysis.
What is the process by which a Hypothesis is tested?
1 - generate a hypothesis
2 - design an experiment
3 - analyse the data using statistical analysis
4 - accept or reject the hypothesis
Describe the difference between a null hypothesis and an alternative hypothesis.
The null hypothesis is always a statement of no difference.
An alternative states that there is a significant difference between two variables.
How does the P-value affect the evaluation of a null hypothesis?
A P-value denotes the likelihood of a null hypothesis being true. The critical level of rejection for a hypothesis is 0.05 i.e. the null hypothesis has a 5% chance of being true.
Therefore, the smaller the P-value the more confident researchers can be about the conclusions drawn from statistical analyses.
What is the difference between and observational study and an experimental study?
Observational study is the conduction of fieldwork to collect data for your intended area of research/hypothesis.
Experimental study comprises of controlling conditions within a laboratory environment in order to collect data surrounding your area of study/hypothesis.
What are the pros and cons for observational and experimental study?
Observational:
- collected observations are more applicable to the real world
- no control over the variable under investigation
- no controlling over factors that may confound the experiment
Experimental:
- full control over all aspects of the investigation
- less applicable to the real world (big disadvantage in biology)
What did Celis-Morales et al research and what were the main findings?
The study investigated the association between active commenting and CVD.
>26,000 participants were studied.
Cycle commenting was associated with a lower risk of CVD, cancer and all cause morality.
Why would you collect a finite sample of a population? What key method would enable you to generalise the findings?
You would collect a finite sample in order to describe a population.
Only when you conduct random sampling can you generalise the findings.
What is spatial sampling and how would you best conduct it?
Spatial sampling is the process of collecting observations in a 2-dimensional framework to determine the contents of a wider geographical area.
The most optimal method for this sampling technique is to produce random number co-ordinates to choose the areas to be analysed using sample squares.
How could you randomly allocate mice to a treatment group?
Tag/code each animal with a random number which has previously generated and allocated a treatment.
How could you fairly measure photosynthesis rates in control and experimental plants? You can only measure four control and four treats plants per 8 hour day.
Plants have a circadian rhythm that affects photosynthetic rates. Therefore, timescales have to be carefully taken into account in order to prevent confounding results.
Alternate measuring between control and treatment plant every hour for 8 hours.
What is investigator bias and how can it be avoided?
Investigator bias, or experimental bias, is where the researcher either subconsciously or consciously influences the outcome of the results. To remove subjective results more than one investigator studying observations can be beneficial.
Why is spatial design important ?
Unknown factors could confound the outcome of the experiment.
What is a Latin square and how do you formulate one?
A Latin square is a n x n array where one aspect only appears once in each row and column; it is used when a researcher wishes to control for the variation in an experiment that can be related to rows and columns.
Design systematically.
