STAT1201, 2Oth of February 2022 Module 1 - Pre-lecture notes Why do we need data analysis and statistical thinking? - The data we observe almost always comes with variability Error variability - Natural variability - differences in subjects (e.g., Toads) - Measurement variability - human inaccuracy (reading meter wrong) Making conclusion Rather than permanent conclusions (e.g., "a stimulus of 2 V gives a higher action potential ... ", it is more scientifically correct to leave the answer to an overall statement allowing room to change in scientific discoveries over time. E.g., "on average a stimulus of 2 V gives a higher action potential ... ". Negative difference - conducting a parallel experiment only to find the opposite conclusion of the original experiment, e.g., Experiment 1 found "on average a stimulus of 2 V gives a higher action potential than a stimulus of 1 V" while experiment 2 found "on average a stimulus of 2 V gives a lower action potential than a stimulus of 1 V" Types of Variables Quantitative Variables (Qualitative Data - however, it is avoided to use this term) - Quantitative - Continuous data that is taken over a range, e.g., height of a person or temperature - Discrete - Only a small number of possibilities such as age in whole years or counting data which is whole numbers Categorical Variables - Categorical - Groups of objects with nominal characteristics, e.g., gender - Ordinal - are categories that have an order, e.g., recording the age group someone falls in, then ordering the groups in the natural orders In this section, we have learned that . The need for data analysis comes from the variability present in data. · Separating the differences between groups from background variability is a fundamental task of statistical analysis.
STAT1201, 2Oth of February 2022 . It is important to be able to identify the types of variables recorded in a study. Data from quantitative and categorical variables will be described and analysed in different ways. Comparative Experiments Simulation experiment - does caffeine influence heart rate? - The sample size of an experiment will depend on the effect we're trying to detect Guidelines of simulation experiment - Measure initial heart rate - Give subject coffee - Measure heart rate again after 30 minutes Why is this experiment not very accurate? - Other ingredients within the drink - Environment - anxiety etc. - Placebo effect 1st revised guidelines - Half of the test subjects drank 250ml of diet coke, while 10 drank 250ml of caffeine-free diet coke - A comparative approach Why is this experiment more accurate than the worst? - It is a comparative study thus if the scientist notices a difference, it is much easier to link the found results to the caffeine rather than the environment or other ingredients Why is this experiment still not as accurate as it could be? - The experiment could be conducted on ordinal subjects such as gender. Say the first 10 were female and the second