hypothesis testing, confidence intervals and power of study Flashcards
Statistics
a way to get information from data
Why we study statistics
- The need to read and interpret the published research of other
- Epidemiology is becoming more quantitative
- Consider how reliable a diagnostic test is
- To understand safety and efficacy of a new drug assessed
- To understand how safety and quality of food for human consumption is assessed
What statistics involves?
- Designing experiments/ a survey/ fieldwork
- Collection of data
- Analysis of data
- Summarising information to aid understanding
- Interpretation of the analyses and drawing conclusions from data
Components of research questions
- Domain or study sample
- Exposure/ determinant
- Outcome
Data
information, such as facts or numbers, collected together to be examined
Variables
an element, feature, or factor that is liable to vary or change. For example: each leaf has some attributes (biological surfaces of the leaf, length, colour, surface area etc) that changes among the leaves
describe the classification of variables
- Categorical or qualitative, divided into nominal and ordinal
- Numerical or quantitative, divided into discrete and continuous
Categorical or qualitative variables
describe a characteristic that can’t easily be measured but can be observed subjectively
Ordinal variables
characteristics with clear ordering
Nominal variables
characteristics with no ordering of the categories (binary variable)
Numerical or quantitative variables
describe a measurable quantity on a well-defined scale
Discrete variables
data that can take only integer values
Continuous variables
the data can have almost any numeric value and can be divided into finer and finer levels
Descriptive statistics (or summary statistics)
summarising your data by using tables, diagrams, summary measures (e.g. mean and standard deviation), identifying the underlying frequency distribution (e.g. data obey a normal distribution)
Inferential statistics
based on data from a sample you are trying to reach conclusions that apply to the entire population
Inferential statistics includes
- population
- Sample
- Focus on your sample and measure/ estimate what you are interested in
- Confidence intervals
- Hypothesis testing
Sample
- every element from the population has the same probability to be in the sample
- Infer conclusions from this samples for the entire population
Confidence intervals
an estimated interval within which an unknown parameter may plausibly lie
The function of confidence intervals
Give you an idea of where the true value of what you are measuring lies and p-value summarise the strength of the evidence against the null hypothesis
Calculating confidence intervals
- Dependant on what you are measuring under the general assumptions
- Estimation
hypothesis
an educated guess about something in the world around you based on facts but has not yet been proved
hypothesis testing
formal procedures to accept or reject statistical hypotheses
null hypothesis
usually denoted by H0….
alternative hypothesis
the alternative hypothesis, denoted by H1, is contrary to the null hypothesis
what us the hypotheses texting usual procedure
- formulate the null and alternative hypotheses
- collect the data and look at them, look for outliers
- formulate an analysis plan
- identify a correct test statistics
- calculate p-value
- interpret and make a decision
p-value