Descriptive Statistics Flashcards
What is the role of descriptive statistics?
- Describe the sample of research participants
- want to convey the average/central tendency and the spread of data
- can summarise data numerically and graphically
What is the second purpose of statistics?
To use data from a sample to tell us about a population
What is the role of inferential statistics?
- estimating the values is a population from a sample
- to determine whether a relationship between variables is real or due to chance
- to see if differences between groups in an experiment/trial are caused by the treatments applied or due to chance
What are the different types of data/variables?
- categorical/binary (e.g. gender, ethnicity)
- continuous (e.g. age, height, weight)
What are the different levels of measurement?
- nominal (binary and categorical)
- ordinal (order or rank categorised)
- interval ( continuos)
- ratio (continuos)
What is nominal data?
- the data is allocated into named categories only, with no implication of size, value or order
- numbers are used to code the categories of a categorical variable
What is ordinal data?
- categorical data that is ordered (intervals may vary) (e.g. Disease stage, degree of pain)
- can’t have means for these results
- categories are represented by a number or words
What is interval level?
- continuous data measured along a scale with equal interval/meaning
- data are represented by a number
- the scale doesn’t have a true zero point (e.g. blood pressure, time of day)
What is ratio level?
- interval data but with the additional property that ratios are meaningful
- zero has a meaning (e.g. age, height)
- highest level of measurement
What is included in central tendency?
- mode
- median
- mean
What is used to show the dispersion of data?
- range
- inter-quartile range
- standard deviation
Why is dispersion important?
Describes how spread out the data is
What is inter-quartile range?
- the range between the lower quartile (1/4 of the way) and the upper quartile (3/4 of the way)
- the range of measurements for the middle 50% of the population whose values lie between the 25% and 75%
What is standard deviation?
The average distance, or deviation, of each value from the mean
How do you describe nominal data numerically?
- data can only be summarised as frequencies or percentages
- central tendency is represented by the mode (bi-modal or multi-modal)
- the spread of data is not applicable
What is a frequency table?
- summarises all of the data visually
- in the text focus on important findings
When are cross tabulations used?
Frequency tables can also be used to illustrate the relationships between categorical variables
How do you describe ordinal data numerically?
- central tendency can be given as the mode and/or median
- short scales are often treated as simply nominal data and displayed using frequency tables
- spread of data can be described via the range and/or inter-quartile range
How is nominal and ordinal data graphically described?
Bar chart
How do you numerically describe interval/ratio levels?
- central tendency can be described using mode, median and mean
- spread of the data can be described via the range, inter-quartile range and/or standard deviation
How do you graphically describe interval/ratio level data?
Frequency distribution (histograms)
Why are histograms used to describe interval/ratio level data?
Used to illustrate the shape of the distribution of values in the data
What is a parametric histogram?
Normal disrribution
What is non-parametric histogram?
Skewed distribution, can be positively or negatively skewed
What is normal/parametric distribution?
- distinctive bell shaped curve with values symmetrical about the mean
- normal distribution: 68% within 1 SD of the mean, 95% within 2 SD, 99.7% within 3 SD
- occurs naturally; many human attributes plot on a normal distribution
What is a skewed/non-parametric distribution?
- the most frequent scores are clustered at one end of the scale
- numbers fall mainly below the mean = distribution is positively skewed
- numbers fall mainly above the mean = distribution is negatively skewed
- mean, median and mode aren’t the same
What is the confidence intervals for the mean?
- an estimated range of values around the sample mean which are likely to include the Unknown population mean
- if many independent samples were taken from the sample population, the a certain percentage of the intervals will include the population mean
- normally 95%
What is the first purpose of statistics?
- to summarise data that has been collected about/from a group of participants