Quantitative data analysis - basic principles Flashcards
What type of statistics is this?
- Summarises data
- Can show averages and spread (variance)
- Can show associations and correlations
Descriptive statistics
What type of statistics is this?
- Makes inferences from the sample to the
population. - Tells us how likely the results are due to the
intervention rather than sampling error - Can provide strength of evidence
- Is used to test hypotheses
Inferential statistics
Match the term with the definition
Variable
A characteristics that can measured and can assume different values
Match the term with the definition
Categorical variable
Refers to a characteristic that is not quantifiable
Match the term with the definition
Numeric
A quantifiable characteristic – values are numbers
Match the term with the definition
Discrete
Countable; measured on a continuum
Match the term with the definition
Continuous
Measured numerically; infinite number
Match the term with the definition
Nominal
Describes a name, label or category
Match the term with the definition
Ordinal
Clear ordering of categories
Match the term with the definition
Interval
A level of measurement that has mathematical properties but no true zero
Match the term with the definition
Ratio
A level of measurement that has mathematical properties and a true zero
Which type of graph would you use to show the following:
- To show the relative proportions of the categorical
data (e.g. how many portions of different meals were
bought in the canteen last week)
Bar chart
Which type of graph would you use to show the following:
- The show what percentage of the whole is made up
by a small number of categories
Pie chart
Which type of graph would you use to show the following:
- A trend over time of a measure.
Line chart
Which type of graph would you use to show the following:
- The strength of relationship between two variable
measured numerically.
Scatterplot
What are some characteristics of a bar chart?
- The Y axis provides a numeric value (which can be
frequency or count) - Used to compare discrete (separate) or categorical
variables. - Would suit showing the preferences of a group for
different types of fruit - There is space between the bars
- Each bar has the same width
What are some characteristics of a histogram?
- The Y axis represents the frequency count.
- Use to show the distribution of frequencies in a data
set. - Would suit showing the distribution of heights of a
group of people. - Each bar represents how many times a data point (or
range of data points) repeats. - There is no space between the bars
- Bars can vary in width from each other
Correlation between two continuous variables can be shown using a scatterplot, or expressed using a number. In Pearson’s correlation
this number is called r.
What values can r take?
-1 to +1
Correlations can be positive, neutral or negative. We are conducting an experiment to determine whether there is a linear relationship
between optimism and life satisfaction. When we analyse the data set we and that the r value is 0.7. What does this mean?
17
There is a strong positive correlation between optimism and life satisfaction.
When we design a research study we need to determine the effect of an intervention, we need to make sure that it is the intervention
that led to the change and not something else. One of the best designs of research to do this is…
- Randomised controlled trial
- Cross-sectional survey
- Qualitative study
- Audit
Randomised controlled trial
We are conducting a study to determine whether reflexology enhances well-being in student nurses. The intervention is reflexology. The outcome measure is well-being.
We will measure well-being using a validated questionnaire before reflexology starts, then we will deliver six sessions of reflexology, and then we will measure well-being again. I want to be sure that reflexology is causing the difference in well-being, not something else I have not measured. What is that unmeasured variable that influences the outcome measure called?
21
Confounding factor
What is the interquartile range?
The range of values of the central 50% of the data.
In a box plot the whiskers extend to…
The lowest and highest values within 1.5 times the IQR from the quartiles, excluding outliers.
