Biomedical Research: Variable Types

• Session Learning Objectives (LOBs):
  • LOB1: Outline different variable types in biomedical research.
  • LOB2: Determine applicability of summary statistics for descriptive data analysis.

Types of Variables

• Variables can be categorized as either numeric or categorical.
• This categorization is more common in the UK.
• Numeric variables describe measurable quantities.
  • Continuous variables: can take any possible value within a given range, like temperature in Celsius.
  • Discrete variables: take specific values within a given range with data gaps, like the number of teeth (integers).
• Categorical variables are sorted into categories based on qualitative description.
  • Normal categorical: variables divided into unordered categories, like the brand of a car.
  • Ordered categorical: variables divided into categories in an ordered manner, like educational level (elementary school, high school, university).
  • Binary: categorical variable with 2 categories, like affected by a disease (yes/no).
• An alternative categorization method, more popular in the US:
  • Numeric variables can be classified as interval scale or ratio scale.
  • Categorical ones can be classified as nominal scale or ordinal scale.

Numeric Variable Scales

• Interval scale: values are graded in equal increments along the whole range, e.g., temperature in Celsius.
• Ratio scale: Same as interval scale, but has a true zero (true absence of quantity), e.g., temperature in Kelvin.

Categorical Variable Scales

• Nominal scale: variables divided into unordered categories, e.g., country of origin.
• Ordinal scale: variables divided into ordered categories, e.g., income categorized as "low income," "middle income," and "high income."
• Variable type depends on how it is defined and used in a particular study.
  • Blood pressure defined as hypertensive (yes/no) is a binary categorical variable. Blood pressure in mm Hg is a numeric continuous variable.

Summarizing Categorical Variables

• Categorical variables can be summarized by proportions.
• Actual numbers are rarely informative and can be misleading.
• Categorical variables can be presented by counting the number of data values in each category of interest
• These counts can be organized into a frequency table.
• Frequency tables display numbers and the proportion for each category.

Presenting Categorical Data with Graphs

• Bar charts and pie charts can be used to present categorical data.

Bar Charts

• Display the distribution of a categorical variable.
• Counts for each category are shown next to each other for easy comparison.
• Bars have the same widths and equal spaces between them.

Pie Charts

• Represent the total population/sample as a circle.
• The circle is sliced into pieces proportional to each category's fraction of the whole.

Homework Task

• Using a provided dataset, summarize the provided smoking status variable and create a bar chart.
• Using a provided dataset, summarize the provided smoking status variable and create a pie chart.
• Use the Excel function "COUNTIF" to calculate observation frequency, then convert that to the frequency to a proportion.