Data Summarisation and Analysis Overview

Why Summarise Data

• Summarising data is essential for data quality monitoring.
• Important for data checking and cleaning processes.
• Helps establish baseline data in studies.
• Necessary prior to conducting complex analyses.

Quantitative Data

• Defined as data that can be measured numerically, encompassing both continuous and discrete types.

Continuous Data

• Exists on a continuum; can take any value within two limits, limited only by measurement accuracy.
• Examples include weight, which is measured on scales allowing for decimal representation.

Discrete Data

• Composed of specific values that do not lie on a continuum, usually represented by counts.
• Example includes the number of previous pregnancies, only whole numbers are relevant.

Ordinal Data

• Quantitative data are inherently ordinal, meaning values can be arranged numerically from smallest to largest.
• Often arises in questionnaire scales where total scores are derived from positive responses.
• Categorical data may also be inherently ordered, like stages of a disease.
• Continuous data may appear discrete due to measurement/reporting methods (e.g., gestational age may be noted in whole weeks but is continuous).

Categorical Data

• Data where individuals are classified into separate categories.

Examples

• Gender: male or female (2 classes).
• Disease status: alive or dead (2 classes).
• Stage of cancer: I to IV (4 classes).
• Marital status: various classifications (5 classes).

Ordering of Categorical Data

• Categories may be numerically coded for convenience, and may imply ordering (e.g., cancer stages).
• Calculating means for ordinal data (like cancer stages) is often unhelpful.

Dichotomous Data

• Type of categorical data with only two classes; also referred to as binary data.

Categorising Continuous Data

• Continuous data can be reclassified into groups for easier reporting.
• Example: reporting birth weight in bands to illustrate how many babies fall within each weight range.