Summary

Check that you can describe what a variable means, identify its representation in R, and check a conversion before using the result.

Before you start: first commands and objects.
Setup: Use R for code examples; no data files are needed.
Your check: Convert a height column while preserving the meaning of its station codes. Practise in the unit summary.

Summary

Three questions describe different aspects of your data. Measurement level describes the meaning of values. Discrete or continuous describes the possible values of a quantitative variable. R’s data type describes how values are stored. One does not automatically determine the others.

Measurement level	Meaning	Example
Nominal	Categories without an inherent order.	Land cover: forest, grassland, water.
Ordinal	Ordered categories whose spacing is not defined.	Erosion severity: low, medium, high.
Interval	Equal differences are meaningful; zero is a reference point.	Temperature in degrees Celsius: a 10-degree difference is meaningful, but 20 °C is not twice as hot as 10 °C.
Ratio	Equal differences and ratios are meaningful; zero indicates absence of the measured quantity.	Length in meters: 20 m is twice 10 m.

The four measurement levels. The diagram’s discrete/continuous branches are an oversimplification: measurement level and whether a quantitative variable is discrete or continuous describe separate properties. Use the table above for the distinction.

A count of trees is discrete; measured tree height is treated as continuous, even if recorded to the nearest centimeter. Numeric category codes, such as 1 = forest, remain categories. The choice of a calculation also depends on the question, sampling, and distribution of the data.

Representation in R

Type	Example	Remember
Integer	3L	Whole-number storage; the L requests an integer literal.
Double	3 or 3.5	Stores numeric values, including whole numbers. R code uses a decimal point.
Character	"003" or "forest"	Text; useful for identifiers with leading zeros.
Logical	TRUE, FALSE	Truth values, also used for selecting data.

Use checks such as is.integer(), is.numeric(), or is.character() to inspect a type. Both integer and double vectors count as numeric. Use conversion functions such as as.numeric() only when the intended meaning supports the conversion.

A short recap in R

recorded <- c("12.5", "18.0", NA)
height <- as.numeric(recorded)
is.numeric(height)
is.na(height)
mean(height, na.rm = TRUE)

The type check returns TRUE. The missing-value check returns FALSE FALSE TRUE, and the mean of the two available heights is 15.25. Removing missing values from a calculation changes which observations contribute; it does not explain why they are missing.

NA marks a missing value; use is.na(x), rather than x == NA, to find it. NaN marks an undefined numeric result, such as 0 / 0, and is also detected by is.na(). Inf represents infinity. NULL represents the absence of an object or component and is not a placeholder for one missing observation.

Try it yourself

Predict the result of converting the recorded heights below to numbers. Then convert only the heights; keep the station codes unchanged. How can you distinguish the introduced missing value from a recorded zero?

station <- c("003", "004", "005")
recorded_height <- c("12.5", "unknown", "0")

Hint

Use as.numeric() on the height text, then inspect is.na() and the original strings. Leave the identifier vector as character data.

Compare a possible solution

height_cm <- as.numeric(recorded_height)
height_cm
is.na(height_cm)
station

R warns about the unrecognised text. Heights are 12.5, NA, 0, the missing-value check is FALSE, TRUE, FALSE, and the station labels still start with zeros. The zero is an observed value; the new NA records an unsuccessful conversion, whose meaning needs the source text.

Ready to continue: you can separate a variable’s meaning from its storage type and explain what changed during a conversion.

Check your understanding

Try these questions before opening the suggested answers.

1. Classify land cover, erosion severity, temperature in °C, and tree count by measurement level. Which quantitative example is discrete?
2. Why should the station identifier "003" usually remain character data?
3. What could happen if a text column containing "unknown" is converted to numeric? How would you check the result?

Suggested answers

1. Land cover is nominal; erosion severity is ordinal; Celsius temperature is interval; tree count is ratio and discrete. Recording a category as a number does not change its meaning.
2. A station code is an identifier. Numeric conversion changes "003" to 3, losing its leading zeros; arithmetic on the code is not meaningful.
3. as.numeric() turns unrecognised text such as "unknown" into NA and warns. Keep the original text, inspect is.na() and compare newly missing positions with the source. An original zero is different from missing information.

Revisit data types and conversion for the checks and special values.