Which node type is commonly used for creating new calculated fields in a dataset?

The derived node type is commonly used for creating new calculated fields in a dataset because it allows users to define expressions and formulas to generate new variables based on existing data. This node enables the manipulation of data by applying calculations, aggregations, or transformations that are essential for enhancing the dataset and preparing it for analysis.

Derived nodes facilitate the creation of complex calculations and can accommodate various data types, ensuring that analysts can tailor their datasets to meet specific analytical requirements. Using this node effectively can lead to deeper insights by emphasizing relationships and trends not immediately identifiable in the raw data.

The other options do not serve the purpose of creating calculated fields. The sample node is primarily used to select a subset of records from a dataset, the input node is used to bring data into the model, and the filter node restricts the data based on specified criteria. These functions are crucial in the data preparation process but do not involve deriving new variables from existing data.