Which statement summarizes multiple observations into a new variable while maintaining its value across iterations?

The SUM statement is designed to summarize multiple observations into a new variable while retaining its value across iterations. When used in a DATA step, the SUM statement enables accumulation of values into the specified variable as the data is processed. This means that each time the DATA step iterates over a new observation, the specified variable retains its previous sum and adds the new observation's value to it.

For example, if you were calculating a running total within a dataset, the SUM statement would keep the running total variable's value from the previous observations, ensuring that it reflects the cumulative value up to the current observation. This functionality is particularly useful in scenarios where you need to perform operations like calculating totals, means, or any other aggregated metrics over a sequence of records.

The other statements have distinct functions: the SAVE statement does not summarize observations; instead, it is used to save the state of a dataset. The CALC statement is not a standard SAS feature; thus, it does not apply in this context. The MERGE statement is used to combine datasets based on a common key, rather than summarizing values within a single dataset. Therefore, the SUM statement stands out as the appropriate choice for the task of summarizing observations into a new variable while preserving its value through