Step 1: Draw a mental timeline for a process.
Picture a process arriving into the system at time $T_{arrival}$ and finally finishing at time $T_{completion}$. Several different metrics measure different slices of this timeline.
Step 2: Identify which slice covers the whole journey.
The metric that spans the entire journey, from the moment the process is submitted all the way to the moment it completes, including any waiting, execution, and I/O time in between, is $Turnaround\ Time = T_{completion} - T_{arrival}$. This matches exactly the wording of the question.
Step 3: Contrast with the narrower metrics.
Throughput only counts how many processes finish per unit time, waiting time counts only the time spent sitting in the ready queue, and response time measures only the delay until the first output appears, none of these span the full submission to completion interval.
Step 4: Conclude.
The full submission to completion interval is called
\[ \boxed{\text{Turnaround time}} \]