In more detail
Phones often record at a variable rate that shifts with light and load, whereas edited or converted files tend to have a constant rate. A mismatch between the declared and measured rate, or dropped and duplicated frames, can indicate editing. Common rates are 24 fps for cinema, 25 and 50 fps in former PAL regions, 30 and 60 fps on phones and the web, and the NTSC broadcast rate of 30000/1001 ≈ 29.97 fps. Each frame lasts 1/rate seconds: 33.3 ms at 30 fps, 40 ms at 25 fps.
Why it matters
Rates and timestamps underpin any timing claim made from a video: how fast a car moved, how long an event lasted, whether a gap was cut out. Using 30 instead of 29.97 fps, or ignoring variable frame timing, skews those measurements. Mixing clips with different rates in an edit also produces judder or duplicated frames that an examiner can spot. See the Video Forensics.
Example
The difference between 30 fps and 29.97 fps adds up:
1 hour at 30 fps = 108,000 frames
1 hour at 29.97 fps = 3600 × 30000/1001 ≈ 107,892.1 frames
difference ≈ 107.9 frames ≈ 3.6 seconds per hour
That 3.6-second drift is why broadcast timecode uses "drop-frame" numbering, and why a speed or duration measured from frame counts must use the exact rate.
Common confusion
Frame rate is not the same as shutter speed or refresh rate: a 60 Hz monitor can play 24 fps video, and a 30 fps clip can be shot with a 1/1000 s shutter.