2400 times a year, divided by 52 weeks a year:
- 2400 / 52 = 46.15 robberies a week
46.15 times a week, divided by 5 days a work week:
- 46.15 / 5 = 9.23 robberies a day
9.23 times a day, over an 8 hour work day:
- 9.23 / 8 = 1.15 robberies per hour, or, 1 robbery every 52 minutes.
So, we have:
- 2400 robberies a year
- 46.15 vs. 44 robberies per week
- 9.23 vs. 9 robberies per day (using 5-day work week)
- 52 vs. 48 minutes between robberies occurring (using 8-hour work days)
And then, working it backwards:
1 robbery every 48 minutes => 10 robberies a day => 50 robberies a week => 2600 robberies a year.
- 1 robbery every 48 minutes
- 10 vs. 9 robberies a day
- 50 vs. 44 robberies a week
- 2600 vs. 2400 robberies a year
IMO, the figures work out fine and aren't too far off (~10% off at most). It does seem though that the statistic is only using the typical operating hours of a business in the US, that being - Monday thru Friday, 8a to 5p - which suggests that all reported robberies within this statistic took place during a bank's working hours (or, at least, to me it does).
That being said, as @Paulie_D has mentioned, the statistic may not be that accurate.