Reproduction of results isn't "add the numbers that they produced to see if they sum to the value they said it did". That isn't replication of science.
1. The data may not support their results. Without it, you can't verify that.
2. The data may be, let's say, 'adjusted' to give better results without admitting it. You may be able to show that by statistical checks, but you can't do that without the data.
Yes, you could completely re-do the experiment, but a) it may be historical data which can't be measured again (e.g. deceleration of a space probe from 1980 to 2000) and b) that may massively increase your costs.