1

I have a suspicion that I hope to get help in proving or disproving. I am either right or wrong, either is acceptable.

I assert that gathering data, to be generalized later, can be gathered in too fine of detail and will produce or encourage the generalization to be too specific to be useful.

Does this sound correct or on the right track? Does any theory or research support this or am I totally off base and wrong?

kalisjoshua
  • 111
  • 1
    Yes, you are likely to get accidental correlations. But gathering data is expensive, and one wants to maximize return. – André Nicolas Aug 18 '13 at 02:20
  • A co-worker of mine suggested Simpson's Paradox and I think that is going to be the closest I am going to find to an answer or explanation of what I am looking for. – kalisjoshua Aug 19 '13 at 02:55

1 Answers1

0

Gathering data always requires a hypothesis. Otherwise, you cannot interpret the data in any meaningful way. You have three choices for any given point of data: supports your hypothesis, denies your hypothesis, has no bearing on your hypothesis. You must properly separate these data points or face incorrect conclusions from your hypothesis.

abiessu
  • 8,115
  • A good statement of principle. However, there are many massive data-gathering exercises, particularly in the medical field, that have yielded useful results not foreseen when the project began. – André Nicolas Aug 18 '13 at 03:21
  • Agreed. Each point of data may be usable in its own right. But the end user must have a point of view or the data will be useless. – abiessu Aug 18 '13 at 03:25