31 thoughts on “From no-data to data: The awkward transition

    • Yikes – “but never would I have thought to double check something as simple as correctly entering a survey”

      The problem is that is a common very harmful view – the data often are initially incorrect in some if not many ways. Just seems too boring to ever be done wrong – not!

      Also why random audits – everywhere but not often enough to get in the way – are always a good idea for those don’t knows you don’t know.

      • +1 to Keith’s comment. The phrase “something as simple as” raises a red flag of naivete. It ain’t simple! The devil is in the details. Etc.

    • I’m now learning that exploring the explanatory power of data given to you [however seemingly complete and incomplete] is quite difficult, even in the simplest of cases. Particularly after being exposed to such diverse viewpoints and approaches in statistics. Nearly every aspect of any research process under scrutiny thereby making the results derived subject to an avalanche of criticism.

  1. Taken literally, “no-data to data” describes pretty much every study I’ve conducted (aside from secondary analyses). I would think “data to no-data” would be a more awkward transition. (And sadly “no-data to no-data” is where I spend a lot of my policy analytic life.)

    But it does make for a nice koan.

  2. I do think this would be an interesting blog post if you were going to write about the instance of theory development without data. Then when a researcher gets data that go against said theory, it’s quite awkward. One would hope they would come forward with the evidence regardless, but there are clearly cases where that does not happen.

  3. Since we’re documenting potential cognitive decline, I looked for my cell phone the other night with the light from my cell phone for several minutes before I realized what I was doing.

        • Phil,

          LOL I was being facetious. Seriously I did run out of my apt with my jeans on backwards. Thank goodness I’m thin. So not noticeable.

          But you have a point in so far as political science types of inferences made. My buddies still think 90’s viewpoints apply today. In hindsight, we were way more ideological than we admit. That’s why I shun terms like liberals, conservatives, leftists, rightists. We all get stuck in some past, present, and future.

        • You’re reading too much into my comment! I was just saying that, just as the climate change deniers can say “nothing has changed, this happened twenty years ago too”, you can deny your cognitive decline via the same reasoning. Putting your jeans on backwards when you were 26 was a genius way to set a benchmark you can use for decades.

        • LOL, No I understood what you meant. Absent-mindedness isn’t necessarily a function of cognitive decline or memory loss even. I think putting on ahirt inside out is more common. Absent mindedly. I did more of that in my teens.

          I was only suggesting that we all have resort to rationales at some juncture.

  4. I once read a statement from a book or a paper which I didn’t remember correctly “one approach to handling problems involving data is to convert them to no data problem.” so, I eagerly wait for this blog post.

  5. Why did Andrew forget what he was going to write? Is it because he first enters a title in the queue and then six months later actually writes the post?

    That may sound silly, but the alternative is pretty odd too: that he writes a post and then waits six months to post it.

    Either way, it is kind of eerie.

    In a few years, when the truth comes out, are we all going to look back and wonder why we didn’t see that something wasn’t right?

    • Terry:

      I write the posts and put them in the queue. For example, I wrote a post the other day and put it in the queue for January. Every once in awhile I start to write a post and don’t finish it until later. It’s very unusual that I just write the title!

  6. Perhaps you wanted to say that one can make a valid prediction just using priors. The prediction is still valid after updating with data. The statistical principle is the same. What changes when data is added. A blog post on this transition would be nice. Lindley writes about it in his book understanding uncertainty.

Leave a Reply

Your email address will not be published. Required fields are marked *