You can now find a thread with all the #SundayStats threads here; it will be updated each week: bsky.app/profile/exce...

Master thread of #SundayStats threads! Bookmark for later ❤️

Next time: 2.6.3 Construct validity! And the end of 'background'.

See you next week for #SundayStats ❤️

Navarro is SUCH a good writer, as previously observed: this paragraph feels like a teaser #SundayStats

Section 2.6 - Assessing the validity of a study

You remember we talked previously about precision vs accuracy (which may be more familiar terms).

Validity is accuracy. How close is your data to right and true (rather than to repeated measures of itself)? #SundayStats

Or you can go to the very beginning of #SundayStats, which starts here: bsky.app/profile/exce...

(It's important to note that a lot of observational studies-- a lot of what you see in studies based on people's electronic health record (EHR) data, for example-- is simply going to have to take confounds into account. I'm sure we'll get there.) #SundayStats

2.5.1 Experimental research

"The... researcher controls all aspects of the study, especially what participants experience during the study. In particular, the researcher manipulates or varies the predictor variables (IVs), and then allows the outcome variable (DV) to vary naturally." #SundayStats

2.5 - Experimental and non-experimental research

"When we make this distinction, what we’re really talking about is the degree of control that the researcher exercises over the people and events in the study," Navarro says. #SundayStats

2.4 The “role” of variables: predictors and outcomes

This, too, may remind you of your middle-school science education: independent vs dependent variables. Just because it's older doesn't mean it can't be confused.

When I taught middle school science, we spent a lot of time on this. #SundayStats

By a similar token, it's possible to have data that is not particularly reliable, but still valid.

(For example, let's say you as a patient have a really fluctuating RDW on your bloodwork but it trends high-ish. Your RDW is still high, even though the measures are not *precise*.) #SundayStats

Reliability = precision
Validity = accuracy

Reliability/precision is asking, "if I take a measurement multiple times, what is the likelihood it will be the same?" Your bathroom scale is going to be reliable / precise: if you step on and off again, it should tell you the same thing. #SundayStats

(We've already spoken about vocabulary changes across fields: this is precision vs accuracy by any other name. You may remember this from your high school textbooks:) #SundayStats

2.3: Assessing the Reliability of a Measurement. We're now thinking about the *quality* of a measure: how do we know it's any good?

We'll judge using two constructs: reliability and validity.

#SundayStats

Giving everyone a minute to grab a drink, their meds, and curl up. If you're anywhere in the eastern US, it's raining right now 🌧️ #SundayStats

Hey everyone! We're back to #SundayStats! We're working from Navarro's learningstatisticswithr.com and this is only our second week, so it's easy to catch up if you want to join in!

Here is where we started last week: bsky.app/profile/exce...

Honestly as we just discussed in #SundayStats, these are all categorical variables so I'm really done with them after this much math :)

Setting these threads up as I do talks at conferences. We're continuing to Background Section 2 here:

bsky.app/profile/exce...

#SundayStats

OK we are committing: #SundayStats