An incredibly useful tool in evaluating and comparing predictive models is the ROC curve.
Its name is indeed strange. ROC stands for Receiver Operating Characteristic. Its origin is from sonar back in the 1940s. ROCs were used to measure how well a sonar signal (e.g., from an enemy submarine) could be detected from noise (a school of fish).
ROC curves are a nice way to see how any predictive model can distinguish between the true positives and negatives. (more…)
In a previous post we discussed the difficulties of spotting meaningful information when we work with a large panel data set.
Observing the data collapsed into groups, such as quartiles or deciles, is one approach to tackling this challenging task. We showed how this can be easily done in Stata using just 10 lines of code.
As promised, we will now show you how to graph the collapsed data. (more…)
Panel data provides us with observations over several time periods per subject. In this first of two blog posts, I’ll walk you through the process. (Stick with me here. In Part 2, I’ll show you the graph, I promise.)
The challenge is that some of these data sets are massive. For example, if we’ve collected data on 100,000 individuals over 15 time periods, then that means we have 1.5 million cells of information.
So how can we look through this massive amount of data and observe trends over the time periods that we have tracked? (more…)
The concept of a statistical interaction is one of those things that seems very abstract. Obtuse definitions, like this one from Wikipedia, don’t help:
In statistics, an interaction may arise when considering the relationship among three or more variables, and describes a situation in which the simultaneous influence of two variables on a third is not additive. Most commonly, interactions are considered in the context of regression analyses.
First, we know this is true because we read it on the internet! Second, are you more confused now about interactions than you were before you read that definition? (more…)
In the past few months, I’ve gotten the same question from a few clients about using linear mixed models for repeated measures data. They want to take advantage of its ability to give unbiased results in the presence of missing data. In each case the study has two groups complete a pre-test and a post-test measure. Both of these have a lot of missing data.
The research question is whether the groups have different improvements in the dependent variable from pre to post test.
As a typical example, say you have a study with 160 participants.
90 of them completed both the pre and the post test.
Another 48 completed only the pretest and 22 completed only the post-test.
Repeated Measures ANOVA will deal with the missing data through listwise deletion. That means keeping only the 90 people with complete data. This causes problems with both power and bias, but bias is the bigger issue.
Another alternative is to use a Linear Mixed Model, which will use the full data set. This is an advantage, but it’s not as big of an advantage in this design as in other studies.
The mixed model will retain the 70 people who have data for only one time point. It will use the 48 people with pretest-only data along with the 90 people with full data to estimate the pretest mean.
Likewise, it will use the 22 people with posttest-only data along with the 90 people with full data to estimate the post-test mean.
If the data are missing at random, this will give you unbiased estimates of each of these means.
But most of the time in Pre-Post studies, the interest is in the change from pre to post across groups.
The difference in means from pre to post will be calculated based on the estimates at each time point. But the degrees of freedom for the difference will be based only on the number of subjects who have data at both time points.
So with only two time points, if the people with one time point are no different from those with full data (creating no bias), you’re not gaining anything by keeping those 72 people in the analysis.
Compare this to a study I also saw in consulting with 5 time points. Nearly all the participants had 4 out of the 5 observations. The missing data was pretty random–some participants missed time 1, others, time 4, etc. Only 6 people out of 150 had full data. Listwise deletion created a nightmare, leaving only 6 people in the data set.
Each person contributed data to 4 means, so each mean had a pretty reasonable sample size. Since the missingness was random, each mean was unbiased. Each subject fully contributed data and df to many of the mean comparisons.
With more than 2 time points and data that are missing at random, each subject can contribute to some change measurements. Keep that in mind the next time you design a study.
Relative Risk and Odds Ratios are often confused despite being unique concepts. Why?
Well, both measure association between a binary outcome variable and a continuous or binary predictor variable. (more…)