Updating report @SalvantoCBS and I did on party switching from Jan-July 2017: cbsnews.com/news/2017s-big…. Analyzing @CBSNews/@YouGov panel data between Feb/Mar '17 and May/Jun '18, substantive results looking quite similar: only 8% switched parties over this roughly 15-month period.
People who switched parties were equally likely to go leftward and rightward. The frequencies of each type of switch (GOP-to-Independent, etc.) are similar to what we found last year (second graphic). But remember, the base is small. (Note: Independents include leaners here.) 
Is switching parties related to changing one's support for Trump? The Nation Tracker includes this item, which is similar to four-part presidential approval. I code first two categories as support and latter two as oppose. Trump had 42% support in the May Tracker by this measure.
Overall, 8% of people changed their support for Trump in these ~15 months (i.e., from support to oppose, or vice versa). That isn't a lot and similar to the probability of switching parties. Changes in support *level* (e.g., 'against now' to 'strongly against') more common at 26%
There's a relationship between switching parties and changing one's support for Trump. Among people who changed their Trump support, 24% switched parties. It's only 7% among people consistent in their support, a much larger group. But neither type of switch happened very much.
Trump support is 90% among Republicans at t2. Counterfactual: if people who changed support had *not* switched parties, GOP support would still be 88%. I'm constrained by a small base of switchers, but I doubt party switches affect Trump's support by more than a couple of points.
