A good theme does one job that a code cannot. It gathers several codes into one clear idea. In this lesson, I want to show how that works with interview data.
The question is simple: how do you move from small code labels to one higher theme?
I start with the codes themselves. Here, the initial codes describe the challenges and benefits of using Eureka curriculum, other curricula, pacing guides, support for students who are struggling, and the work of carrying out an instructional vision. Each code points to a local problem. It names a specific part of practice.
That is the first step in analysis. A code stays close to the data. It helps sort and mark what people said. A theme sits one level above that. It pulls several codes together when they keep pointing to the same tension or pattern.
In this case, the higher theme is “Clash between ambitious instruction and perception of student capacity.” That theme is more abstract than any one code. It does not name one curriculum or one support method. It names the shared conflict underneath them.
Here is why that matters. The codes are about separate things on the surface. One code is about Eureka curriculum. Another is about pacing guides. Another is about helping students who are still struggling. But all of them can point to the same deeper issue. Teachers are trying to teach in a demanding way, while also judging how much students can take in the moment.
That is the move from description to abstraction. The theme does not erase the codes. It gathers them. It says, in effect, these are all signs of the same larger pattern.
I can make that plain in a small example. Suppose an interview says a teacher likes the structure of a pacing guide, but also feels pressure to slow down because students seem unable to keep up. That single interview segment could be coded in more than one way. It might fit “Challenges and benefits of using pacing guides.” It might also fit “Supporting Currently Struggling Students.” If the same interview also shows tension around keeping the instructional plan ambitious, it starts to support the broader theme of clash.
This is where qualitative research becomes subjective in a useful way. The data do not hand over a single theme with a label on it. The researcher has to construct one. That construction should still be disciplined. It should stay tied to the coded material, not float away from it.
So in task (b), one constructed theme could be this: teachers describe instruction as a constant push between keeping lessons rigorous and responding to what they think students can handle. The initial codes connect to that theme because each one marks a version of that push. Curriculum choice, pacing, student support, and instructional vision all become parts of the same story.
I find it helpful to think of the theme as a sentence that can stand on its own. A code is a tag. A theme is a claim about what the tags mean together. The theme should be broad enough to cover several codes, but not so broad that it loses the tension in the data.
That balance matters in interview work because the point is not to pile up labels. The point is to make a careful claim from the labels. A theme is one way to do that. It gives shape to repeated meaning without pretending the data are simple.
This also shows why the theme has to be higher-level than the codes. If it stayed at the same level, it would only repeat them. A theme must do more work. It must connect separate bits of interview evidence into one idea that explains them.
I would also note one limit. A theme is not proof that every speaker meant the same thing in the same way. It is an analytic construction, and it should be treated that way. The value is in how well it organizes the data that support it.
Once a reader sees this pattern, the method becomes easier to follow. Small codes mark what was said. A theme shows what those codes add up to. That is how interview data become a plain, usable interpretation.
That is also the promise behind The Source List: one digital source worth knowing, one search tip, and one honest limitation.