The worst moment in a group meeting is often not when your advisor points out a problem, but when they ask, "Where did this number come from?" You look down and scroll through your phone for half a minute, then say, "I'll go back and confirm."
The problem usually isn't that you prepared too little, but how you prepared. This article breaks down a very simple habit: go through your materials yourself before the presentation, in a fixed order.
01 Four checkpoints, and the order cannot be changed
▲ Figure 1 | From conclusions to uncertainties, the four checkpoints build on one another
The first checkpoint is conclusions. Before looking at the materials, write down the three conclusions you really want to make. This step sounds redundant, but it is actually the most important—if you don't define the conclusions first, you won't know which evidence to look for, and it easily turns into "put up every chart I have."
The second checkpoint is evidence. Attach the strongest one or two pieces of support to each conclusion. Note: the "strongest," not "all." Three pieces of medium-strength evidence are not as good as one solid piece.
The third checkpoint is sources. Label the source behind every number clearly, and it's even better if you can specify the record ID or literature location. This checkpoint is specifically for "being questioned."
The fourth checkpoint is uncertainties. List the parts you can't explain clearly on their own. This checkpoint is the easiest to skip, because instinctively you don't want to show uncertain areas. But it is precisely this part that best demonstrates how well you know the materials—knowing where you don't know is itself information.
The order of the four checkpoints cannot be changed. Only after you have conclusions do you know what evidence to look for; only after you have evidence do you know where to look for sources.
02 When questioned, where do the two kinds of materials differ?
▲ Figure 2 | With the same three questions, two preparation approaches produce completely different responses
The difference in how materials are prepared becomes concentrated and exposed the moment you are questioned.
Loose materials refers to the state of relying on last-minute searching: when asked for a source, you scroll through your phone looking for a screenshot; when asked about the definition, you can't remember whether it was this version or the previous one; when asked about a comparison, you can only answer, "I'll confirm that." This is not a memory problem; it's that the materials themselves have never been organized.
Materials with evidence attached are in a completely different state: sources can be traced directly to a record ID; definition changes are written in the footer and visible at a glance; baselines and experimental groups sit side by side in the same table, and the comparison relationship is ready-made.
The difference is not how much you remember, but whether the materials themselves can be pointed to item by item. The former depends on your memory; the latter depends on the structure of the materials. What MuchSci does here is very simple: gather scattered records under unified fields, so that "going to find a record" becomes a searchable action rather than an act of recall.
03 "Whether it exists" and "whether it can be verified" are two different things
▲ Figure 3 | Both are called self-checks, but what they check is not actually the same thing
Many people do self-checks that examine form: whether heading levels are consistent, whether charts have units labeled, whether sources are listed.
Of course these need to be checked, but they only answer "whether it exists." What really helps when you are questioned is another layer of questions—whether it can be verified.
Take the same item, "chart units": a formal self-check looks at "whether units are labeled," while a verifiability self-check looks at "whether the values can be recalculated back." Likewise for "data sources": the former looks at "whether they are listed," the latter at "whether they can be located in the original records." Likewise for "uncertain parts": the former tends to avoid writing them as much as possible, while the latter requires a separate page listing them clearly.
The difference between the two layers is that a formal self-check makes you look finished, while a verifiability self-check lets you answer when questioned.
▲ Figure 4 | Comparison of two organizing methods: piling up by time vs. indexing by conclusion. The difference is not how much you remember, but whether others can use it.
▲ Figure 5 | Example structure of a self-check list: the status only describes "whether I can verify it myself" and does not evaluate whether the conclusion is good or bad
The four status labels in the list are worth noting: Verifiable / In doubt / Unable to verify / Lead. Note that they all describe "whether I can verify it myself," not "whether this conclusion is correct." Whether a conclusion is right or wrong is an academic judgment; whether I can verify it is my own business—mix the two together, and a self-check becomes self-comfort.
04 The order for compressing materials into an outline
▲ Figure 6 | From a pile of materials to an outline, four steps
When it comes to actually doing it, the order is roughly:
First, pour out the conclusions. Without looking at the materials, write down from memory the three things you really want to say. This step forces you to face what you actually want to express.
Then attach evidence. Follow each conclusion with one or two of the strongest pieces of support. Leave aside the ones you can't attach; don't force them.
Mark the uncertainties. List the parts you can't explain clearly on a separate page; don't hide them.
Go back and make trade-offs. At the end, take another look—for conclusions that cannot be supported by evidence, seriously consider deleting them. A presentation with three conclusions that hold up is better than one with eight conclusions, five of which are hanging in the air.
MuchSci's place in this process is mainly the third and fourth checkpoints: organizing the source labeling and uncertain items of materials into a searchable structure, so that when you are asked, you can locate the specific record directly instead of relying on recall. You still do the organizing work yourself; it handles the mechanical work of gathering scattered records and aligning definitions.
The content of this article has also been organized into a set of vertical graphics, making it easy to flip through in fragmented time or keep in your phone's photo album.
05 Boundaries
① This article discusses material preparation habits and does not constitute a promise of results. No matter how clearly the materials are organized, it cannot guarantee that the conclusions hold, let alone guarantee that the presentation will go well.
② The self-check list does not make judgments for you. The list is only responsible for marking "what has not yet been verified"; whether a conclusion can be published or should be written into a paper still must be judged by the researcher.
③ Organizing data and results must comply with the current regulations of your institution and the journal. This article discusses ways to organize public information and your own records, and does not cover the specific requirements of data processing standards.