NR-518 Week 5 Epidemiologic Study Designs and Causal Reasoning: How to Write It

What NR-518 Week 5 asks for

The territory is the design ladder and the language that sits on each rung. Descriptive work says who, where and when. Analytic observational work adds a comparison group: cross-sectional studies photograph exposure and outcome together, case control studies start from the outcome and look backwards at exposure, and cohort studies start from exposure and wait. Experimental work assigns the exposure and is the only rung that removes the reasons two groups differed before anyone intervened. The measures follow from the rung: odds ratios where the design started from cases, risk and rate ratios where the design started from exposure and followed people forward.

Graduate writing about design is not a taxonomy exercise. The reasoning that scores is a chain: the question, the design able to answer it, the measure that design produces, the alternative explanations that remain, and the strength of claim those alternatives permit. Confounding, selection and recall of past exposure are the three usual suspects, and a paper that names one and says which direction it would push the result has done more analytic work than a paper that lists all three abstractly.

Deliverables at this stage usually ask you to read published work and account for it, either as an appraisal of one study or as a short synthesis of several. The rubric rows tend to reward accuracy about what a design permits far more than enthusiasm about a finding. If your section runs a discussion this week, settle the design and the measure before drafting, since posts do not reopen once submitted in Canvas and a misnamed design is quoted back at you all session.

The NR-518 Week 5 method, step by step

Six moves that take a published finding and turn it into a claim you can defend at exactly the strength the evidence supports.

1. Read your week's rubric for the verb it wants

   Describe, appraise, compare and evaluate ask for different work. An appraisal row wants the flaws named and weighed. A comparison row wants two designs set against each other on a stated criterion. Reduce each row to its verb before you read a single article, and let the verbs become your headings in the printed order.

2. Identify the design from the methods, not the title

   Articles describe themselves loosely. Read the methods and answer two questions: did the investigators start from the outcome or from the exposure, and did anyone assign the exposure. Those two answers place any study on the ladder without needing the authors to name it for you.

3. Name the comparison group out loud

   Write one sentence saying who the exposed or affected people were compared against and how those comparators were found. A study with no comparison group cannot support an association at all, and where the comparators were assembled differently from the cases, you have located the paper's main weakness in a single line.

4. Read the measure of association with its interval

   Report the point estimate and the confidence interval together, and say in words what the interval covers. An interval that includes the null value tells you the data is consistent with no difference, and writing that plainly is more accurate than reporting a promising point estimate on its own.

5. Give the alternative explanations a paragraph each

   Take confounding, selection and measurement in turn, and for each say whether it is plausible here and which way it would bend the result. Vague acknowledgement earns little; a named confounder with a stated direction earns the analysis row.

6. Set the claim at the level the design can pay for

   Close with a sentence that states what may now be said. Observational work supports occurred more often among and was associated with. Assignment of exposure supports reduced and caused. Writing the ceiling explicitly stops the conclusion from drifting above the evidence, which is where most marks are lost at this stage.

A layout and word budget for a study appraisal

Sized for a piece of roughly 1,100 to 1,300 words. This is our drafting frame rather than a university-issued form, and your week's rubric outranks it wherever the two disagree. Scale each target proportionally if your assigned length differs.

Evidence craft for design and causal writing

Put the design and the sample in front of every finding. Write that a cohort of 8,400 adults was followed for six years, then the result. In a paper about evidence, provenance is not background material, it is the sentence that tells your reader how much weight the next clause can bear.

Match the measure to the design. A study that started from cases produces an odds ratio; a study that followed exposed and unexposed people produces risk or rate ratios. Reporting the wrong measure name for a design tells a grader the methods section was skimmed, and it is a fast row to lose.

Keep relative and absolute effects side by side. A doubling of risk means something different when the baseline is two in a hundred thousand than when it is two in ten. Give both wherever the source allows, because population health decisions are made on the absolute figure.

Reserve causal verbs for designs that earn them. Association is what observational work supplies, however large the study and however strong the mechanism sounds. Where you want to argue causation from observational evidence, do it explicitly through stated criteria such as temporality, strength, consistency and dose response, rather than by quietly upgrading the verb.

Five mistakes that cost points in this week's territory

• Naming the design from the abstract. Authors describe their own work generously. The methods section decides the rung, and a misplaced study invalidates every judgment that follows.
• A point estimate with no interval. Reporting the ratio and dropping the interval hides exactly the uncertainty the appraisal rows are asking you to weigh.
• Confounding mentioned but never directed. Saying that confounders may exist costs nothing and earns nothing. Name one and say whether it inflates or deflates the finding.
• Temporality assumed from a snapshot. Cross-sectional data cannot say which came first, so any sentence implying sequence has outrun the design.
• Statistical significance read as importance. A tiny difference can be significant in a large study and irrelevant to a population program, and this course grades the second judgment.

Before you submit

• The design is stated with the methods detail that identifies it
• The comparison group is described, including how it was assembled
• Every measure of association appears with its confidence interval
• At least one alternative explanation is named with a stated direction of bias
• No verb in the conclusion exceeds what the design supports
• Every reference appears in the text and every in-text citation appears in the list

Questions students ask about this stage

Keep going