NR-503 · Week 4 of 8 · Study designs and measures of association

NR-503 Week 4 Study Designs and Measures of Association: How to Write It

The short answer

NR-503 Week 4 moves from describing a pattern to testing whether an exposure and an outcome travel together. That means the analytic designs, cohort, case control, cross sectional and the experimental trial, and the measures each one can produce: relative risk from a cohort, an odds ratio from a case control, and prevalence ratios from a survey taken at one moment. The graded skill is matching the measure to the design and reading the number without inflating it. Your section may print this as NR 503 or NR503; it is the same course. Chamberlain publishes no syllabi outside Canvas. The placement here is our teaching judgment from the course's catalog arc; your section's rubric decides what your week actually asks.

NR-503 Week 4 grading scale at Chamberlain, the criterion levels this assessment is scored on, from Chamberlain Tutors
How Chamberlain grades NR-503 Week 4, visualized by Chamberlain Tutors.

What NR-503 Week 4 asks for

Each design answers a different question and pays a different price. A cohort follows people who differ in exposure and waits for the outcome, which gives you incidence in both groups and therefore relative risk, at the cost of time and size. A case control starts from people who already have the outcome, finds comparable people who do not, and looks backward at exposure, which is efficient for rare outcomes and yields an odds ratio rather than a risk. A cross sectional survey measures both at once, which is fast and cannot tell you which came first. A trial assigns the exposure, which is the only design that licenses causal verbs, and is often impossible for the exposures population health cares about.

Reading the measure is the other half. A ratio of one means no association. Above one means the exposed group had more of the outcome; below one means less. The confidence interval matters more than the point estimate for judging what the study established, because an interval that includes one is compatible with no association at all, and a very wide interval is telling you the estimate is unstable regardless of where its centre sits.

Deliverables at this point in an eight-week session frequently take the form of an article appraisal: a published study read and evaluated against the design and measure it used. If your section runs a discussion this week, draft it separately, since posts do not reopen once submitted in Canvas and a design named incorrectly is a visible error.

The NR-503 Week 4 method, step by step

Six moves that read a study for what it can support rather than for what it concluded.

  1. Name the design from the methods, not from the abstract

    Abstracts describe results; methods describe how people were selected and when things were measured. Read who was recruited, on what basis, and at what point exposure and outcome were each recorded. That sequence names the design more reliably than any label the authors used.

  2. Establish the direction of time

    Ask whether exposure was known before the outcome occurred. If it was, the study can speak to sequence. If both were captured at one moment, it cannot, and every later sentence in your appraisal has to respect that limit.

  3. Check that the measure fits the design

    Relative risk requires incidence in both groups, which a case control cannot provide. An odds ratio approximates relative risk only when the outcome is uncommon. Say which measure the paper reported and whether the design could support it, since this is the single most scored judgment at this stage.

  4. Read the interval before the estimate

    Look at whether the interval crosses one and how wide it is, then look at the point estimate. Reading in that order stops you from repeating a headline figure that the study's own precision does not support.

  5. Ask who was left out

    Selection is where analytic studies most often go wrong. Who was eligible, who declined, who was lost during follow-up, and how were controls chosen. A study that lost a third of its cohort is telling you something about its own result.

  6. Write the conclusion the design licenses

    Close your appraisal with a sentence stating what this study can support and what it cannot, in the study's own population. That sentence is the appraisal, and everything before it is the working that justifies it.

A layout and word budget for a study appraisal

Sized for an appraisal of roughly 1,300 to 1,600 words. The frame is ours rather than anything the university issues; where your week's rubric prints its own headings, use those and borrow only the proportions.

SectionWhat belongs in itWord target
The study and its questionCitation, the exposure, the outcome and the population studied, in plain sentences before any judgment.140 to 180
Design identifiedThe design named from the methods, with the evidence for that identification quoted or described.180 to 220
Sampling and selectionWho was eligible, how participants or controls were chosen, and who was lost along the way.220 to 270
Measure of associationThe measure reported, whether the design supports it, the point estimate and the interval read together.280 to 340
Threats to the findingSelection, measurement and timing problems the design leaves open, each tied to something specific in the paper.250 to 300
What it supportsThe claim this study licenses in this population, and the claim it does not, in explicit terms.180 to 220

Evidence craft for analytic studies

Report design and sample before any result. In a case control study with 240 cases and 480 matched controls is the phrase that gives the number after it meaning. A finding quoted without its design is an assertion wearing a decimal point.

Say whether an interval crosses the null. A ratio of 1.4 with an interval running from 0.9 to 2.1 has not established an association, and writing it as though it had is the most consequential misreading available at this stage. One clause reporting the interval prevents it.

Match the verb to the design, every sentence. Observational work supports was associated with, occurred more often among and predicted. Caused, reduced and prevented belong to designs that assigned the exposure. This rule is easy to keep in the methods paragraph and easy to lose in the conclusion.

Keep relative and absolute effects together. A doubling of a rare outcome is still rare. Where the paper reports both a ratio and the underlying frequencies, give the reader both, because a relative measure alone can make a small absolute difference sound like a public health emergency.

Five mistakes that cost points in this week's territory

  • Design named from the abstract. Authors describe their own work loosely, and a design misnamed at the start makes every judgment after it wrong.
  • Relative risk attributed to a case control study. The design cannot produce it, and this error is the fastest way to show a grader that the measures were not understood.
  • A point estimate quoted without its interval. Precision is part of the finding, and dropping it converts an uncertain result into a confident one.
  • Causal verbs over observational data. The most common deduction in appraisal writing, and it usually appears in the final paragraph rather than in the methods discussion.
  • Limitations copied from the authors. Repeating the paper's own limitations paragraph is not appraisal. Say what you found, tied to specific numbers in the tables.

Before you submit

  • The design is identified from the methods section with the evidence stated
  • The timing of exposure and outcome measurement is explicit
  • The measure of association is checked against what the design can produce
  • Every ratio appears with its confidence interval and a reading of that interval
  • Selection, losses and control choice are all addressed
  • The closing claim stays inside what the design and the population support

Appraising a study this week?

Send the article, the prompt and the rubric. A premium original appraisal comes back in 24 to 48 hours with the design identified from the methods and every ratio read alongside its interval.

Questions students ask about this stage

When can an odds ratio be read as a relative risk?
When the outcome is uncommon in the population being studied, conventionally taken as somewhere below about ten percent, the odds ratio approximates the relative risk closely enough to be discussed in similar terms. As the outcome becomes more common the two diverge, with the odds ratio moving further from one than the risk ratio does, which means a paper reporting an odds ratio for a frequent outcome and discussing it as though it were a risk has overstated the effect. In your appraisal, say which situation applies rather than treating the two measures as interchangeable or as unrelated. If the paper gives you enough information, note the outcome frequency and state whether the approximation holds. That one sentence shows the measure was understood rather than recognized, and it is exactly what an appraisal row is looking for.
Which design should I choose if the assignment asks me to propose one?
Let the outcome frequency and the time available decide, then say why in the paper. For a rare outcome, a case control design is usually the only feasible choice, because a cohort would need an impractically large sample to accumulate enough events. For a common outcome with an exposure you can observe going forward, a cohort gives you incidence in both groups and a risk ratio, which is easier to interpret. For a question about how much of something exists right now rather than what leads to what, a cross sectional survey is honest and cheap, provided you write the limitation about sequence rather than hiding it. A trial is the strongest design and is often unavailable for population health exposures on ethical or practical grounds, and saying that explicitly is part of the answer rather than an excuse.
The study reports a significant result. Is that enough to act on?
Statistical significance answers a narrow question about whether a result of that size would be unlikely under a no-association assumption. It says nothing about whether the effect is large enough to matter, whether the population resembles yours, or whether the design could establish sequence at all. Read three things together before you write a recommendation: the size of the effect in absolute terms, the width of the interval, and how closely the study population matches the group you care about. A significant finding in a sample nothing like your population supports a hypothesis rather than an action. In a graduate paper the sentence that scores is the one that names the gap plainly, saying what the finding supports for that population and what would have to be shown before it could guide practice in yours.

Keep going

Online now