HS340 Unit 5 Bias and Confounding Memo Example

Reviewed by Elspeth Marlowe, MSN, RN Epidemiology Purdue University Global Free custom sample in 24 to 48h

Residence hall students at a composite university were recorded with influenza-like illness 1.7 times as often as off-campus students, and the HS340 Unit 5 bias and confounding memo tells the campus health director why that figure runs high. Purdue Global HS 340 asks BS in Health Science students to separate a confounder from a distortion in the data. The memo shows that only a quarter of hall residents were vaccinated against 58 percent off campus, and that stratifying by vaccination brings the ratio to about 1.30 in both groups. It then shows that hall residents who fell ill visited the nearby clinic far more often, which inflates the hall count further. It recommends a hall vaccination push before any occupancy cap.

CourseHS340 Epidemiology
UnitUnit 5
Paper typeBias and confounding memo
LengthAbout 780 words, 5 pages double-spaced
FormatAPA 7 student paper
SchoolPurdue University Global
ProgramBS in Health Science
UpdatedOctober 2026

Free sample paper for HS340 Unit 5

1

Why the Residence Hall Figure Runs High: A Unit 5 Bias and Confounding Memo to a Campus Health Director

[Student Name]

Purdue University Global

HS340: Epidemiology

Unit 5 Bias and Confounding Memo

[Instructor Name]

[Date]

The university, its students and every count are composite and chosen to illustrate the methods; the bias and confounding sources cited are real.

What this part is doingNaming both problems in the lead gives the director the conclusion before the numbers, which is how a memo to a busy reader should open.
2

Memo Header

To: Director, Student Health Services From: Epidemiology intern, Office of Campus Health Date: [date] Re: Influenza-like illness in residence halls compared with off-campus housing, last season

Bottom line: residence halls do carry somewhat more influenza-like illness than off-campus housing, about 30 percent more, not the 71 percent the clinic records suggest, and the true gap is probably smaller still.

Background

Of [4,000] undergraduates followed through last season, clinic records show [176] of [1,600] hall residents with influenza-like illness, 11.0 percent, and [154] of [2,400] off-campus students, 6.4 percent. The crude risk ratio is 11.0 divided by 6.4, or 1.71. That figure has been quoted in a proposal to cap hall occupancy during flu season.

The Confounder: Vaccination

Vaccination lowers the risk of influenza-like illness and is very unevenly spread between the two groups. Only a quarter of hall residents, [400] of [1,600], had a recorded flu vaccine, against 58 percent of off-campus students, [1,392] of [2,400]. A confounder must be linked to the exposure, linked to the outcome and not lie on the path between them (Jager et al., 2008). Vaccination meets all three: it differs by housing, it affects illness, and living in a hall does not cause a student to go unvaccinated, though it may make the clinic less convenient.

Splitting the students by vaccination status removes that imbalance.

Unvaccinated students: hall residents [156] of [1,200], 13.0 percent; off-campus students [101] of [1,008], 10.0 percent. Risk ratio 1.30.

Vaccinated students: hall residents [20] of [400], 5.0 percent; off-campus students [53] of [1,392], 3.8 percent. Risk ratio 1.31.

Within each vaccination group the gap is about 30 percent, and the crude figure of 71 percent appears only when vaccinated and unvaccinated students are pooled, because the halls hold far more of the unvaccinated.

What this part is doingThe highlighted sentence shows the director exactly where the extra 41 points came from, which is the core of a confounding argument.
3

Because the two stratum-specific ratios are almost identical, a single adjusted figure of about 1.30 is a fair summary, the situation in which a pooled stratified estimate is appropriate (Celentano & Szklo, 2019).

The Distortion: Who Gets Recorded

The clinic records count students who came to the clinic. They do not count students who were ill and stayed home. An end-of-term survey of a random [400] students asked about flu-like symptoms during the season and about clinic visits. Among students reporting symptoms, [62] percent of hall residents had visited the clinic, against [38] percent of off-campus students. The clinic sits in the middle of the halls; off-campus students have to drive, park and miss class to reach it.

That is information bias, specifically differential ascertainment: the outcome is detected more completely in one exposure group than the other (Delgado-Rodríguez & Llorca, 2004). It inflates the hall figure. If the survey's visit rates hold for the whole population, correcting for them would lower the adjusted ratio of 1.30 further, perhaps close to 1.0. The survey is small, and its own recall of symptoms is imperfect, so the memo does not offer a corrected figure, only the direction.

What This Means for the Proposal

An occupancy cap is a costly response to a gap that is real but modest, and possibly smaller than 30 percent. Two cheaper steps follow directly from the analysis. First, raising vaccination in the halls from 25 percent toward the off-campus level would close much of the crude gap, since unvaccinated students in any housing fall ill at more than twice the rate of vaccinated ones. Second, future comparisons should use a symptom survey that reaches both groups equally, not clinic records.

What this part is doingTurning the analysis into two cheaper actions shows the director why the memo matters, not just what the numbers are.
4

Limits

The vaccination records may miss students vaccinated at pharmacies off campus, which would mean more off-campus students were vaccinated than recorded. Influenza-like illness is a symptom definition, not laboratory-confirmed influenza. And the analysis adjusts for only one confounder; differences in age, class year or chronic illness between the groups were not examined.

Recommendation

Do not cite 1.71 as the effect of hall living. Cite about 1.3, with the caveat that unequal clinic use probably inflates even that. Fund a hall vaccination push before considering occupancy limits. A clinic in the hall lobbies during the first two weeks of the fall term, when students are moving in and the clinic is quiet, would reach the largest number of unvaccinated students at the lowest cost, and next season's figures would show whether the gap narrows.

5

References

Celentano, D. D., & Szklo, M. (2019). Gordis epidemiology (6th ed.). Elsevier.

Delgado-Rodríguez, M., & Llorca, J. (2004). Bias. Journal of Epidemiology and Community Health, 58(8), 635-641. https://doi.org/10.1136/jech.2003.008466

Jager, K. J., Zoccali, C., MacLeod, A., & Dekker, F. W. (2008). Confounding: What it is and how to deal with it. Kidney International, 73(3), 256-260. https://doi.org/10.1038/sj.ki.5002650

The Purdue Global HS340 Unit 5 instructions

The fifth unit typically asks you to explain, to a nonspecialist decision maker, why a reported association may be misleading, distinguishing confounding from bias and showing at least one correction. Memos usually run about two pages. Instructors look for a confounder identified with the three conditions it must meet, a stratified or adjusted result that shows its effect, and a source of bias named precisely, such as differential ascertainment, with its likely direction. Many sections reward a clear bottom line at the top and a recommendation that follows from the analysis. Keep the arithmetic visible but brief, and explain terms in plain words. Cite methods sources for the definitions of confounding and bias.

How the HS 340 Unit 5 example is put together

A To and From block heads the memo, followed by a bottom line giving the corrected size of the gap. Background reports the crude counts and ratio and notes the proposal that prompted the analysis. The confounding section shows the vaccination imbalance, applies the three conditions for a confounder, and gives the stratified counts and ratios for unvaccinated and vaccinated students; the highlighted sentence explains where the extra gap came from. A separate section on the records describes an end-of-term survey showing unequal clinic use and names it differential ascertainment, with its direction. Sections on what this means for the proposal, the analysis's limits and a recommendation close the memo, pointing to vaccination rather than an occupancy cap.

Where the points sit in the HS340 Unit 5 rubric

Bias and confounding memos tend to be graded on correct identification of the confounder, a correct stratified or adjusted analysis, correct identification and direction of bias, clarity for a lay reader and a sound recommendation. The analysis of confounding usually carries the largest share of points, since showing the ratio change after stratification is the core skill of the unit. Bias credit goes to naming the specific mechanism and saying which way it pushes the estimate. Clarity is rewarded when the director could act on the memo without a statistics course. The recommendation earns points when it follows from the numbers. Format, including a header and a bottom line, and correct citations for definitions round out the marks.

Common HS340 Unit 5 mistakes, and how to avoid them

Students frequently call every problem bias, or every third variable a confounder, without checking the conditions. Another common slip is adjusting for something on the causal pathway, which removes part of the real effect. Some memos stratify correctly but never say what the stratified result means for the decision at hand. Others give a corrected figure with more confidence than the data allow. A memo that buries its conclusion on the second page loses a busy reader. If your scenario is different, perhaps a workplace injury comparison or a hospital infection rate, send it and we can help you find the confounder, the distortion and a bottom line a director can use.

Get an HS340 Unit 5 example written to your instructions

Whatever association the HS340 Unit 5 scenario presents, the memo can separate the distortion from the third factor in the same way. Attach the scenario, any tables, the named reader and the rubric. The first custom sample is free, arrives within 24-48 hours, and gives each problem its own evidence and its own fix. The paper above is an original model document written by our desk, not a submitted student paper and not an official Purdue University Global document.

Request this unit free All HS340 units Finished samples

More HS340 papers and BS in Health Science samples

HS340 Unit 5 questions, answered

What is the quickest test for whether something is bias or confounding?

Ask whether the problem would still exist if every measurement were perfect. A confounder would: vaccination really does differ between hall and off-campus students and really does lower illness. Bias would not, because it lives in the way a study selected or recorded its participants. Uneven clinic use disappears once illness is counted the same way for everyone, which marks it as information bias.

Why does the example stratify rather than adjust with a model?

Stratification shows the confounding happening, which suits an introductory memo and a nontechnical reader. Two strata with identical ratios of 1.30 make the point more plainly than a coefficient would. A regression model would reach a similar answer, and some sections expect one, but the director needs to see why the figure moved, and a small table shows that directly.

Where can I find a free HS340 Unit 5 sample paper?

You can read the whole Unit 5 bias and confounding memo on this page, explaining why a residence hall flu figure overstates the gap.

What three conditions make a variable a confounder?

It has to travel with the exposure, affect the outcome in its own right, and sit outside the chain that runs from exposure to outcome.

What is differential ascertainment?

A form of information bias in which the outcome is detected more completely in one exposure group than in the other, for example because one group lives closer to the clinic.