Midterm review: Health Research: Purposes, Data Types, and Study Designs

What is health research?

Systematic inquiry used to generate or apply knowledge that improves health.

What are the five purposes of health research?

Description, association, prediction, causation, and evaluation.

What does description ask?

What is happening?

What does association ask?

What things are related?

What does prediction ask?

What might happen?

What does causation ask?

Does one thing cause another?

What does evaluation ask?

Does something work?

What type of data does quantitative research use?

Numerical data analyzed with statistics.

What type of data does qualitative research use?

Words, narratives, experiences, and other textual data.

What reasoning is associated with quantitative research?

Deductive reasoning.

Which generally uses larger samples?

Quantitative research.

Which generally uses smaller, purposefully selected samples?

Qualitative research.

Which uses open-ended questions, interviews, and focus groups?

Qualitative research.

Which commonly uses structured questionnaires and closed-ended questions?

Quantitative research.

Quantitative analysis usually produces what?

Statistics.

Qualitative analysis usually produces what?

Themes and theories.

What is the target population?

The entire population the researcher ultimately wants the findings to apply to.

What is the accessible population?

The portion of the target population the researcher can actually reach.

What is the sample?

The people who actually participate in the study.

What is the order?

Target population → Accessible population → Sample.

What is generalizability?

The extent to which findings can apply beyond the study sample.

What type of validity asks, "Who can we generalize these findings to?"

External validity.

What is a nominal variable?

Categories with no meaningful order.

Easy memory cue for nominal?

NAME.

Example of nominal data?

Blood type.

What is an ordinal variable?

Categories with a meaningful order or ranking.

Easy memory cue for ordinal?

ORDER.

Example of ordinal data?

Mild → moderate → severe pain.

What is discrete data?

Countable numerical values.

Easy memory cue for discrete?

COUNT.

Example of discrete data?

Number of hospital visits.

What is continuous data?

Numerical data that are measured.

Easy memory cue for continuous?

MEASURE.

Example of continuous data?

Blood pressure.

What is validity?

Whether something actually measures what it is intended to measure.

What is reliability?

Whether a measurement produces consistent results.

Can something be reliable but not valid?

Yes. It can consistently produce the same incorrect measurement.

What are the three major observational designs from your slides?

Cross-sectional, case-control, and cohort.

What separates observational from experimental research?

Observational researchers observe existing exposures; experimental researchers assign an intervention/exposure.

What is a cross-sectional study?

Exposure and outcome are measured at approximately the same point in time.

Memory cue for cross-sectional?

SNAPSHOT 📸

What does a cross-sectional study commonly estimate?

Prevalence.

Major limitation of cross-sectional research?

Temporality is difficult to establish—you may not know which came first.

What does a case-control study start with?

The outcome/disease.

What happens after cases and controls are identified?

Researchers look backward for previous exposure.

Memory cue for case-control?

OUTCOME FIRST → LOOK BACK ⬅️

What are cases?

People who have the outcome/disease.

What are controls?

People who do not have the outcome/disease.

What type of disease is case-control especially useful for?

Rare diseases.

Major biases in case-control studies?

Recall bias and selection bias.

What does a cohort study start with?

Exposure status.

What happens next in a cohort study?

Exposed and unexposed groups are followed toward an outcome.

Memory cue for cohort?

EXPOSURE FIRST → OUTCOME ➡️

What are the two major types of cohort studies?

Prospective and retrospective.

What does prospective cohort mean?

Exposure is identified and participants are followed forward into the future.

What does retrospective cohort mean?

Existing historical records are used to reconstruct exposure and subsequent outcomes.

Major bias/problem in cohort studies?

Loss to follow-up.

What makes research experimental?

The researcher assigns an intervention/exposure.

What is an RCT?

Randomized controlled trial.

What is the key feature of an RCT?

Participants are randomly assigned to groups.

Why is randomization important?

It helps reduce bias/confounding and improves comparability between groups.

Which design provides the strongest evidence for causal inference?

Randomized controlled trials.

What is quasi-experimental research?

An intervention is introduced, but random assignment is not used.

Biggest clue for quasi-experimental vs. RCT?

No randomization = quasi-experimental.

What is a field trial?

A preventive intervention tested in a real-world, generally nonclinical setting.

What is a community trial?

An intervention assigned or implemented at the community/group level.

What is recall bias?

Participants inaccurately remember past information or exposures.

Which design is particularly vulnerable to recall bias?

Case-control.

What is selection bias?

The way participants enter or remain in a study creates systematic differences.

Easy question for selection bias?

"WHO got into the study?"

What is attrition/loss-to-follow-up bias?

Participants drop out over time, potentially making the remaining group systematically different.

Which design is especially vulnerable to loss to follow-up?

Cohort studies.

What is confounding?

A third factor that is associated with both the exposure and outcome and distorts their apparent relationship.

Example: Coffee appears associated with heart disease, but coffee drinkers are also more likely to smoke. What could smoking be?

A confounder.

Major experimental-study problems from your comparison slides?

Attrition and noncompliance.

Major cross-sectional bias concern from your comparison slides?

Survivor bias.

What does correlational research examine?

Associations or relationships between variables.

Can correlation establish causation?

No.

What does a positive correlation mean?

As one variable increases, the other tends to increase.

What does a negative correlation mean?

As one variable increases, the other tends to decrease.

What does zero correlation mean?

No apparent relationship between the variables.

Pearson's r is generally used for what?

Continuous, normally distributed variables with a linear relationship.

Spearman's rho is generally used for what?

Ranked/ordinal data.

Major limitation of correlational research?

It cannot establish causality and may be affected by confounding.

What is a systematic review?

A structured synthesis of research using a comprehensive search and explicit inclusion/exclusion criteria.

What is a meta-analysis?

Statistical pooling of quantitative results from multiple studies.

Key clue for meta-analysis?

Combines numbers/effect sizes statistically.

What is a scoping review?

A broad review that maps available evidence and identifies gaps.

Systematic review vs. meta-analysis?

Systematic review systematically synthesizes evidence; meta-analysis statistically pools results.

Major benefit of evidence synthesis?

It summarizes the overall body of evidence to inform practice, guidelines, and policy.