quantitative research Flashcards

Question

what is a cross sectional study?

Answer 1

Data is collected at a single point in time to provide a snapshot of a population. Collect data on exposure and disease simultaneously Grouping of subjects by exposure and an outcome status after collecting data. Sampling - based on demographic characteristics – age, ethnicity, location The main tool is a survey / Questionnaire. Compute prevalence (%) to estimate the magnitude of a problem.

Answer 2

Quick, easy to conduct, and inexpensive Measure prevalence of disease Describe patterns of disease occurrence Can study several diseases or exposures

Answer 3

Selection bias Retrospective recall of timing of exposure (recall bias) Cannot argue any temporal relationship between the two variables (exposure must precede disease to be causal)

Answer 4

Select subjects based on the disease status (YES/NO) Have the disease (Cases) or not (Controls) Compare groups to determine how many in each group had (in the past)the exposure – (Odds ratio).

Answer 5

Relatively inexpensive & fast Select cases from disease registries or records Great for rare diseases Can study multiple exposures to one disease

Answer 6

Appropriate control group is hard to find Potential for high recall bias Weak in establishing temporality: Can assume, but cannot confirm the temporal sequence of exposure and disease

Answer 7

Subject selection based on exposure (exposed vs. unexposed) [e.g., exposure to radiation] Follow people prospectively until disease occurs or the study ends –TEMPORALITY Compare disease incidence between exposed and unexposed (Relative Risk/Ratio)

Answer 8

Data is gathered for the same subjects repeatedly over a period. It can extend over years or even decades. The same individuals are observed over the study period.

Answer 9

Temporal sequence Useful when studying rare exposures You can study multiple disease outcomes Can directly compute disease risk (incidence)

Answer 10

Expensive due to follow-up period. Tracking subjects is a challenge (dropouts, lost). Potential for loss to follow-up Dropouts may introduce selection bias. (Attrition Bias)

Answer 11

Considered the ‘GOLD standard’ among all designs. Researcher manipulates exposure. Treatment group & Control group / Placebo or Current treatment. Establish a temporal relationship - exposure and then assessing a defined outcome.

Answer 12

Blinding : Single/double/triple Randomization: Random assignment of subjects to groups. Use of Placebo.

Answer 13

- blind - randomised - placebo

Answer 14

Reduced confounding (randomization) Reduce bias by blinding Temporal sequence Strongest evidence of a direct effect

Answer 15

Not as much control over outside forces Possibility for non-compliance with assigned treatment Loss to follow-up Expensive design and long follow-up

Answer 16

surveys and questionnaires observations methods experimental methods

Answer 17

Structured tools used to collect large amounts of data from participants. Example: Patient satisfaction surveys with Likert scale responses from 1 to 5.

Answer 18

Quantitative observation involves systematically recording behaviours or events as they occur. Example: Counting the number of times hand hygiene is performed in a shift.

Answer 19

Involves manipulating variables in a controlled setting to observe outcomes. Example: Administering a new drug to one group and a placebo to another.

Answer 20

Instrument error Observer error Participant misunderstanding Calibration issues Example Blood pressure cuff not calibrated = false hypertension diagnosis.

Answer 21

Missing data Incomplete responses Data entry errors Inconsistent definitions Poor data quality = poor clinical decisions.

Answer 22

Descriptive statistics (means, prevalence) Inferential statistics (t-test, chi-square, correlation) Effect size Confidence intervals

Answer 23

Extreme values far from other observations Can be real OR data errors Example: One patient weight recorded as 600 kg instead of 60 kg.

Answer 24

Indicate data entry error Distort the mean Suggest non-equivalent groups at baseline Influence statistical results Important: The mean is sensitive to outliers. The median is more robust.

Answer 25

Correlation ≠ causation Overgeneralisation Ignoring confounders

Answer 26

the consistency of a measurement

Answer 27

Test–retest reliability Inter-rater reliability Internal consistency

Answer 28

Inconsistent scoring = inconsistent care Poor reliability weakens evidence Reliability affects treatment guidelines

Answer 29

stability over time

Answer 30

stability across users - do different people get different scores

Answer 31

stability within the tool do all question in the survey measure the same thing

Answer 32

does the instrument measure what it claims to measure

Answer 33

Content validity Criterion validity Construct validity

Answer 34

internal validity Are results trustworthy within the study? External validity Can we apply results to our patients or in other locations?

Answer 35

does the tool cover all parts of the problem? - fall risk assessment

Answer 36

does it match a 'gold standard'?

Answer 37

does it measure the right idea?

Answer 38

systematic error that distorts results

Answer 39

Selection bias Recall bias Measurement bias Attrition bias Publication bias Confirmation bias

Answer 40

Attrition bias occurs when participants drop out unevenly between groups Dropouts differ systematically from those remaining

Answer 41

Statistical power Smaller sample = reduced ability to detect real effects Internal validity Groups may no longer be equivalent External validity Final sample may not represent original population

Answer 42

Incorrect treatment decisions Harm to patients Misallocation of resources Poor guideline development

quantitative research Flashcards

(66 cards)