week 1-6 Flashcards
what should be used to represent categorical data
A) box plot
B) bar chart
recall! categorical data represents characteristics or groups!
B) bar chart
as it clearly represents proportions or frequencies of DIFF categories
How can diversity impact random error in measurement?
A) Increasing diversity can decrease random error.
B) Diversity has no impact on random error.
C) Diversity only affects systematic error.
D) Increasing diversity can increase random error.
D) Increasing diversity can increase random error.
- random error shifts each measurement from its true value by a random amt and in a random direction
- increased diversity
→ increased variability
⇒ increase random error
How does increasing sample size typically affect random error?
A) It has no effect on random error.
B) It increases random error.
C) It systematically alters the type of random error.
D) It reduces random error.
D) It reduces random error.
- random error shifts each measurement from its true value by a random amt and in a random direction
- larger sample size
→ average out fluctuations (i.e. they cancel each other out)
⇒ average closer to the true score
What is a key characteristic of systematic errors?
A) They skew measurements consistently away from the true value.
B) They can easily be addressed via statistical inference.
C) They are an inherent part of all measurement processes.
D) They shift each measurement by a random amount.
A) They skew measurements consistently away from the true value.
systematic errors occur in the same direction and same magnitude every time a measurement is taken
← due to error in measurement tool or process
Which of the following is a method used to address systematic errors?
A) Confidence intervals
B) Study designs
C) Statistical inference
D) Hypothesis testing
B) Study designs
only way to reduce systematic error as
- study designs are like blueprints for how you will conduct research
- well-chosen study designs proactively eliminates or minimises various sources of systematic error
includes random sampling
and standardised protocols and training
A researcher plans to sample 500 people to determine the average age in a country, but a co-researcher suggests using the national birth and death registry instead. Why might this suggestion eliminate the need for statistical inference?
A) Because the researcher would save the work, time and expense of the sampling.
B) Because the registry data is more prone to systematic error.
C) Because the registry data would include only a sample of the population.
D) Because the registry data contains information on the entire population.
D) Because the registry data contains information on the entire population.
Statistical inference is only needed when
we do NOT have acces to entire population
and therefore need to draw conclusions about the population based on data from a sample
A researcher is determining whether people in Town A have a different mean BMI compared to people in Town B. Which of the following represents the null hypothesis (HO)?
A) People in Town A have a lower mean BMI than people in Town B.
B) There is no difference in the mean BMI between Town A and Town B.
C) People in Town A have a higher mean BMI than people in Town B.
D) There is a difference in the mean BMI between Town A and Town B.
B) There is no difference in the mean BMI between Town A and Town B.
Null hypothesis (H0) is the hypothesis that suggests theat there is NO effect or difference bet your observations
What will the H1 be if you are comparing between 3 groups
H1 will be the hypothesis that
the mean of AT LEAST 1 group is different from the others
H0 will still be the same,
i.e. there is NO difference in the mean across the 3 groups
In hypothesis testing, what does it mean to ‘reject the null hypothesis’?
A) The researcher failed to analyze the data correctly.
B) There is insufficient evidence to say that the null hypothesis is false.
C) There is sufficient evidence to say that the null hypothesis is false.
D) There is insufficient evidence to support the alternative hypothesis.
C) There is sufficient evidence to say that the null hypothesis is false.
There is insufficient evidence to say that the null hypothesis is false
= NOT rejecting H0
When is it necessary to check for normality, while comparing 2 interventions?
A) Before identifying the outcome variable.
B) Before running statistical tests.
C) If the outcome variable is numerical.
D) After identifying independent and dependent variables.
C) If the outcome variable is numerical.
determines whether parametric or non-parametric tests are used afterwards
* numerical = parametric tests
← parametric tests assume normality for their vailidity
Steps of hypothesis testing
1. Identify the (…),
and determine if it is (…) or (…).
2. Determine the number of groups in the (…),
and determine if they are (…) or (…)
(if there are 2 groups).
3. If the outcome variable is numerical, check for normality.
4. Determine test to be used.
- Identify the outcome variable,
and determine if it is numerical or categorical. - Determine the number of groups in the independent variable,
and determine if they are independent or paired
(if there are 2 groups). - If the outcome variable is numerical, check for normality.
- Determine test to be used.
usually if numerical, will involve “mean …” (e.g. mean age)
but if categorical, will involve proportions (e.g. percentage of ppl)
A study aims to evaluate the impact of a hand hygiene training program on reducing hospital-acquired infections (HAIs). What is the most appropriate statistical test to evaluate the decline, six months before vs six months after the program was implemented?
A) ANOVA
B) Two-Sample t-test
C) Chi-Square test
D) Paired t-test
D) Paired t-test
- Outcome variable
= number of HAIs, numerical - Independent variable
= 2 groups, paired - Thus paired t-test
Paired groups are usually from same individual or group,
just before and after
Which factor does NOT influence the width of a confidence interval?
A) Systematic error.
B) Confidence level.
C) Sample size.
D) Variability in the sample data.
A) Systematic error.
- Confidence level:
greater confidence level = greater width
as to be more confident that CI contains true population parameter,
you need to increase the “net” and allow more random errors - Sample size:
greater sample size = smaller width - Variability:
greater variability = greater width
what does the width of confidence interval indicate
amount of random error or uncertainty
⇒ larger width of CI = greater uncertainty/random error
A study estimates the mean systolic blood pressure in a population to be 120 mmHg and a 95% confidence interval of 115 to 125 mmHg. What best describes the meaning?
A) There is a 5% chance the true population mean is outside range of 115-125 mmHg.
B) 95% of individuals’ systolic blood pressure lies between 115 and 125 mmHg.
C) We are 95% confident that the true population mean lies between 115 and 125 mmHg.
D) The sample mean is exactly 120 mmHg.
C) We are 95% confident that the true population mean lies between 115 and 125 mmHg.
- confidence interval involves POPULATION mean,
NOT INDIVIDUALS
⇒ (B) is wrong - confidence interval is about confidence, NOT CHANCE
⇒ (A) is wrong
A study comparing systolic blood pressure (BP) between two groups, had a 95% confidence interval for the mean difference of 1-5mmHg. Which of the following is true?
A) We can say with 95% confidence that the true mean range is within the range.
B) There is no way to estimate the true mean difference
C) 95% are absolutely certain all patients taking medication X will be between that range.
D) In future studies, systolic BP range won’t be between 1-5mmHg.
A) We can say with 95% confidence that the true mean range is within the range.
Confidence interval is only for current study,
and does not indicate anything about FUTURE STUDIES
⇒ (D) is wrong
Suppose the p-value were 0.01, what can we interpret from this?
A) There’s a 1% chance of observing a result as was shown.
B) Effectiveness of treatment is guaranteed
C) Drug is inefffective.
D) Effectiveness of the drug can’t be measured.
A) There’s a 1% chance of observing a result as was shown.
Thus, usually, if p < 0.05
-> less than 5% chance of observing a certain result
IF NULL HYPORTHESIS (H0) IS TRUE
=> reject H0
when does a result achieve clinical significance
when ALL the estimates are above the minimally clinically important difference (MCID) threshold
Mrs. Tan, a 65-year-old patient on warfarin, needs a tooth extraction. She is worried about bleeding risks. Which PICO element does ‘the risk of bleeding’ represent in this scenario?
A) Comparison
B) Outcome
C) Intervention
D) Population
B) Outcome
A researcher uses the PICO framework to investigate whether continuing warfarin during dental extraction affects thromboembolic risk compared to discontinuing it for patients with mechanical heart valves. What is the ‘Comparison’ component?
A) Continuing warfarin.
B) Risk of thromboembolic events.
C) Patients with mechanical heart valves.
D) Discontinuing warfarin.
D) Discontinuing warfarin.
When would a researcher use the OR operator in a database search?
A) To combine similar concepts, broadening the search.
B) To search for terms in a specific journal only.
C) To exclude irrelevant search terms from the results.
D) To combine different concepts, narrowing the search.
A) ‘OR’ is used to combine similar concepts or synonyms, thus broadening the search.
What does using the truncation symbol (*) allow you to do?
A) Locate cited references for a specific article.
B) Search for phrases rather than individual words.
C) Sort results by publication date.
D) Find variations of a word with the same root.
D) Find variations of a word with the same root.
A researcher is investigating the impact of diet on both diabetes and hypertension. Which search statement would likely retrieve the most relevant results?
A) diet OR (diabetes AND hypertension)
B) diet AND diabetes AND hypertension
C) diet OR diabetes OR hypertension
D) diet AND (diabetes OR hypertension)
D) diet AND (diabetes OR hypertension)
researcher doesn’t want impact of diet on patient with both diabetes AND hypertension,
but rather impact of diet on patient with diabetes
AND impact of diet on patient with hypertension
When evaluating the sources of infromation, what does CRAAP stand for?
A) Credibility, Resources, Audience, Accuracy, Purpose
B) Currency, Relevance, Autonomy, Accuracy, Purpose
C) Currency, Relavance, Authority, Accuracy, Purpose
D) Credibility, Research, Authority, Appraise, Purpose
C) Currency, Relavance, Authority, Accuracy, Purpose
Incidence vs prevalence
- incidence: frequency (or number) of NEW cases
- prevalence: frequnecy (or number) of total CURRENT cases in a defined population
Which type of study assesses exposure and outcome at the same time?
A) Case-controlled study
B) Randomized controlled trial
C) Cross-sectional study
D) Cohort study
C) Cross-sectional study
⇒ makes it challenging to establish whether exposure preceded outcome
In a case-control study, what is compared?
A) Participants with the outcome to a matched group without the outcome
B) Exposure and outcome measurements taken at the same time
C) Participants with prior exposure to those without the possible risk factor
D) Participants randomly assigned to different intervention groups
A) Participants with the outcome to a matched group without the outcome
⇒ association bet exposure and outcome then determined by comparing exposure histories bet the 2 grps
Which study design involves measuring exposure in the present and following participants over time to observe outcomes?
A) Case-controlled study
B) Cohort study
C) Cross-sectional study
D) Randomized controlled trial
B) Cohort study
Participants are identified and categorized based on their exposure status before the outcomes occur.
NOT RCT as RCT is more ASSIGNING exposures randomly to participants,
then following them to observe outcomes
Which of the following best describes ‘nominal’ data?
A) Ranked categories where the intervals between them may not be consistent
B) Numerical data that can take any value within a range
C) Categorical data where categories have no inherent ranking or order
D) Data with a clear order where intervals are consistent
C) Categorical data where categories have no inherent ranking or order
e.g. types of fruits or colours
“NO-minal data has NO inherent ranking”
What differentiates ordinal data from nominal data?
Choices
A) Ordinal data has consistent intervals between values; nominal data does not
B) Ordinal data has a natural order or ranking, while nominal data does not
C) Ordinal data is numerical, while nominal data is categorical
D) Ordinal data lacks categories, unlike nominal data
B) Ordinal data has a natural order or ranking, while nominal data does not
Ordinal data involves categories that can be ranked, but the intervals between them are NOT necessarily equal or consistent
(e.g., educational levels - primary, secondary, tertiary)
Which of the following is an example of continuous numerical data?
A) Number of patients visiting a clinic each day
B) Number of adverse events reported
C) Customer satisfaction scores
D) Temperature of a patient
D) Temperature of a patient
Continuous numerical data can take any value within a range, including fractions and decimals
What is the primary reason for blinding in a study?
A) To increase the number of participants in the study.
B) To accommodate participants with visual impairments.
C) To remove (interviewer/observer) bias.
D) To make the study results more complex.
C) To remove (interviewer/observer) bias.
In the hypertension study, the Usual Care group had a proportion of 0.31 with controlled blood pressure, and the BP monitoring and Web services training group had a proportion of 0.36. What is the relative difference?
A) About 26%
B) About 21%
C) About 16%
D) About 11%
C) About 16%
The realitve difference between these two groups is calculated by finding the absolute difference between the 2 grps
and then dividing by the “original” group (which is 0.31)
what is one choice of charts that is different for continuous and discrete data
- continuous data: boxplot display
as it summarises the spread of the data - discrete data: frequency table
as it lists distinct values along with their occurrence counts
what does a 95% confidence level mean
if we take multiple random samples and construct confidence intervals, about 95 out of 100 intervals will contain the true population parameter.
what is the null hypothesis value for
1. mean difference
2. odds ratio
- mean difference: 0
(i.e. 0 = no difference) - odds ratio: 1
(i.e. 1 = exposure does not change risk of outcome)
in relation to null hypothesis value, when do we reject H0?
when confidence interval does NOT contain null hypothesis value
- e.g. 95% CI for mean difference is (2.3, 5.7)
→ It does NOT include 0
- e.g. 95% CI for odds ratio is (1.25, 1.85)
→ It does NOT include 1
