SAMPLE SIZE

to avoid bias in interpreting results

Why do we define the sample size before starting a study?

the results cannot be generalized to the population

What is one consequence of including very few subjects in a study regarding generalizability?

the sample will not represent the size of the target population

Why can results from a study with too few subjects not be generalized?

the study may not be able to detect the difference between test groups

What is a potential issue if very few subjects are included, related to detecting differences?

making the study unethical

What ethical concern arises if a study with too few subjects cannot detect a real difference?

put more individuals to the risk of the intervention

What is one ethical concern with studying more subjects than required?

waste precious resources, including the researchers’ time

Besides ethical concerns, what resources are wasted if more subjects than needed are studied?

crucial

How important is the calculation of an adequate sample size in any clinical study?

calculate the optimum number of participants

What is the process of calculating adequate sample size aiming to determine?

to be able to arrive at ethically and scientifically valid results

What is the goal of calculating the optimum number of participants in a clinical study?

Acceptable level of significance, Power of the study, Expected effect size, Underlying event rate in the population, Standard deviation in the population

What are the general factors that affect the sample size for any study?

expected drop-out rate, an unequal allocation ratio, and the objective and design of the study

What are some additional factors that can be considered when calculating the final sample size?

usually dealt with on a study-by-study basis

How are additional factors like drop-out rate typically handled when calculating sample size?

level of significance

What statistical concept is commonly known as the "p" value?

prior to starting a study

When do we set an acceptable value for the level of significance (p-value)?

we are ready to accept that the probability that the result is observed due to chance (and NOT due to our intervention) is 5%

What does accepting a p<0.05 as significant mean?

detecting a difference 5 out of 100 times when actually no difference exists

In simpler terms, what does a p-value of 0.05 mean in terms of accepting false positive results?

false positive result

What is it called when a study detects a difference when actually no difference exists?

p = 0.05 or p = 0.01

What p values are conventionally accepted in studies?

5% (or 1%) chance of erroneously reporting a significant effect

What does the conventionally accepted p value of 5% (or 1%) mean?

detection of a "false positive" test

What is the level of significance (p-value) equivalent to the detection of?

detection of a "false negative" test

What statistical concept refers to the detection of a "false negative" test?

decide what is the false negative rate we are willing to accept

What must be decided regarding false negatives to make a study adequately powered?

β

What is the statistical letter used to refer to the false negative rate?

proportion of positive instances that were erroneously reported as negative

How is the false negative rate defined in statistics?

(1 –β)

How is the "power" of the study calculated?

probability of failing to detect a difference when actually there is a difference

What is the power of a study defined as?

decrease

How does the power of a study change as the chances of committing a Type II error decrease?

Type II error

Committing a false negative error is also known as committing what type of statistical error?

80%

What power percentage do most studies usually accept?

we are accepting that one in five times (that is 20%) we will miss a real difference

What does accepting a power of 80% mean?

depending on the study

Can the conventional 80% power level be modified?

randomized controlled trials

In what type of study is the expected effect size usually accounted for?

indicators of the claimed effect size

In day-to-day examples like weight loss or score improvement, what do values like 10 kg or 10% represent?

the difference between the value of the variable in the control group and that in the test drug group

In statistics, how is the "effect size" defined?

absolute difference or the relative difference

How can the difference between control and test groups (effect size) be expressed?

10 kg

In the weight loss example where the control group lost 10 kg and the test group lost 20 kg, what is the absolute effect size?

50%

In the weight loss example, what is the relative reduction with the test intervention if the control group lost 10 kg and the test group lost 20 kg?

based on previously reported or preclinical studies

How can the effect size be estimated before conducting a study?

less

If the effect size is large between study groups, is the required sample size generally more or less?

large

If the effect size is small between study groups, is the required sample size generally more or less?

a smaller sample would be needed

If studying the association between smoking and lung cancer, where a large effect size has been shown, would a larger or smaller sample be needed?

a larger sample size would be required

If studying the association between smoking and brain tumor, where the effect size is unknown or small, would a larger or smaller sample size be needed?

Underlying event rate in the population

Besides level of significance, power, effect size, and standard deviation, what other factor listed is extremely important for sample size calculation, especially in observational studies?

observational studies

In what type of study is the underlying event rate in the population usually accounted for?

Unlike the level of significance and power

How is the underlying event rate different from the level of significance and power in terms of how it's determined?

estimated from previously reported studies

How is the underlying event rate of a condition determined?

Standard deviation (SD or Σ)

What statistical measure is important for calculating sample size and represents the dispersion or variability in data?

measure of dispersion or variability in the data

What does the standard deviation represent?

anticipate the variation in the measures that are being studied

What does an investigator need to do regarding standard deviation when calculating sample size?

smaller

Would a smaller or larger sample size be required if the population is more homogenous?

a smaller variance or standard deviation

What characteristic of a homogenous population leads to requiring a smaller sample size?

the difference between the two groups would be masked by the inherent difference between them because of the variance

If studying weight effect in a population with large weight variation (large SD), why would a larger sample size be needed?

get a tighter and more homogenous group

If taking a sample from a population with a smaller weight range (smaller SD), what kind of group would naturally result?

reducing the standard deviation

What happens to the standard deviation when sampling from a more homogenous group?

therefore the sample size

Besides reducing standard deviation, what does sampling from a more homogenous group also reduce?

manually

How was sample size often calculated previously?

a slew of different formula

What was needed to calculate sample size manually?

depending on the study design and the factors identified

What did the specific formula needed for manual sample size calculation depend on?

rely on software and apps

In the advent of newer technology, what can investigators use to help calculate sample size?

determine which study design we are implementing

What is still very important to do, even when using automated calculations?

automated calculations still rely on the different study designs

Why is it still important to determine the study design when using software for sample size calculation?

OpenEpi

What specific software is recommended for epidemiologic statistics?

free and open source software for epidemiologic statistics

How is OpenEpi described?

run from a web server or downloaded and run without a web connection

How can OpenEpi be accessed or used?

compatible with recent Linux, Mac, and PC browsers, regardless of operating system

What is the compatibility of OpenEpi in terms of operating systems and browsers?

Understand the principles discussed

What is important to understand about sample size calculation, even if using software like OpenEpi?

Bias in interpreting results

What is avoided by defining the sample size before a study?

Generalize results to the population

What capability is lost if a study includes too few subjects?

The study becomes unethical

What is an ethical implication of a study being unable to detect a difference due to low sample size?

Wasting precious resources and putting individuals at risk

What are the negative consequences of studying more subjects than required?

Optimum number of participants required

What is the main outcome of the sample size calculation process?

Ethically and scientifically valid results

What kind of results are achieved by calculating an adequate sample size?

Probability the result is due to chance

What does the p-value represent in relation to the intervention?

False positive

What type of result is indicated when a study detects a difference that doesn't exist?

Type I error

Committing a false positive error is also known as committing what type of statistical error?

False negative

What type of test detection does 'power' relate to?

β

What Greek letter represents the false negative rate?

1-β

How is power expressed mathematically?

Missing a real difference

In simple terms, what does the power of a study relate to?

The difference between the value of the variable in the control group and the test group

How is effect size defined statistically?

Based on previous studies

How is the effect size typically estimated for a study?

Less sample size

What does a large expected effect size imply about the required sample size?

Larger sample size

What does a small expected effect size imply about the required sample size?

Observational studies

In which type of studies is the underlying event rate particularly important for sample size?

Prevalence rate

What is another term for the underlying event rate of the condition under study?

From previously reported studies

How is the underlying event rate determined?

Measure of dispersion

What aspect of data does standard deviation quantify?

More homogenous

If a population has a smaller standard deviation, what can be said about its homogeneity?

Smaller sample size

What does a smaller standard deviation generally allow for in terms of sample size?

Variability in measures being studied

What must an investigator anticipate regarding the data when considering standard deviation for sample size?

Manual calculation with different formulas

Before newer technology, how was sample size calculated?

Study design and influencing factors

What determined the specific formula needed for manual sample size calculation?

Software and apps

What technological advancements aid sample size calculation now?

The study design

Even with automated tools, what critical piece of information is still needed for calculation?

OpenEpi

What is the recommended free and open source software for epidemiologic statistics?

It can be used online or downloaded for offline use

How flexible is OpenEpi in terms of access?

To avoid bias

One reason to determine sample size beforehand is to avoid what?

Cannot be generalized

What happens to the results if the sample is too small to represent the target population?

Put subjects at risk unnecessarily

What is an ethical problem with having too large a sample size?