Getting Ready To Estimate Sample Size Flashcards
(29 cards)
Are ways of calculating sample sizes true to reality?
No, they only are an estimate.
What is the purpose of calculating sample size?
It gives us a feel of the study. It often reveals that the research design is not feasible, or that different predictor or outcome variables are needed.
Is a research hypothesis and research question the same?
No
Research hypothesis is more specific to the study.
What makes a good hypothesis?
Must be based on a good research question.
Should be simple, specific, and stated in advance.
Can hypothesis contain more than one predictor or outcome variable?
Yes
What should a specific hypothesis not leave out?
Who the subjects will be, what the predictor and outcome variables will be, and how the test of statistical significance will be applied.
What can lead to over interpreting the importance of the findings in a research study?
Hypotheses that are formulated after examination of the data
Should alternative hypotheses be one-sided or two-sided?
Two-sided
What do grant and manuscript reviewers expect, an one-sided or two-sided hypothesis?
Two-sided
Does the research hypothesis need to have the same “sides” as an alternative hypothesis?
No
The research hypothesis can be one-sided but he alternative should always be two-sided.
What type of statistical error is a false-positive?
Type 1 error
What type of statistical error is a false-negative?
A type 2 error
An investigator does not know the exact effect size of a study, so what must she do?
She must choose the size of the association she expects to find.
What is effect size?
It is the association the researcher believes she will find in the study.
When calculating effect size, what should the investigator do first?
She should first try finding data from the literature in related areas to make an informed guess.
If there are no datum for calculating effect size what should a researcher do?
Perform a pilot study
When more than one predictor or outcome, which hypothesis do we use to calculate sample size?
The effect size for the most important hypothesis
What is one of the important aims of sample size planning?
To keep alpha and beta at acceptably low levels without making the study unnecessarily expensive or difficult
What result is better to report if the test was not significant (P>.05) probably due to low sample size?
The results, although suggestive of an association, did not achieve statistical significance (P=.08)
Do statisticians and epidemiologists still favor the P value/hypothesis testing to report precision of the study results?
No, they are now frequently using confidence intervals. Only for sample size planning for analytical studies do p values continue to be the standard.
If there is no association between the predictor and the outcome and we reject the null hypothesis, which type of error have we committed?
A type 1 error
If there is an association between the predictor and the outcome and we fail to reject the null hypothesis, which type of error have we committed?
A type 2 error
What is alpha?
It is the level of statistical significance in a study.
What is the probability of making a type 2 error called?
Beta
