Statistics Flashcards
What purpose does a basic knowledge of statistics serve for urologists?
It allows urologists to make informed decisions about treatments, read the medical literature, and conduct their own research and quality improvement studies.
Figure 1: Example of Shrinking Confidence Intervals with Increasing Sample Size
Explain the relationship between sample size and the estimation of the population.
As the sample size increases, the sample becomes a better estimate of the population. The larger the sample size, the more precise the statistics and the reduction in sampling error.
: Define “standard error.”
Standard error represents the standard amount of error expected in a sample given the sample size. It’s a measure of sampling error.
What do confidence intervals represent?
Confidence intervals represent the best estimate where a specific percentage (e.g., 95%) of the sample means would fall around the population value for a specific distribution if multiple samples were selected.
How do the widths of confidence intervals change as sample size changes?
Smaller samples produce wider confidence intervals due to increased sampling error. As the sample size increases, the width of the confidence intervals decreases.
Describe the significance of the confidence interval in relation to the IQ example provided.
In the first scenario with n=9, the confidence interval ranged from 99 to 119 (95% CI: 99-119). Since this range contains 100, there’s no statistical difference. However, with a larger sample of 100 students, the 95% CI shrank to 107-111, indicating a significant difference from the average IQ of 100.
Explain the relationship between a 95% confidence interval and its associated p-value.
A 95% confidence interval corresponds to a p-value of p < 0.05, meaning there’s only a 5% chance that scores outside the 95% CI belong to that specific distribution.
What’s the primary issue with small sample sizes when it comes to statistical significance?
Small sample sizes can result in wide confidence intervals that might be too broad to detect or see differences that are actually present, leading to a Type II error.
Define “statistical power” and explain its importance in research.
Statistical power is the chance of finding statistical significance when a true difference is present. It translates into the likelihood of detecting statistical significance. The power of a study is primarily determined by the sample size, with larger studies having the power to detect small differences due to narrower confidence intervals.
How does the concept of “statistical power” relate to sample size and the resources needed for research?
Larger studies have the power to detect small differences since the width of their confidence intervals will be narrow. However, large sample sizes can be expensive, so researchers plan the number of subjects needed before starting a study, considering the predicted effect of the study and the desired width of the confidence interval.
What is the chance of finding statistical significance for an appropriately powered study?
An appropriately powered study will have an 80% chance of finding statistical significance if the predicted differences are produced by the study.
Define the term “dependent variable” (DV).
The variable being measured for change, influenced by the treatment or other variables in the study, also called the outcome variable.
What is the “independent” variable in research?
It’s either the variable manipulated by the researcher (e.g., group assignment in experimental designs) or the variable(s) believed to influence the dependent variable.
In a study comparing a new drug vs. a placebo for treating erectile dysfunction, identify the independent and dependent variables.
Independent Variable: Group assignment (new treatment vs. placebo). Dependent Variable: Measure of erectile function.
In a study analyzing the impact of age and diabetes on urinary flow rate, specify the independent and dependent variables.
Independent Variables: Age and diabetes. Dependent Variable: Urinary flow rates.
What is an “operational” definition in research?
It’s how a researcher chooses to define or measure the variables in the study.
How can erectile function be measured in a continuum?
Using the Erectile Function Domain (EFD) of the International Index of Erectile Function (IIEF), where lower scores indicate poorer function and higher scores indicate better function.
When using the EFD, how can erectile function be operationalized as a continuous variable?
: EFD scores range from 6-30.
: EFD scores range from 6-30.
By determining if a subject has ED or not based on the EFD. A score less than 26 indicates ED, whereas a score equal to or greater than 26 means no ED.
What does it mean when a variable is measured as a binary variable?
The variable is defined in a way where only two outcomes are possible (e.g., a condition exists or it doesn’t).
How does the measurement of the dependent variable influence the choice of statistical tests?
The way the dependent variable is measured will decide the type of statistical tests used to analyze the data.
Which statistical measures are associated with continuous variables?
Continuous variables produce descriptive statistics such as means, standard deviations, and variances.
What are the statistical tests appropriate for analyzing data with a continuous dependent variable?
T-tests, Analysis of Variance (ANOVA), correlation coefficients, and linear multiple regression.
Which descriptive statistics are produced by binary variables?
Frequency (or the number of subjects in each category), often reported as percent.
Can binary outcomes report statistics like means, standard deviations, or variance? Why or why not?
No. Since binary outcomes are reported in percentages, it is not possible to calculate such statistics.
Fig 2. Decision Tree for Statistical Tests
