Biostatics 3 Flashcards
(108 cards)
1
Q
How is statistics divided?
A
Statistics is divided into descriptive and inferential statistics.
2
Q
What is the purpose of descriptive statistics?
A
Presenting, organizing, and summarizing data.
3
Q
What methods are used in descriptive statistics?
A
Methods include calculating measures of central tendency (mean, median, mode), measures of spread (range, variance, standard deviation), and using graphical representations (histograms, bar charts, pie charts).
4
Q
What is the purpose of inferential statistics?
A
Drawing conclusions about a population based on data observed in a sample.
5
Q
What methods are used in inferential statistics?
A
Methods include hypothesis testing, confidence intervals, and regression analysis.
6
Q
How do descriptive and inferential statistics differ?
A
Descriptive statistics summarize and present data, while inferential statistics make predictions or inferences about a population based on a sample.
7
Q
When would you use descriptive statistics?
A
Use descriptive statistics when you need to describe the basic features of the data in a study.
8
Q
When would you use inferential statistics?
A
Use inferential statistics when you want to make predictions or generalizations about a larger population from a sample.
9
Q
Give an example of descriptive statistics.
A
Calculating the average age of students in a class.
10
Q
Give an example of inferential statistics.
A
Estimating the average age of all students in a school based on a sample of students.
11
Q
What is the purpose of statistical inference?
A
To make conclusions about a population based on data observed in a sample.
12
Q
What do statistical tests allow us to do?
A
They allow us to make statistical inferences.
13
Q
What question does statistical inference help answer?
A
Is the result observed in our sample likely to be a true reflection of the result in the population?
14
Q
What is a study population?
A
The defined patient population we are interested in.
15
Q
What is a sample?
A
A subset of the study population from which data is collected.
16
Q
What are sample statistics?
A
Measures computed from a sample, also known as point estimates.
17
Q
What are true population parameters?
A
Fixed values that are unknown but represent the true characteristics of the entire population.
18
Q
What is the goal of statistical inference?
A
To estimate unknown population parameters and determine the likelihood that the sample results reflect the true population characteristics.
19
Q
What is the basis of statistical tests?
A
Hypothesis testing.
20
Q
What do we start by assuming in hypothesis testing?
A
There is no relationship between the variables in the population.
21
Q
What is the null hypothesis (H0)?
A
The assumption that there is no relationship between the variables.
22
Q
What is the alternative hypothesis (H1)?
A
The assumption that there is a relationship between the variables.
23
Q
Give an example of a null hypothesis in a study about pregnancy planning and age
A
There is no difference in age between women who plan their pregnancy and women who do not plan their pregnancy.
24
Q
Give an example of an alternative hypothesis in a study about pregnancy planning and age.
A
There is a difference in age between women who plan their pregnancy and women who do not plan their pregnancy.
25
What is the goal of hypothesis testing?
To gather statistical evidence to support or refute the null hypothesis.
26
Why is hypothesis testing important in statistical analysis?
It helps determine whether observed data provide sufficient evidence to conclude that a relationship exists in the population
27
What is a Type 1 error?
A Type 1 error occurs when we incorrectly reject the null hypothesis.
28
What is another name for a Type 1 error?
False positive.
29
What does a Type 1 error imply?
It implies finding a relationship or effect when there is none.
30
What is a Type 2 error?
A Type 2 error occurs when we fail to reject the null hypothesis when it is false.
31
What is another name for a Type 2 error?
False negative.
32
What does a Type 2 error imply?
It implies not finding a relationship or effect when there is one.
33
Why is it important to understand Type 1 and Type 2 errors?
To accurately interpret the results of hypothesis testing and understand the potential risks of incorrect conclusions.
34
What is the p-value?
The p-value is the probability of finding the observed relationship, or one more extreme, assuming the null hypothesis is true.
35
What threshold is commonly used to determine statistical significance?
A threshold of 0.05.
36
What does it mean if p ≤ 0.05?
We reject the null hypothesis, and the p-value is said to be "significant."
37
What does a significant p-value (p ≤ 0.05) imply?
It implies there is less than a 5% probability that the observed relationship is due to chance if the null hypothesis were true.
38
What does it mean if p > 0.05?
We do not reject the null hypothesis, and the p-value is said to be "not significant."
39
What does a non-significant p-value (p > 0.05) imply?
It implies there is a high probability that the observed relationship could occur by chance if the null hypothesis were true.
40
Why is the p-value important in hypothesis testing?
It helps determine whether the observed results provide enough evidence to reject the null hypothesis.
41
What is the consequence of a low p-value (≤ 0.05)?
It indicates strong evidence against the null hypothesis, suggesting a true effect or relationship in the population.
42
What is the consequence of a high p-value (> 0.05)?
It indicates insufficient evidence to reject the null hypothesis, suggesting the observed relationship could be due to chance.
43
What is statistical power?
The probability of correctly finding a relationship when the null hypothesis is false.
44
Why is statistical power important?
It indicates the likelihood of detecting an actual effect or relationship in the population.
45
How does sample size affect statistical power?
Larger samples have more statistical power, while smaller samples have less.
46
How does effect size influence statistical power?
Larger effect sizes are easier to detect and require less power, while smaller effect sizes are harder to detect and require more power.
47
How does the level of significance affect statistical power?
A higher level of significance (e.g., 0.05) requires less power, while a lower level of significance (e.g., a smaller p-value) requires more power
48
What is the effect of a large sample size on power?
Increases power, making it easier to detect true relationships.
49
What is the effect of a small sample size on power?
Decreases power, making it harder to detect true relationships.
50
What happens when the effect size is large?
It is easier to find the effect, requiring less statistical power.
51
What happens when the effect size is small?
It is harder to find the effect, requiring more statistical power.
52
What should be considered to ensure adequate statistical power?
Consider the sample size, effect size, and desired level of significance to ensure the study is well-powered to detect true relationships.
53
What does a p-value less than 0.05 indicate?
The sample size and power were adequate to detect the difference.
54
What does a p-value greater than 0.05 indicate?
We cannot say that the null hypothesis is true, but there is not enough statistical evidence to reject it.
55
What are possible reasons for a non-significant p-value?
1. The study may be underpowered to detect the real difference.
2. The sample size may be too small to detect the real difference.
3. There may truly be no difference.
56
How should p-values be reported?
P-values should be reported numerically, not just as "significant" or "not significant."
57
Why is it important to provide confidence intervals along with p-values?
Confidence intervals give an indication of the precision of summary statistics.
58
What should be considered along with statistical significance?
Consider both statistical significance and clinical significance.
59
What is the best approach for reporting statistical results?
Report both the p-values and confidence intervals to provide a complete picture of the data's statistical and clinical significance.
60
What is a confidence interval?
Another method for statistical inference, measuring the precision of our result in the sample in relation to the expected truth in the population.
61
What do confidence intervals account for?
They account for sampling/random error.
62
How do you interpret a confidence interval for a mean age?
If the mean age of women is 28 years with a 95% CI of 25-31 years, we are 95% confident that the true mean age of the population lies between 25 and 31 years.
63
What does the 95% confidence level mean?
It means that 5% of the time, the true population mean will not lie within the sample's 95% confidence interval just by chance.
64
What do wide confidence intervals indicate?
Lack of precision. Example: 4.2 (0.3-16.2).
65
What do narrow confidence intervals indicate?
Good precision. Example: 0.4 (0.2-0.6).
66
How does sample size affect confidence intervals?
Larger samples are more likely to be representative of the population, resulting in more precise, smaller intervals.
67
Why are confidence intervals important?
They provide a range within which we expect the true population parameter to lie, offering insight into the precision and reliability of our estimates.
68
What indicates a non-significant result in terms of confidence intervals?
If the 95% confidence interval around your measure of association includes the null value, the p-value will be >0.05, and the result will not be statistically significant.
69
How is statistical significance determined when testing the difference in mean gestational age?
If the 95% CI includes 0, there is a non-significant difference in the mean gestational age.
70
How is statistical significance determined when measuring prevalence, risk, or odds ratio?
If the 95% CI includes 1.0, there is a non-significant association between the exposure and the outcome.
71
What is the relationship between confidence intervals and p-values?
A 95% confidence interval that includes the null value corresponds to a p-value >0.05, indicating non-significance.
72
Why are both confidence intervals and p-values important?
Confidence intervals provide a range of plausible values for the population parameter and indicate the precision of the estimate, while p-values indicate the probability of observing the data if the null hypothesis is true
73
What should you consider when choosing a test to test your null hypothesis?
Consider the type of the dependent variable, whether the grouping variable is independent or paired, and the distribution of the data if the dependent variable is numerical.
74
What types of dependent variables might you have?
Numerical or categorical.
75
How do you determine if the grouping variable is independent or paired?
Independent groups are different and unrelated (e.g., men vs. women). Paired groups are related (e.g., before and after measurements in the same subjects).
76
What should you check if the dependent variable is numerical?
Check the distribution of the data (normal or skewed).
77
What do parametric tests assume?
They make assumptions about the distribution of the data (e.g., normal distribution).
78
What do non-parametric tests assume?
They do not make assumptions about the distribution of the data.
79
When should you use parametric tests?
When the data distribution meets the assumptions of the test (e.g., normal distribution for numerical data).
80
When should you use non-parametric tests?
When the data distribution does not meet the assumptions of parametric tests or is unknown.
81
What test is used to determine if a numerical variable is normally distributed?
Shapiro-Wilk test.
82
What does the Shapiro-Wilk test do?
It tests whether the distribution of the variable is different from a normal distribution.
83
What test is used for comparing a numerical variable across two independent groups if the distribution is normal?
T-test or Student's T-test (parametric test).
84
What does the T-test compare?
It tests for the difference in means.
85
Why does the T-test not consider the rest of the distribution?
Because the characteristics of a normal distribution are fixed.
86
What test is used for comparing a numerical variable across two independent groups if the distribution is skewed?
Mann-Whitney test or Wilcoxon rank-sum test (non-parametric test).
87
What does the Mann-Whitney or Wilcoxon rank-sum test compare?
It tests for differences in the overall distribution of the variable.
88
What test is used for comparing a numerical variable across more than two independent groups if the distribution is normal?
One-way ANOVA (parametric test).
89
What does the One-way ANOVA test for?
It tests for the difference in means.
90
What limitation does the One-way ANOVA have?
It tells you whether there are any differences, not which specific groups are different.
91
What is needed if the One-way ANOVA finds a difference?
Additional tests (post-hoc tests) to see which groups are different.
92
What test is used for comparing a numerical variable across more than two independent groups if the distribution is skewed?
Kruskal-Wallis test (non-parametric test).
93
What does the Kruskal-Wallis test for?
It tests for differences in the overall distribution of the variable.
94
What limitation does the Kruskal-Wallis test have?
It tells you whether there are any differences, not which specific groups are different.
95
What is needed if the Kruskal-Wallis test finds a difference?
Additional tests (post-hoc tests) to see which groups are different.
96
What are paired groups?
Groups that are related to each other, e.g., the same people measured at two different points in time.
97
What test is used for comparing a numerical variable across two paired groups if the distribution is normal?
Paired T-test (parametric test).
98
What does the Paired T-test compare?
It tests for the difference in means.
99
What test is used for comparing a numerical variable across two paired groups if the distribution is skewed?
Wilcoxon signed-rank test (non-parametric test).
100
What does the Wilcoxon signed-rank test compare?
It tests for differences in the overall distribution of the variable.
101
What does the Chi-squared test compare?
It tests whether the observed frequencies are different from the frequencies that would be expected if there were no differences between the groups.
102
When should you use Fisher's exact test instead of the Chi-squared test?
When the expected frequencies are less than 5.
103
How can you visually assess the relationship between two numerical variables?
Use a two-way scatter plot.
104
What test is used for linear relationships if the variables are normally distributed?
Pearson’s correlation (parametric)
105
What is the range of Pearson’s correlation coefficient (r)?
It ranges from -1 through 0 to +1.
106
What test is used if the variables are not normally distributed?
Spearman’s correlation coefficient (non-parametric).
107
What is the null hypothesis for testing the relationship between two numerical variables?
There is no relationship between the two variables in the population
108
What is the alternative hypothesis for testing the relationship between two numerical variables?
There is a relationship between the two variables in the population.
