proportions Flashcards

1
Q

What are the four conditions of a Binomial random variable?

A

Fixed number of trials (๐‘›)

Independent trials

Two possible outcomes per trial

Same probability of success for each trial

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

How is the sample proportion (๐‘ฬ‚) estimated?

A

๐‘ฬ‚ = (Number of successes) / (Total number of trials)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What is the formula for the standard error of the sample proportion?

A

๐‘ ๐‘’(๐‘ฬ‚) = sqrt[๐‘ฬ‚(1โˆ’๐‘ฬ‚)/๐‘›]

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What is the formula for a 95% confidence interval for ๐‘?

A

๐‘ฬ‚ ยฑ (1.96 ร— ๐‘ ๐‘’(๐‘ฬ‚))

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Why do we assume a Normal approximation for the sample proportion?

A

Because the sample sizes are large, allowing the Binomial distribution to be approximated by a Normal distribution.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What R function is used to compute a confidence interval for proportions?

A

binconf(x, n, alpha=0.05, method=โ€asymptoticโ€) from the Hmisc package.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What happens when the sample size is too small for a Normal approximation in confidence intervals?

A

The confidence interval may extend below zero, which is not possible for probabilities. A different method, such as Wilsonโ€™s interval, should be used.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What is Wilsonโ€™s confidence interval formula used for?

A

It is used for small sample sizes to ensure the confidence interval does not go below zero or above one.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What is the formula for the standard error of the difference in two proportions?

A

se(p^1โˆ’ p^2)= sqrt[((p^1(1 - p^1) / n1) + (p^2 (1 - p^2) / n2)]

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What are the null and alternative hypotheses for comparing two proportions?

A

Null Hypothesis (๐ป0): No difference between proportions (๐‘๐ด - ๐‘๐ถ = 0)

Alternative Hypothesis (๐ป1): There is a difference (๐‘๐ด - ๐‘๐ถ โ‰  0)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Why is a hypothesis test used to compare two proportions?

A

It determines whether the observed difference between two proportions is due to chance or represents a true difference.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Why is Wilsonโ€™s interval preferred for small sample sizes?

A

It avoids impossible probability values (e.g., negative probabilities) and provides more accurate confidence intervals when ๐‘ฬ‚ is close to 0 or 1.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What does the test statistic measure in a two-proportion z-test?

A

It measures the observed difference between sample proportions as a ratio of the standard error, helping determine statistical significance

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

How is the test statistic for comparing two proportions calculated?

A

dataestimateโˆ’hypothesizedvalue / standard error

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What does a test statistic of 4.82 indicate in a z-test?

A

It means the observed difference is 4.82 standard errors away from the null hypothesis (zero difference), suggesting strong evidence against ๐ป0

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What is the p-value for a test statistic of 4.82 in a z-test?

A

The probability of obtaining such an extreme value (or more) under ๐ป0 is very small, around10โˆ’6, leading to rejection of ๐ป0

17
Q

How is the confidence interval for the difference in two proportions calculated?

A

(p^1โˆ’p^2)ยฑ(zร—se(p^1โˆ’p^2))

18
Q

What are the three sampling situations for comparing proportions?

A

Situation A: Independent samples (e.g., comparing two countries).

Situation B: One sample, mutually exclusive categories (e.g., voting choices).

Situation C: One sample, multiple response options (e.g., survey with multiple answers).

19
Q

How is the standard error calculated for Situation A (independent samples)?

A

sqrt [((P^1(1-P^1))/ n1)+ ((p^2(1-p^2) / n2)]

20
Q

When comparing survey responses from two countries, which sampling situation applies?

A

Situation A (independent samples), since each person belongs to only one countryโ€™s sample.

21
Q

How does the choice of standard error formula impact results?

A

If the wrong formula is used, confidence intervals and hypothesis tests may be incorrect, leading to misleading conclusions.

22
Q

What is the formula for the standard error of the difference between two proportions?

A

se(p^1โˆ’ p^2)= sqrt [(P^1 + P^2 - ( P^1-P^2)^2) /n]

23
Q

When should Situation B be used in sampling?

A

Situation B is used when one sample is asked a single question with mutually exclusive response options, such as โ€œagree,โ€ โ€œdisagree,โ€ or โ€œdonโ€™t know.โ€

24
Q

How are statistical odds calculated?

A

Odds= p(success) / p(failure) = p / 1โˆ’p

25
Q

What does an odds ratio (OR) greater than 1 indicate?

A

It indicates that the intervention group has higher odds of success compared to the control group.

26
Q

How is an odds ratio (OR) calculated?

A

ฮธ= oddsingroup1/ oddsingroup2โ€‹

26
Q

What does an odds ratio (OR) less than 1 indicate?

A

It indicates that the control group has higher odds of success compared to the intervention group.

27
Q

Why do we use the log of the odds ratio to construct confidence intervals?

A

Because the distribution of the odds ratio is highly skewed, and taking the log makes it approximately normal.

28
Q

What is the formula for the standard error of the log odds ratio?

A

seOR= sqrt [ 1/n11 + 1/n12 + 1/n21 + 1/n22 ]

29
Q

How do you obtain a confidence interval for an odds ratio?

A

Compute log(ฮธฬ‚)
use: log(ฮธ^)ยฑz1โˆ’ฮฑ/2ร—seOR
โ€‹Exponentiate the lower and upper limits to return to the odds ratio scale.

30
Q

What does an odds ratio of exactly 1 indicate?

A

It indicates no difference between the two groups.

31
Q

What is the formula for the pooled sample proportion when testing the difference between two proportions?

A

p^ = (x1 + x2) / (n1 + n2)

where ๐‘ฅ1 and ๐‘ฅ2 are the number of successes in each sample.

32
Q

How do you interpret a confidence interval for the difference between two proportions?

A

If 0 is in the interval, there is no significant difference.

If the interval is entirely positive, ๐‘1>๐‘2
โ€‹
If the interval is entirely negative, ๐‘1<๐‘2

33
Q

Why are odds used instead of probabilities in logistic regression?

A

Because odds have mathematical properties that allow for a linear relationship with predictor variables on the log scale.

34
Q

What does a log odds ratio of 0 mean?

A

It means the odds ratio is 1, indicating no difference between the two groups.

35
Q

What transformation is used to make the odds ratioโ€™s distribution approximately normal?

A

The natural logarithm (log transformation) is applied to the odds ratio.