In: Statistics and Probability
On a survey taken in 2018 and reported in the GSS data file, the question, "How much of your communication is via text, mobile phone, or internet?", was asked.
The summary results are as follows: Among 202 people who are 18-29 years old, 69 reported, "All or almost all of it." Among 166 people who are 40-49 years old, 50 reported, "All or almost all of it."
a. Find the 97% confidence interval for the difference between the proportions of those two age groups that answerd, "All or almost all of it." Include the name of the routine and the margin of error in your answer.
b. Does this data indicate that the population proportion of this issue is different between the two groups? Explain.
We want to know about the population proportion difference. That means these are binomial proportion differences. Since n > 30, we can use normal approximation.
Here the success is to answer yes to almost or all of it.
18-29 (1) | 40-49 (2) | |
x | 69 | 50 |
n | 202 | 166 |
Sample | 0.3416 | 0.3012 |
pooled | 0.3234 |
a. Find the 97% confidence interval for the difference between the proportions of those two age groups that answerd, "All or almost all of it." Include the name of the routine and the margin of error in your answer.
=1 - 0.97 = 0.03
Therefore the critical value at
=
= 2.17009.............using normal percentage tables with p = 0.015
Margin of error = Critical value * SE
= *
= 2.17009 * 0.0024
Subsituting the values
This means that we are 97% confident that true proportion difference lies within this range.
b. Does this data indicate that the population proportion of this issue is different between the two groups? Explain.
We can answer this using confidence interval testing for difference.
Test:
Null: There is no difference
Alternative: There is a difference
Test Statistic:
(0.038, 0.0428)
Since the interval does not include '0', we can say that there is a difference between the proportions at 3% significance.