Question

A researcher wants to know if more education will have a later effect on how much they watch television. They conducted a survey of 2000 people which asks about education level-- measured by whether the respondent has a college degree-- and the number of hours of TV viewed in last week.

The researcher provides you with the mean hours of TV viewed for the respondents without a college degree (20.9) and for the respondents with a college degree (18.3).

• What other information, if any, would be required to conclude that the estimates are really different between the two groups?
• Given the data that was collected, the most appropriate statistical test to use to test the hypothesis is probably?

Answer 1

We have to groups, i.e. TV viewed for the respondents without a college degree (20.9) and for the respondents with a college degree

It is clear that only sample mean for both groups is provided.

We need sample standard deviation for each group along the individual sample sizes. Significance level is needed to test the final result for the difference between the two groups.

If we have sample size, sample standard deviation, significance level and sample means for each group, then the most appropriate hypothesis test check for the significant difference between the means is two independent samples t test. Two independent samples t test is most appropriate because the population standard deviations are unknown.