9.35 Nikkei 225. If we know how the New York stock market

performs today, can we use that information to predict whether the

stock market in Japan will go up or down tomorrow? Or is the

movement of the Japanese stock market today better predicted by

how the Japanese market did yesterday? The file Markets contains

data from two stock markets for 56 days. The variables recorded are

Date; Nik225ch (the one-day change in the Nikkei 225, a stock

index in Japan); DJIAch (the one-day change in the New York-

based Dow Jones Industrial Average from the previous day); Up (1

or 0 depending on whether or not the Nikkei 225 went up on a

date); and lagNik (the one-day change in the Nikkei 225 from the

previous day). Thus if we want to predict whether the stock market

in Japan will go up or down on a Tuesday, we might use the

Monday result from Japan ( lagNik) or the Monday result from

New York ( DJIAch)? remembering that when it is Monday

evening in New York, it is Tuesday morning in Japan.

a. Fit a logistic model with Up as the response and DJIAch as

the predictor. Is DJIAch a significant predictor of the

direction the Nikkei 225 will go the next day? Explain the

basis for your conclusion.

b. Fit a logistic model with Up as the response and lagNik as the predictor. Is lagNik a significant predictor of the

direction the Nikkei 225 will go the next day? Explain the

basis for your conclusion.

c. Compare the models in part (a) and part (b). Which

variable, DJIAch or lagNik, is a more effective predictor of

where the Nikkei 225 is going the next day? Explain the

basis for your decision.

An analysis of the Business School graduates found that 211 out of 310 randomly selected graduates used a statistical inference technique during their first year of emplyment. (a) Calculate a 90% confidence interval for the proportion of graduates who used a statistical inference technique within the first year of their employment. Present your work. (b) Interpret the confidence interval. H (c) If a new follow-up study is to be undertaken, what sample size should be taken in order to estimate the true proportion of graduates who used a statistical inference technique in the first year of employment within 2%, with a confidence interval of 90%? (d) If a new follow-up study is to be undertaken, what sample size should be taken in order to estimate the true proportion of graduates who used a statistical inference technique in the first year of employment within 2%, with a confidence interval of 99%?An analysis of the Business School graduates found that 155 out of 302 randomly selected graduates used a statistical inference technique during their first year of employment. (a) Calculate a 90% confidence interval for the proportion of graduates who used a statistical inference technique within the first year of their employment. Present your work. (b) Interpret the confidence interval. (c) If a new follow-up study is to be undertaken, what sample size should be taken in order to estimate the true proportion of graduates who used a statistical inference technique in the first year of employment within 3% with a confidence interval of 90%? (d) If a new follow-up study is to be undertaken, what sample size should be taken in order to estimate the true proportion of graduates who used a statistical inference technique in the first year of employment within 3% with a confidence interval of 95%?Select True or False from each pull-down menu, depending on whether the corresponding statement is true or false. 1. In a sample of 500 students at a university, 12% of them are accounting majors. The 12% is an example of statistical inference. 2. A summary measure that is computed from a sample to describe a characteristic of a population is called a statistic. 3. Conclusions and estimates about a population based on sample data are not always going to be correct. For this reason measures of reliability, such a significance level and confidence level, should be built into the statistical inference. 4. A local cable system using a sample of 500 subscribers estimates that forty percent of its subscribers watch a premium channel at least once per day. This is an example of statistical inference as opposed to descriptive statistics