Research Methods for Managerial Decisions and Survey Instrument

Research Methods for Managerial Decisions and Survey Instrument

By

Mollee Kikumoto

Kelly LaFrance

Jennie McClure

Alvin Trotman

Instructor: Jason M. Etchegaray, Ph.D.

Abstract

After running the "Research Methods for Managerial Decisions" simulation Team B will further explore the multiple regression model and how it relates to Coffee Time predicting weekly revenue more accurately using normal values and lagged values. The difference between the two models will also be explained. This paper will also look at Coffee Time's adverting spending to travel agents, as well as provide the key decision maker in the simulation with recommendations to the challenge Coffee Time has been given. Lastly, Team B will compile a survey that will allow the team to address a problem that requires a decision to be made based on several factors including demographics, operationalizing the problem, the hypotheses, scale development, levels of measurement, instrument design, and face validity.

Introduction:

All Team B members ran the Research Methods for Managerial Decisions and Survey Instrument

Simulation multiple times in order to gain complete data. The multiple regression model is explained

using both normal and lagged values. Tourism was a huge consideration for Coffee Time and Team B

explains why this is an unwise decision in the upcoming paragraphs. Coffee Time comparatively prices

their sales based on the leading competition and markets accordingly. Team B makes suggestions

based on these actions and encourages Coffee Time to spend more than the leading company in all

areas of promotion. A survey instrument has been implemented by Team B to collect data to address

the business decision of : "should Coffee Time provide sandwiches along with coffee."

a. Laura wanted to build a multiple regression model based on advertising expenditures and coffee times price index. Based on the selection of all normal values she obtained the following:

1) Multiple R = 0.738

2) R-square = 0.546

By using lagged values she came up with the following:

1) Multiple R = 0.755

2) R-square = 0.570

Explain the differences in using these different models. How could CoffeeTime further optimize this model?

The Coffee Time simulation allowed Laura to build and analyze different models in order to predict Coffee Time's weekly revenue more accurately. Laura first decided to select all normal values. By selecting the three normal values she was able to determine that 54.6% of the variation in predicted weekly revenue is explained by Coffee Time's weekly advertising expenditure, Coffee Time's price index, and the estimate on Quick Brew's weekly advertising expenditure. Using the normal model it allows Laura to have two independent variables used, which are known variables, but also takes into consideration a dependent variable which is Quick Brews weekly adverting expense. Without having direct access to Quick Brew's weekly advertising expense statement ahead of time, like say an employee would have, Coffee Time has decided to enlist a media monitoring firm to give them an estimate of Quick Brew's advertising. The normal model doesn't allow Coffee Time to have exact numbers ahead of time for Quick Brew's advertising expenditure so they must wait until after the fact to view it.

When Laura uses lagged values, which are values that are known, and are considered all to be independent variables, Laura would be expected to see that a higher percentage of coffee times revenues could be explained by Coffee Time's advertising expenditures. Using the lagged model allows Laura to have all independent variables and view past data. Knowing Coffee Time's weekly advertising expenditure, Coffee Time's price index, and the estimate on Quick Brew's weekly advertising expenditure by viewing past data Laura was able to determine that 57% of the variation in predicted weekly revenue is based on these three independent variables. This percentage is higher then what was determined using the normal model.

Coffee Time could further optimize this model by using a combination of normal values and lagged values. For example, if Coffee Time used their weekly advertising expenditure and their price index as both normal values and lagged values, and then used the estimate on Quick Brew's weekly adverting expenditure as a lagged value, they would then have a 75.6% variation, which is much greater then using all normal values or all lagged values. The perfect scenario would be to factor in all independent variables and isolate the one dependent

variable. That would allow Laura to come up with the highest percentage of variation in predicted weekly revenue as it correlates to adverting expenditures and price index.

b. Tourism is one consideration for Coffee Time's future. A survey of 1,233 visitors to Mumbai last year revealed that 110 visited a small cafe during their visit. Laura claims that 10% of tourists will include a visit to a cafe. Use a 0.05 significance level to test her claim. Would it be wise for her to use that claim in trying to convince management to increase their advertising spending to travel agents? Explain.

Hypothesis: 10% of the tourists will include a visit to a cafй.

Significance Level: 0.05

HO: pi > 10%

H1: pi < 10%

Given the information of the example, it was determined that a one-test