In defining a hypothesis for the study, the researcher of the study took into consideration the variables to form a predication in the mind of the affected changes likely to occur if the independent variables are varied across the scale. The multiple regression is used to find the future value amount of the sales when provided with data concerning the adverts, radio plans, and airplay image budgets. The three variables mentioned above affect the direction of sales for the firm. The reason is that they have a direct relationship, which can be equated to linearity. By combining the diverse linear regression models in the case study resulted in a multiple regression model, which was used in the determination of the budgeted sales in the firm. The following is a simple illustration of linear regression analysis:
Y=Mx+C.
Y represents the sales in values
M is the constant which is applied in the determination of the future values for the sales and constant amounts represented by C; thus, fixed, variable costs. The above regression utilizes the idea of two known variables. There is the case where the independent variables are more than two, such as the case presented. The following is an example of the regression model applicable to knowing whether the hypothesis can be rejected or accepted. In most cases, the hypothesis is stated as a null hypothesis. If one accepts or rejects the null hypothesis, then the multiple regression acts as a predictive tool for knowing variable quantities once the due process has taken its course. The general regression for the case is as follows:
Y = C + X1+ X2 + X3
Where;
Y = sales
X1 = adverts
X2 = radio plans
X3 = airplay/ brand images
C- Constant
The hypothesis utilized in the study is the null hypothesis, and they are stated as below:
Null hypothesis
H01: The various adverts placed in the print and social media do not have a major effect on the number of sales realized by the enterprise.
H01: The radio plans instituted by the firm did not have a significant impact on the sales revenue generated for the year.
H01: The branding of the different images of company products and services had a little or no impact on the number of sales for that period.
Outlining the hypothesis
Apparently, the figures obtained for the t-test indicated a p-value of 0.00, which prompted the rejection of the null hypothesis. The value is acceptable at a significant level of 0.05. This means that radio plans, branding, and adverts have a primary impact on company sales for a specific time. The reason is that the three marketing techniques tend to provide the clients with broader network coverage of educating them of the different services and commodities manufactured by the organization. The direct relationship established by the variables
creates an element of linearity and collinearity, which forms one of the assumptions of multiple regression models to hold. The three hypotheses support the fact that multiple regression was the best appropriate method of predicting future trends of the market to inform management of whether profit will increase or not to the following years or months.
Independent and dependent variable measurements
Sales ΒDependent variable
Air play Images, advert and radio plans forms part of the independent variables.
Sales is not a constant figure and is measured by the amount of profit the firms make for the financial years. The sales can be in terms of quantifiable volumes or measured in ratios such as the return of the various items of the balance sheet or at the rate of which the stock is moving. Any changes in the form of the advert to place to the public has the significant effect of bringing massive changes into the organization sales. The reason is that these independent figures are not affected by other factors in the market as they privately institute forces on the sales variable.
Assessment of outliers
In the determination of the best line of fit, some variables may become outliers because they do not fit on the straight-line fit. For the purpose of using linear regression, coordinates were plotted on the Cartesian plane to get the liner line for the research. The outliers are mostly few in a scenario where the IV and DV are plotted on the graph. Most of the variables plotted, except for the incident of a number of plays on, forms a pattern that is not clearly reflected in the value output result. The assessment of outliers on the study involves two points which were probed further for clarification purposes.
Assumptions of multiple regression
The first postulate of the regression is that all the extra points will have a regular distribution pattern. This means that a linear expression model or line will be formed once the points are plotted onto the graph. Secondly, in the case of multiple regression, the issue of multicollinearity should not exist because the results would not be accurate in that case. The IV variables should not have a high correlation with one another as expressed on the table results of the case study. Thirdly assumption is that the error term of all the IV is the same across all the levels. In testing for error term similarity, a graph could be used where the residuals which had a specific standard are plotted against the discrete values (Field, 2017).
Tests for addressing the assumption
In most times, the multicollinearity, the correlation matrix is used in testing whether the IV is more or less correlated. The following is an example of the test explained above:
Variables Sales Adverts Airplay images Branding images