REFLECTING ON MY CLASS LEARNING

It was an engaging week on learning real-life simulation methods in solving business problems and decisions on how issues can be tackled.

In regression analysis, we need to understand how certain things are done and the subsequent effects or impacts on the decisions we make. Having to learn on linear regression was insightful. It is used in predicting the values of both dependent and independent variables.

In making decisions as managers, we need to take certain steps. The relationship between price and cost are part of the consideration to tend towards.

There are two major drivers for revenue- price and volume, we have better control over price than volume.

There is the sensitivity of price and volume.

We talk about the correlation analysis. It is the relationship between two variables. It tells us the direction and strength of the relationship or associations between two variables.

The basic concepts of correlation are:

1. Direction of the relationship
2. Strength of the relationship.

The direction of the correlation relationship can be strong positive or strong negative relationship. However, with the strength of the relationship, it can be strong or weak.

It is important to distinguish between direction and strength. The need for direction as the most important thing we need to look out for.

We look at the implication of increasing price and possible effect.

The correlation relationship is divided into two:

• Graphical Method – Scatter Diagram
• Algebraic Method – Karl Pearson’s Coefficient.

The variance can be quantitative in nature, example of quantitative is number of cars etc. Whereas for product moment correlation. Examples of Pearson product correlation is software.

The spearman rank correlation is used for a qualitative correlation. Example of qualitative correlation is colours, perception etc.

Limitations with correlation analysis are:

• Correlation does not imply causation.
• Correlation does not separate or differentiate between dependent and independent variable.
• A high correlation can be by mere chance as there may be no logical connection.
• The correlation cannot be used to forecast or predict.

Regression analysis shows the followings:

• Relationship between two or more variables.
• Shows the cause and effect
• Can be used to forecast and predict.

Regression means going backward. That is studying the past behaviour. It is also about pattern recognition. It is pattern behaviour of the past.

As previously discussed, correlation does not separate between dependent and independent variables.

In regression, a model must be correctly specified (Y = a +bx). Y is called independent variables and X is dependent. A dependent variable is affected by the change of another variable while the independent variable is not dependent on any variable. Example of dependent variable is productivity, and the dependent is labour. Productivity is measure by output while labour is measured by man-hour.

There are so many assumptions in a regression analysis.

A simple regression model is where one has a single dependent and independent variable.

While a multiple regression model one has a single dependent and multiple independent variables.

#MMBA-4