Question

Describe a research effort where you could use a Multiple Regression analysis. It could be something related to work productivity, or perhaps a student’s performance in school.

List three variables (X1, X2, X3) you’d include in a Multiple Regression Model in order to better predict an outcome (Y) variable. For example, you might list three variables that could be related to how long a person will live (Y). Or you might list three variables that contribute to a successful restaurant. Your Regression Model should have three variables that will act as “predictors” (X1, X2, X3) of a “criterion” (Y’). Note that the outcome or criterion variable (e.g. how long a person would live, or the success/profit made by a restaurant measured) in must be a “Measurement” variable, that is something that is measured on a scale like inches, pounds, IQ, lifespan, stock value, etc. But that the predictors (X variables) can be either a measurement variable OR a categorical variable such as gender, political party, location, etc.

Answer 1

Suppose in a class , training on effective utilisation of time in competitive examinations is given. The training is conducted by experts in the field who have wide exposure and vast experience. The training consists of solving previous question papers and other practice problems. The training is conducted for 40 hours. To judge the effectiveness of training , examinations are conducted before and after training. The number of students who took the coaching are 50. The score of students are recorded for three tests before and after training. To estimate the effectiveness of coaching the analyst has developed a simple model where marks of students are taken as dependent variable. The independent variables are 1) dummy representing whether marks correspond to before or after coaching. 0 represents before coaching and 1 represents after coaching.

2) IQ level of students is classified in to three classes - below average, average and above average.

3) effort put by students is guaged by number of hours studied.

Expected signs of coefficient of the three variables are positive.