Multiple Regression: Get How to Use It in Analyzing Your Data
Multiple regression is a very useful statistical tool in analysis using data gathered by scientists especially those whose research involve many variables or factors (multivariate). Data analysis using multiple regression requires a simple understanding of the nature of the variables or factors studied to get how to relate these variables with each other.
How will one get how to use multiple regression and how will data be used in the analysis? This requires an understanding of the basic statistical concepts. In studies that involve multiple variables, at least two things should be clear in mind. That is, the researcher must be clear about which variables are dependent or independent.
Dependent and Independent Variables
In studying statistics as a tool useful in understanding data, some students of statistics cannot get how to understand two basic concepts: dependent and independent variables. Analysis using statistical tools will be worthless unless these terms are well understood.
But first, the term "variable" must be defined. A variable is a phenomena in nature that is subject to variation, or a quantity that can assume any set of values. Age, for example, is a variable. It can assume a value of 0 to the maximum life span man can reach, say 120. Other variables are gender (male and female), civil status (single, married), race (Caucasian, African, Asian), eye color (blue, green, brown, etc.) or height (assume a range of shortest to tallest).
Which of these factors will be the dependent or the independent variable in trying to find out relationships between variables? This will now depend on the research question. The dependent variable is the variable that changes when the independent variable is manipulated. In other words, the dependent variable is the "response" or "outcome" of the independent variables which are the "stimuli" or "manipulated" variables.
Use of Dependent and Independent Variables in Multiple Regression Analysis
This can be made clear by a sample research study which aims to find out which factors correlate with height. In the above-mentioned variables, we may include age, gender, and race as the independent variables. The research question is "Is there a significant relationship between height (dependent variable) with age, gender, and race (independent variables)?
This is the type of question that multiple regression can provide answers to. The computer output of multiple regression analysis will reveal which of the independent variables highly correlate with height. The researchers get how age, gender and race relate with height. But the exact relationship can be provided by multiple regression results.
In the analysis, using computers is the only fast and feasible way to get how to determine the degree of relationships are among multiple variables. Multiple regression analysis is provided as an add-in in Microsoft Excel. There are also many statistical software companies that always include multiple regression analysis as part of their programs. Data analysis using softwares like SPSS, Statistica, SAS, among others have their own unique ways of encoding data.
How Will Data in Categories Be Encoded?
How will data in categories be encoded for multiple regression analysis? Each statistical software has its own way prescribing how data will be encoded properly in the columns and rows. They provide a help section which will show how data will be organized for analysis. Of course, the data should be in numbers, not words. These are called dummy variables. To get how to encode data, you may refer to this resource.