KEY FEATURES: Introduces students to developing research questions and shows their importance in driving research design. Rarely taught topics, such as how to enter and clean data, offer students information missed in both research methods and statistics courses. Shows how to write up survey results for academic, business and nonprofit reports to alleviate the confusion students feel about how to write up findings. Rigorous treatment of sampling focuses on many sampling issues from probability theory to weighting. Offers the process of actually conducting a survey with advice on administering surveys, incentives, and improving response rates.

How Do We Interpret Linear Regression With More Than One Predictor Variable?

How Do We Interpret Linear Regression With More Than One Predictor Variable?

A great strength of regression is that it allows us to include multiple independent variables in the regression equation. We can regress income on both education and gender simultaneously. The equation is Income = a + b1*education + b2*male. We would want to do this because gender is related to both income and education. Therefore, the parameter estimate for the regression of income on education alone is too big; it is confounded with gender. Figure 91.1 shows what this means. If we regress income on education only, the parameter estimate will be too large because it will include areas “a” and “b”. Area “b” indicates an area of overlap between gender, ...