Research Design Inferential Stats & Regression PADMN 6289// WEEK 11

Inferential stats 03 Regression 02 Examining relationships recap 01 This Videocast

Examining Relationships (Recap) 01

Examining Relationships Differences in measures of central tendency (means, medians, modes) Measures of association • Appropriate correlation analysis for level of measurement (i.e. nominal, ordinal, interval); Johnson, pg. 208 • Direction of relationship • Strength of relationship (correlation coefficient = r); Johnson, pg. 206 Crosstabs • Controlling for a third variable in crosstabs

Detecting Relationships Analysis types Example criteria Differences between means or medians ✓ Found differences ✓ Controlling for rival explanations (confounding variables) Correlations ✓ Direction of relationship ✓ Strength of association (Johnson, pg. 206) Crosstabs ✓ 10% difference (or more) between groups (Johnson, pg. 199) ✓ Controlling for rival explanations (confounding variables) Statistical significance ✓ Alpha level

Regression Analysis 02

Simple Regression Hours studying GPA IV DV

Simple Regression = intercept + + error bx Y

Simple Regression = a + + e bx Y + b = mx Y intercept Regression coefficient Slope

Hours studying GPA Simple Regression intercept Regression coefficient

Simple Regression Simple Regression: Hours Studying & GPA Hours studying 0.03 *** Constant (aka intercept) 0.04 R2 0.05 N 101 Note: ***p<0.01, **p<0.05, *p<0.10 Regression coefficient • Positive • Statistically significant Model fit • 5% of variance explained Intercept Number observations

Hours studying GPA intercept Regression coefficient

Hours studying Classes attended GPA GPA

GPA Hours partying Hours studying GPA

Multiple Regression Hours studying GPA Classes attended Hours partying

Multiple Regression Multiple Regression: Hours Studying & GPA Hours studying 0.03 *** Classes attended 0.01 Hours partying -0.01** Constant (aka intercept) 0.06 R2 0.15 N 99 Note: ***p<0.01, **p<0.05, *p<0.10

Inferential Stats 03

Indicates whether a finding can be generalized to a population, within a specified degree of confidence Statistical Significance The confidence level (chosen by the researcher) at which a finding is to be considered statistically significant 95% confidence level; a=0.05 90% confidence level; a=0.10 Alpha Level “Probability value,” or probability of a specific result, calculated for a specific result and compared with the chosen alpha level. P-Value If a calculated p-value falls below (is smaller than) the alpha level, the result in question is considered a statistically significant finding.

Understanding Associations & Statistical Significance Association Statistical Significance Interpretation Strong Significant Relationship; Generalizable Strong Not significant Relationship; Not generalizable Weak Significant No/weak relationship; Generalizable Weak Not significant No/weak relationship; Not generalizable

Research Design Inferential Stats & Regression PADMN 6289// WEEK 11