70 likes | 107 Views
Inferential Statistics. Inferential statistics. Use the sample to make inferences about the population Use at least two variables at a time Can never be sure that results in sample are 100% representative of population
E N D
Inferential statistics • Use the sample to make inferences about the population • Use at least two variables at a time • Can never be sure that results in sample are 100% representative of population • Instead, statistics tell us the probability that results in sample are true in population
Significance Levels • p<.10 • This means that the probability that the finding was by chance is less than 1 in 10. • This indicates that there was a 'trend' in the data. It is not considered a significant finding, but it is encouraging enough to continue with the line of investigation. • p<.05 • This means that the probability that the finding was by chance is less than 5 in 100. • This indicates that there was a significant finding in the data. You should certainly continue with this line of investigation. • p<.01 or less • This means that the probability that the finding was by chance is less than 1 in 100. • This means that there was a highly significant finding in the data. You should continue with this research with considerable confidence.
Inferential Statistical Tests: Which Test? * For Correlation and Chi-Square analyses, variables are not distinguished as IVs or DVs.
High GPA Low Self-Esteem Does GPA Correlate With Self-Esteem? • Does GPA tend to increase in people with high self-esteem? • Correlation coefficient indicates the degree of association between the two variables.
Perfect Moderate Weak + - Two Features of a Correlation Direction Strength
Chi-Square Expected Results for Random Relationship between Gender and Ability to Drive What if this is what we really found: