POLS 3650 Lecture Notes - Lecture 15: Null Hypothesis, Interval Ratio, Statistical Significance
Document Summary
Much is the same as comparing a sample to the population: Dv nominal or ordinal = compare proportions; dv interval/ratio = compare means. We can express the comparison in a difference or a ratio. Main difference: we are comparing two groups that are both drawn from a sample. Consequences for statistical significance: doubles the possibility of sampling error and this is further increased if the sample size is small. Example: attitudes towards elizabeth may (interval/ ratio) H1: women are likely to have more positive feelings about elizabeth may than men. H2: women are likely to have the same or more negative feelings about elizabeth. This is a directional hypothesis so we will conduct a one tail test. Visualize data with charts and tables, also calculate necessary measures of central tendency and dispersion. Need to know this because these are samples and we need to know if it is a genuine reflection of the population.