Here it is…. …Your Super Duper Faster than Ever Pooper Scooper Review of Statistics and Data Analysis!
Types of Data Nominal – identifies categories e.g. gender, color, yes/no Ordinal – identifies rank order e.g. class rank Interval – falls in a number line with an absolute zero (which means “the absence of”) e.g. height and weight Ratio – falls in a number line, but 0 is just another number e.g. temperature
Figure 1.10 A skewed distribution Myers: Psychology, Eighth Edition Copyright © 2007 by Worth Publishers Descriptive Statistics Measures of Central Tendency mean, median, mode (Which is easily skewed?) If the 3 are =, you have bell curve. If not, a skew. Positive skew = scores pull mean higher Negative skew = scores pull mean lower Question: Which skew is this an example of?
Measures of Variation Range (crude estimate) Standard Deviation – the higher the number, the greater variability of scores Table 1.4 Source: Myers, Psychology, Eighth Edition, Copyright © 2007 by Worth Publishers
Inferential Statistics In order to generalize from samples…: Use representative samples Use larger sample sizes Observations with low variability are more reliable Statistical Significance – the likelihood that the difference in samples’ means is due to chance Goal is a significance of.05 or less In other words, it is only 5% likely the difference is due to chance (or 95% likely the difference may be attributed to the Independent Variable)
Figure 1.3 Scatterplots, showing patterns of correlation Myers: Psychology, Eighth Edition Copyright © 2007 by Worth Publishers Correlation Correlation Coefficient – a statistical measure of the strength of the relationship between variables e.g (number indicates strength, sign indicates direction of correlation) Ranges from to Positive Correlation – two factors increase or decrease together (“DIRECT” correlation) Negative Correlation – one factor increases as the other decreases (“INVERSE” correlation)
Figure 1.5 Three possible cause-effect relationships Myers: Psychology, Eighth Edition Copyright © 2007 by Worth Publishers Correlation does NOT equal Causation! People with low self-esteem are more likely to report depression. Why?
Experimental Design Between-subjects design: Compares one group of people to another Most common method Efficient method, but individual differences of the people may influence the results Within-subjects design: Compares participants to themselves Pre-test Ind. Variable Post-test Less efficient and more time-consuming, but resistant to the possible effects of individual differences
For More Practice… Use the PsychSim Tutorials on the textbook website! textbook website textbook website Click on “PsychSim Tutorials” and then Click on Chapter 1, “Descriptive Statistics” and/or “Correlation Analysis”