PSYC217 Chapter 13 - Inferential Statistics Introduction - Researchers are interested in inferring whether results obtained represent what we would find in the entire population from which that sample was drawn. - Inferential statistics: statistics that estimate whether the results observed based on sample data are generalizable to the population from which that sample was drawn. - Results of any given study are based on a sample of participants drawn from larger populations. · Researchers rarely study entire populations > simply too large, making it impossible to gather data from every single member of that population. - Inferential statistics are a way to help us infer whether a specific result observed in a sample reflects what we would observe in the population. - NHST -> most common form of inferential statistics (null-hypothesis significance testing) · Identified as any technique that results in the calculation of a p-value (probability value). Inferential Statistics: Ruling out Chance - Groups are made equivalent by controlling all other variables · Use techniques like random assignment or within-subjects design. - If groups are equivalent except for manipulated > any differences in DV are assumed to be caused by manipulated IV. " To be valid experiment must be well-designed and free of threats to internal validity. · Random error will be responsible for some difference between groups, even if IV had no effect on DV. - Inferential statistics are a way to judge whether the difference between means reflects a real effect of the IV observed within the population, or simply random error. - When inferential statistics conclude differences > statistically significant · Statistical significance does not indicate importance of an effect, or how meaningful it is. Statistical Significance - Statistically significant: within NHST framework, observing an outcome has a low probability of occurrence (defined as p-value less than .05), assuming null hypothesis is correct. - After collecting data and calculating descriptive statistics must determine if values are statistically significant. . "Are differences due to random error, or do they reflect a real effect of the experimental manipulation that would be observable in the population?" · Researchers analyze data using statistical tests -> different kinds of statistical tests depends primarily on the study's design and type of data · Logic underlying the use of any statistical test rests on statistical theory, which is grounded in probability theory. - Goal of any statistical test > inform a judgment about whether an effect observed in a sample is good evidence of a real effect in the population. " Decide how willing you are to be wrong if there is an effect in the population > significance level/alpha level (also known as threshold for statistical significance), signified by alpha (a). · For NHST statistics, we calculate a probability value (p-value) and then check whether this value is larger or smaller than alpha (significance level). · Smaller than alpha > result is statistically significant. - Most likely to obtain statistically significant results (i.e., p < . 05) when you have large