PSY 307 – Statistics for the Behavioral Sciences Chapter 1What is Statistics  A tool for discovering relationships and patterns in data.  A way of describing and comparing behavior of groups.  A way of presenting and supporting arguments empirically.Two Kinds of Statistics  Descriptive statistics – tools for organizing and summarizing observations.  Tables, graphs, averages  Inferential statistics – tools for generalizing beyond the actual observations.  Hypothesis tests (t-test, ANOVA)Two Kinds of Inferential Studies Population (unknown scores) Sample (known scores) Random Sample Survey – SOC/PSY 444 Volunteers (unknown scores) Treatment Group (known scores) Control Group (known scores) Compare Experiment – PSY 433 Random AssignmentTwo Types of Data  The type of statistic used depends on whether data are numbers or words (or codes).  Quantitative – any single observation is an amount or a count.  Ranks – relative standing, order  Qualitative – any single observation is a class or category.Numerical Codes  When data consists of classes or categories, sometimes numbers can be used to replace category names.  Male = 1, Female = 2  Yes = 1, No = 2What is a Variable  Variable – a characteristic or property that can take on different values.  Constant – an observation that has only one value.  Any single observation in a data set.Types of Variables  Discrete – isolated numbers separated by gaps.  Most counts  Continuous – values have no restrictions.  Weight, reaction times  Rounded off continuous variables are approximate.  Rounding creates gaps.Types of Variables Nominal Ordinal Scale Categories Ranks Interval/Ratio Discrete Discrete Continuous Words or codes Numbers Numbers Male/female Old/new Yes/No 1st, 2nd, 3rd Fresh, Soph Prvt, Cpl, Sergeant Age Temperature Response time Correct answers Words can be coded using numbers Arrange in order then replace rank with number Continuous values become discrete when rounded off ExamplesVariables in Research  Independent variable – the variable manipulated by the investigator.  May be used to define groups.  Used to test hypotheses about causation.  Dependent variable – the variable measured, counted, or recorded by the investigator.  The outcome.  Confound – an uncontrolled variable that varies with the independent variable.Two Kinds of Studies Number of Communication Breakdowns Active-Listening Training No Active-Listening Training Number of Communication Breakdowns Independent Variable Dependent Variable Pre-existing Score for Active Listening Number of Communication Breakdowns First Variable Second Variable Are the two variables related? Is this a real or transitory difference? Experiment Observational StudyMultiple Dependent Variables  Sometimes variables of interest cannot be manipulated (e.g., sex, poverty) but only measured.  Correlation studies -- studies with multiple dependent variables.  Goal is to identify relationships among the dependent variables measured.  Often used in observational