CHAPTER SEVEN:Naturalistic MethodsProblem with naturalistic methods is that they do not answer WHY such behaviors occurNaturalistic Research: designed to describe and measure the behaviors of everyday lives as a source of ideas and hypothesesMay be the only way to collect dataResearcher does not interveneEcological validity (high)- extent to which research is conducted in situations similar to everyday life experiences of participantsObservational Research: making observations of behavior and recording these behaviors in an objective mannerUnacknowledged ParticipantChance to get intimate informationResearcher may change the situationPoses ethical questionsAcknowledged ParticipantPotential for REACTIVITY and influence of personal relationshipsEthically appropriateMay need habituation to get more accurate data observationsAcknowledged or Unacknowledged ObserverNo active participationMay be more accurate and objectiveOperationalizations, Data Collection and AnalysisObjectivityReliability and validitySystematic Coding Methods (What to Observe and Record)Behavioral categoriesTime and spacePeopleCase Studies: descriptive methods of one or more individual’s behaviorsPROSIntensive focus on one or a few instancesSource for ideas and hypothesesAfford the study of rare phenomenaPersuasive and motivational valueCONSAnecdotal and controlledBased on reconstructions of the pastLack of generalizabilityCHAPTER EIGHT:Hypothesis Testing and Inferential StatisticsCorrelational and Experimental ResearchInvestigate and observe relationships between variablesPredictExplainHypothesis Testing: determine the likelihood or probability that the data observed reflect a real relationship between variables (H1) rather than a relationship due to chance (H0)Inferential Statistics: use a sample to draw inferences about the true state of affairs based off of these proceduresHypothesis Testing ChartDevelopment of research hypothesisSet alpha (a=.05) as PROBABILITY THRESHOLDCalculate POWER to determine sample size and collect data. (probability that the test will reject null and avoid type 2 error)Collect statistic and p-value PROBABILITY OF DATACompare p value to a.P<.5 REJECT NULLP>.5 FAIL TO REJECT NULL (ACCEPT NULL)Sampling Distributions: way to specify what the observed data would look like if the research hypothesis was not true. Distribution of ALL possible outcomes of raw scoresBinomial Distribution: for events that have two equally likely possibilitiesNull Hypothesis: research hypothesis cannot be tested directly because we cannot prove something to be true since we do not know what the data should look like but we CAN prove something wrong.Test null to indicate what will happen by chance and hope to REJECT it.Probability and Statistical Significance:P-value: likelihood of observed statistic to occur on the basis of the sampling distribution. Calculate the probability of a mean at least as large as the samplesP<a then REJECT the null and deem significance as data deviates from what is expected under sampling distributionP>a then ACCEPT the null and deem insignificantInferential Errors:Type 1: reject the null when in fact it is trueDependent on alpha level. The lower the alpha the less likely this will happenType 2: failing to reject the null when it is false and missing a true relationshipMore common when power is LOWRandom errorsSmall effect sizePower: probability that the researchers will be able to reject the null properlyEffect size: size of relationship between variablesLarger=stronger, smaller= more missableSample Size: as sample size increases so does the power of the testStatistical significance: p=(effect size)(sample size)CHAPTER NINE:Correlational Research DesignsCorrelational Research: assessing the association between two or more variables (Quantitative and nominal variables)Quantitative Variable Associations:Regression Line: line of best fir that minimizes the squared distancesScatterplot:indicates scores of predictor vs. Outcome variablesLinear Relationship: when relationship can be easily seen through a straight line (can be positive or negative)Correlational Coefficients:Quantitative variablesPearsons RLinear relationships between quantitative variables-1≤ r ≥ + 1 direction and magnitude of associationCoefficient of determination (the proportion of variance)Effect size = rNominal Variables:Contingency coefficient--> total # individualsComparing observed frequencies vs. Expected frequenciesValue and statistical significance of coefficient and do NOT contain info about the pattern of the associationAssociations:Strong (linear) moderate no relationshipForms:LinearCurvilinearIndependentMultiple Regression: simultaneously considers the influence of more than one predictor on a single outcome (increases prediction ability and allows correlations between MULTIPLE variables)r=zero order correlationsR= multiple correlation coefficient (ability for predictor variables TOGETHER to predict the outcome variables)B= regression coefficient (UNIQUE effect of each predictor variable holding constant the effect of the other predictor variables)Correlation and CausalityReverse causation: causal relationship is the opposite of what is has been hypothesized asCommon causal variables: 3rd variable not part of research that cause both the predictor and outcome variable and produce a correlationSpurious Relationship: explain away the relationship between the predictor and outcome variablesExtraneous variables: reduce the likelihood of finding a correlation between the predictor and the outcome variables. Variable other than the predictor that produce the outcome variableMediating variables: variable caused by predictor variable that in turn causes the outcome variable. Explains WHY the relationship between the two variables occursUsing Correlational Data to Test Causal ModelsLongitudinal research: same individuals measured over time and time period is long enough so that changes in variables of interest could occur (can be limiting)Path Analysis: represents associations among a set of variablesControlling for Causal Variables: correlational data can be used to rule out reverse causation and common causal variable influences. Uses BOTH the predictor and potential common causal variable to predict the outcome variable.Structural Equation Analysis: tests if the observed relationships among a set of variables conform to a theoretical prediction about how those variables are causally relatedDesigned to