1Group AnalysisMoving from the IndividualReport to the Group Analysis• What relationships did you find to beimportant in your two cases?• Does this work suggest variables thatshould either be eliminate from furtherconsideration or new variables that shouldbe considered?• Do they obtain across the 6 or 8 cases?Next Steps• Create the matrix that will form the basisfor your analysis.• The rows will be your cases.• The columns will represent your variables• Each cell then will contain the value of thatvariable for that case study.Implications• All cases have to use the same variables.• All cases have to have each variable measuredin the same way. Using the same data source isbest, but not absolutely essential.• Empty cells (Missing data) are a real problem.Try to get complete data. Make sure to sharedata sources.• If can’t get the data, you can either drop thevariable from the analysis or measure it in sucha way that everyone can get the data.2Techniques for QuantifyingQualitative Variables4Right20%Right0Australia3Humpback65%Humpback0US2Sperm80%Sperm1Canada1Orca90%Orca1UKSuccess ofrecoveryWhaleSpeciesRecoveryRateWhaleSpeciesKyotoRatificationCountryNote• When using a ranking, you need to documentyour technique such that others could followyour procedure and get the same ranking. Itcannot be purely subjective.• Sometimes you can turn a qualitative variableinto an continuous variable.• For example if the variable concerns theexistence of a policy (normally a 1 or 0), youcould use the number of years the policy hasbeen in force.• Or if it is a tax policy you could include the taxrate as a variable (cents/gallon).The Next Steps:Using your Matrix• Sort the matrix so the cases are orderedby the value of the dependent variable.• Observe the pattern, if any, that emerges.• Is that pattern broadly consistent with thehypothesis?• Are there individual cases that seem to beoutliers?Sorting your group matrix00410D00110C11311B11211AEV: HabitatProtectionEV:AwarenesscampaignEV:Public $EV:RegulationDV:RecoverySpecies3Analyzing Explanatory Variables244Right433Humpback122Sperm311OrcaRegulationBiologicalSuccess ofrecoveryWhaleSpeciesIdentifying Outliers I34Right43Humpback22Sperm11OrcaRegulationSuccess ofrecoveryWhaleSpeciesIdentifying Outliers II88Honduras77Ghana26France65Egypt44Denmark33Canada52Bulgaria11AustriaEV: IncomeDV: SustainabilityCountryWhat does this pattern suggest?88Honduras77Ghana66Egypt55Bulgaria34Canada13France42Denmark21AustriaEV: IncomeDV: SustainabilityCountry4Deepening the Analysis• Is your original hypothesis supported?– Yes: explain the mechanism, provide detailedevidence from your cases– No: why not? What there another variable that is moreimportant? Was there an intervening factor or acombination of factors?– Yes, but some cases did not fit: why, what did welearn from the “outliers?” Why, do you think, are theyoutliers?• What other factors/variables did you find to beimportant. Were they important across all cases,or in some cases? Explain and illustrate.The Steps: Summary• Harmonize all variables between cases.• Quantify and document all variables• Create the matrix• Sort the matrix by the dependent variable• Identify patterns and outliers• Test consistency of these patterns with your hypothesis orhypotheses.• Explain the findings including the consistency with hypothesisincluding some sense of what can be learned from the outliers.Comment (to the extent you can) on the strengths or weaknesses ofthose findings.• Draw out the implications of your findings.• Suggest areas where further research might be particularly