CMN 102 Chapter Notes - Chapter 8: Statistical Hypothesis Testing, Internal Validity, External Validity
Document Summary
The web has been used for conducting experiments. Natural experiments: experiments that occur in regular course of social events. A variable can be independent for one experiment and dependent for another. Both variables should be defined operationally: pretesting & post testing. Pretesting: subjects measured in terms of dependent variable. Then exposed to independent variable stimulus: experimental & control groups. Experimental receives the stimulus but the control does not. Using control group helps to detect any effect of the experiment. Degree of resemblance in probability sampling is function of sample size. Number the subject, then select by random # table. Or assign even and odd and separate by category. Should use when large pool, then probability laws apply. Uses quota matrix of all relevant characteristics. Description of experimental group should be similar to control group. Used when conditions for experiment can"t be met. 3 types: one-shot case study- group of subject measured on dependent variable.