SOCIOL 5 Lecture Notes - Lecture 15: External Validity, Confidence Interval, Null Hypothesis
Document Summary
Central tendency pick one for each variable. R = 0 = no association with variable. In a college campus (shows bias) for example. Different phases in cities to show how proportional their representation was. How they unbiasedly chose people in hospitals. How they made samples representative as possible. Law of large numbers= bigger sample = mean mathes sample more closely. Asking only 10 people for mean or average of class not as clear as asking 100 people. When you graph means of sample (taking many sample) ask 10 people and other groups. Plotting it then there should be a normal distribution (bell curve symmetrical in histogram) Y axis is people in counts or frequency. X axis = percentage (modal in peak) (ex. Percentage of scores through the class people may have got) Cannot interview entire pop so sample as much as you can. Strong evidence still can have little bit of error. Wide or high confidence level and interval.