Page 55 - Quantitative Data Analysis
P. 55

Quantitative Data Analysis
                                              Simply Explained Using SPSS


               Type I Errors

               1.      Type I Error occurs when we find an effect or relationship in our
                       sample that is not in the population.
               2.      Have a significant result when in fact it is not significant.
               3.      Reject a true null hypothesis.
               4.      The probability of a Type I Error = α.
               5.      The conventional α value is .05.
               6.      Since the probability of a Type I Error is α, the researcher has
                       direct control over how likely a Type I Error would occur.
               7.      Using α= .05 means that even when no difference exists in the
                       population, 5% of random samples will show a significant
                       difference.
               8.      To reduce type I errors, simply use a smaller α, (i.e. .01 or .001).

               Type II Errors

               1.      Type II Error occurs when we do not find an effect or relationship
                       in our sample that exists in the population.
               2.      Have a non-significant result when in fact it is significant.
               3.      Fail to reject a false null hypothesis.
               4.      Probability of a Type II Error = β
               5.      Type II Errors (β) depends on the size of the difference in the
                       population and the sample size.
               6.      A similar measure is power (1 − β), the probability that a study
                       will produce a statistically significant result if the research
                       hypothesis is true.
               7.      The conventional level for power is .8, this is what researchers
                       strive for when planning a research study.

               How to increase power

               1.      Increase sample size
               2.      Intensify or prolong treatment
               3.      Increase the type I error rate



               The Theory and Applications of Statistical Inferences           39
   50   51   52   53   54   55   56   57   58   59   60