Alternative to NHST is confidence intervals and power analysis. A confidence interval gives an estimate of range of values which includes the unknown population parameter. Confidence intervals are more informative as compared to hypothesis test results (rejection or acceptance of Null hypothesis) since they provide a range of possible values for the unknown parameter.
Consider Confidence Interval for a Mean, Confidence interval for normal around the mean is given by mean±z_(α/2) SD
The true mean of the population is unknown. Confidence intervals tells us the range in which our true, unknown population mean will lie and we are corresponding percentage of confidence. If alpha = 0.05, then I am 95% confident that the true, unknown population mean is between the Lower confidence limit (LCL) and the upper confidence limit (UCL). This means I will be correct 95% of the time and wrong 5% of the time (the alpha error).
The width of the confidence interval is related to the confidence level, standard error, and sample size (n). The higher the percentage of confidence desired, the wider the confidence interval. The larger the standard error, the wider the confidence interval. The larger the n, the smaller the standard error, and hence a finer confidence interval.
Secondly power analysis, which is an important aspect of experimental design. It allows to determine the required sample size to detect an effect of a given size with a given degree of confidence.
Login | Sign Up
