Research Methodology in Social Sciences
Dr. David Sichinava
October 26, 2018
Parameter vs. Estimate
- Very often, we have to infer about population based on relatively small number of observations (sample)
- For instance, usually in the media polls are covered as parameters but they are estimates
Biased and Unbiased Estimates
- A good estimate has a sampling distribution which is centered around the parameter and has small standard error
Confidence Interval
- We are never sure how precise are our estimates, therefore it is safer to assume a particular margin that contains true population parameter by a high probability, say - close to 1
- This margin is called confidence interval. Confidence interval has a form: Point Estimate +/- margin of error (multiple of a standard error)
- Confidence intervals are chosen conventionally, usually, 95% or 99%
Confidence Interval for Proportion and Mean
Confidence Interval for Proportion and Mean
Sample size
- \( n = {\sigma}^2 * ({z}/M) \)