Standard deviation, standard error and （标准差,标准误差）.pdfVIP

“Inference by Eye“ Standard deviation, standard error and confidence intervals as error bars Gabriela Czanner PhD CStat Department of Biostatistics Department of Eye and Vision Science 10 October 2012 MERSEY POSTGRADUATE TRAINING PROGRAMME Workshop Series: Basic Statistics for Eye Researchers and Clinicians Inference by eye “Inference by eye is the interpretation of graphically presented data” - Cumming, Finch 2005 Two goals to draw a picture • To enhance description of data • To visualize results of statistical inference On first seeing the figure on left, what questions should spring to our mind? • What is the centre? • What are the error bars i.e. the antenas in the picture? • 1 Standard deviation, 2 x Standard deviation (or 1.96), • Standard errors, Confidence interval (2 x Standard error), • Range i.e. min and max Outline • Fundamental statistical concepts • Mean • Standard deviation • Normal distribution • Standard error • Confidence interval for the mean • What error bars should we use? • References Fundamental statistical concepts We aim to learn about population. • Population is theoretical concept used to describe the entire group. • Characteristics of a population can be described by quantities called parameters. We choose sample of subjects and collect data on the sample. • Data are used to provide estimates of population Inference about parameters. populations of like • We describe the sample using graphs and individuals (or eyes) numerical summaries. Called descriptive statistics • We use data to make inference about the para