Flashcards
research skills handout 2
Cards in this deck (12)
- 1why do we keep the term standard error when its esentially a standard deviation (3)i) While the standard error is a form of standard deviation, it is a very special
form. It is the standard deviation of a hypothetical distribution that is never
actually observed. Values for the standard error usually need to be computed using a theoretically derived formula.
ii) Standard errors and standard deviations are put to different uses. Standard deviations are descriptive tools that indicate the dispersion in a sample. A standard error is an inferential tool, which measures the precision of estimates of population parameters.
iii) The fact that a standard error is a form of standard deviation can readily give rise to confusion. Standard error
is the term that has been widely used for the standard deviation of the distribution of sample means and to change nomenclature now may cause even greater confusion. - 2what is standard errorstandard deviation of sample means
- 3outline standard errorif you were to take several samples from the same population all of the same size not all of them would give the same sample mean. if you were to plot a histogram of the sample means you go they would form a normally distributed curve. said curve would form around the middle/mean of the population mean. said curve woudl have a standaard devaition telling you how spread out you would expect sample means to get from the population mean. this standard deviation is the standard error
- 4standard error gives a measure ofhow far estimates should be expected to depart from true value
- 5in practace standard error is worked out howstandard deviation/ the square root of n
- 6using the formular for standard error explain how increasing sample betters our estimate of population meanincreased sample size menas increased n. increased n means increased value of square root of n. as such bigger denominator in: SE = SD/ Square root of n. as such value of SE decreases. Smaller SE means better estimate
- 7alternative name for the 95% confidence intervealinterval estimate of mew
- 8alternative name for sample meanpiont estimate of mew
- 9why do you not alwasy want SE at smallest popssibkleit might not be of any benifit to increase localisation of mew to a smaller area. i.e. knowingwhere mew is in a 10 figure range may be just as benifitial as knwoing it in a 0.1 range. as such decreasing SE any further is not needed. to decrease SE further weould take more time and resourse
- 10T or F: the sample means from a normally distributed popuklation will always give a normal distribution themselvesT
- 11T or F: the sample means from a NOT normally distributed popuklation will always give a normal distribution themselvesF. most the time will but not always
- 12why is the following not corredt for dictating the 95% confidence interval (2): M +/-2 (sigma/square root n)1. its not technically 2 its 1.96 that is correct for dictatingw ehre 95% of the data pionts should lie
2. as we have to use SD not sigma. SD is not a perfect guess of sigma. as such the values gained will be wrong. increasing the value in front of the fraction decreases this impact. how much you increase it by is deteremined by how good an estimate SD iss of sigma. In other words its determined by n