To: Methodology of medical scientific research saves all take an examination ofing to training class student the 3rd part

Brian Shaffer Missing Ohio State Medical Student

9, when the statistical description method that chooses data of medical scientific research, what issue should consider?

The statistical description method of medical data has statistics law, statistical plan law and statistical index way, the specific feature that the suitable scope according to different method and medical data answer when specific choice has a consideration.

Statistics is the method of description of a kind of statistic that concerns with form form conveys the amount between the thing compendiously, it can replace the expatiatory character in medical data to narrate, and facilitate read and mutual intercrop is compared. The statistics in medical scientific research has fixed format and relatively unified tabulation demand, should try to notice when the choice.

Statistical graph is the method of description of a kind of statistic that will convey quantitative data with the geometrical graph such as dot, line, face, body, but size of amount of expression of ground of intuitionistic, figure and compare each other. When choosing, notice statistical graph also has corresponding n/arc drafting demand. Common statistical plan has 9 to plant (see 56 pages, browse by oneself please) , apply at different medical to study a data respectively.

When law of choice statistic index has a description, differ according to the type of medical data, chooses descriptive sex target also differs somewhat. To metric data, use average commonly (arithmetical average, geometry average, median, harmonic average, mode) will describe its to center a trend, use mutation index (extreme difference, standard deviation, variance, standard by accident, divide a span number, coefficient of variation) will describe its disperse level; To computation data, often use relative number (rate, than) have a description.

10, can be the error divided which two kinds be? What characteristic do each have?

The error is to show what the institute obtains is actual determine the difference between value and real value. The origin according to the error, property and whether can accuse, the error can be divided for systematic sum of errors random error.

Systematic error is when be being investigated or be being measured, as a result of some kind affirmatory reason causes, it can make findings slants big or slant small. Systematic error is mixed in fixed size direction, can appear when repeat investigation or be measured, but if use unified level, raise business technology level, enhance working sense of responsibility, systematic error can be eliminated.

Random error includes error of sampling of accidental sum of errors. Accidental error is caused; Error of sampling is the difference between the statistical quantity that the expression in pointing to sampling research comes out and statistical quantity or statistical quantity and parameter, cause by individual variability commonly. The size of random error and direction are fixed still, cannot eliminate. But of random error distributinging have the law, show normal school. Through raising a test the frequency can reduce accidental error hair, through increasing sample size can reduce error of sampling. Error of sampling can undertake estimating according to statistical method.