Method of clinical medical statistic (2)

"Universum Medical - Holistic Medical Center

2. The normal state sex of data examines

The object that statistic analyses is statistical data, the type that identifies statistical data correctly is reasonable the principal premise that chooses statistic to analyse a method. Normal school is one of type of distribution of common metric data data, the analytic method of metric on medical data asks data distributings is normal state or approximate normal state, because this needs,measure destiny to examine according to having normal state sex to primitive independence.

Through the frequency of scale data distributinging histogram judges sex of data distributinging normal state qualitatively. Such graphical judgement is strict normal state sex anything but examine, the information that it provides is the important complement that examines to normal state sex only.

The commonly used normal state sex in software examines to basically have 3 kinds of methods:

1) computation is integrated interconnected system is metric

Like moment law, Xiapiluo - power Er overcomes Shapiro-Wilk law (W examines) , Dagesidinuo D′Agostino law (D examines) , Shapiro-Francia law (W′ examines) .

2) of normal school plan to add up to goodness to examine

Like Pearson χ 2 examine, logarithm likelihood is compared examine, Kolmogorov Kolmogorov-Smirov law examines.

3) graphic representation law (Normal Probability Plot of normal state probability graph)

If divide digit to pursue (Quantile Quantileplot, abbreviation QQ pursues) , 100 minutes of digit (Percent Percent Plot, abbreviation PP pursues) change probability graph with stability (Stablized Probability Plot, abbreviation SP pursues) etc.

1) SPSS

Undertake normal state sex examines in SPSS, can use " Descriptive Statistics " the Explore method that describes sexual statistic method.

In Explore method, guide above all the data that wants an analysis, be like " Vital_HR " to " DependentList " in independent list. Must choose here " Normality Plots Withtests " normal state sex examines graphic representation, ability gets normal state sex analyses a result.

The eventuate that analysis of normal state sex examines is expressed twice, two graphs. Two watches are " Case Processing Summary " and " Tests OfNormality " . Two graphs are " Normal Q-Q Plot Of Vital_HR " and " Detrended Normal Q-QPlot Of Vital_HR.

Among them " Testsof Normality " the watch examines for normal state sex result. There are two kinds to examine in the watch result.

Tests Of Normality

Kolmogorov-Smirnova

Shapiro-Wilk

Statistic

Df

Sig.

Statistic

Df

Sig.

******

.100

80

.046

.967

80

.036

A. Lilliefors Significance Correction

2) SAS

Undertake normal state sex examines in SAS, need inputs process designing statement, following place show:

Procunivariate Normal Plot; carries out Univariate process and make normal state sex examine, normal state probability graph

　Var X; analyses variable X

　Pace of Run; process ends

3) result analysis

SPSS and SAS give Shapiro - Wilk (w to examine completely at the same time) examine with Kolmogorov -Smirnov(D) two interconnected system are metric. SPSS regulation: When ≤ of N of ≤ of example content 3 5000 when, the result examines with Shapiro - Wilk (W) embarrassed, when example content N>5000 results examine with Kolmogorov - Smirnov (D) to allow. And SAS regulation: When ≤ of example content N 2000 when, the result examines with Shapiro - Wilk (W) to allow, when example content N>2000, the result examines with Kolmogorov - Smirnov (D) to allow. By from the watch 2 knowable, exemple of this N = 110 of data example content, be worth affirmatory probability P by the W of SPSS and SAS plan cipher out>0. 05, cannot reject to examine suppose, can be normal school of this data obedience. From this, can pass examine the normal state sex of data of result test and verify.