Training manual SAS

 

 

 

 

Training Manual

 

 

 

 

Data Analysis using SAS

 

 

 

 

Sujai Das

 

 

 

 

 

 

NIRJAFT, 12 Regent Park, Kolkata - 700040

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1. Introduction

SAS (Statistical Analysis System) software is comprehensive software which deals with many problems related to Statistical analysis, Spreadsheet, Data Creation, Graphics, etc. It is a layered, multivendor architecture. Regardless of the difference in hardware, operating systems, etc., the SAS applications look the same and produce the same results. The three components of the SAS System are Host, Portable Applications and Data. Host provides all the required interfaces between the SAS system and the operating environment. Functionalities and applications reside in Portable component and the user supplies the Data. We, in this course will be dealing with the software related to perform statistical analysis of data.

 

Windows of SAS

1. Program Editor : All the instructions are given here.

2. Log : Displays SAS statements submitted for execution and messages

3. Output : Gives the output generated

 

Rules for SAS Statements

1. SAS program communicates with computer by the SAS statements.

2. Each statement of SAS program must end with semicolon (;).

3. Each program must end with run statement.

4. Statements can be started from any column.

5. One can use upper case letters, lower case letters or the combination of the two.

 

Basic Sections of SAS Program

1. DATA section

2. CARDS section

3. PROCEDURE section

 

Data Section

We shall discuss some facts regarding data before we give the syntax for this section.

 

Data value: A single unit of information, such as name of the specie to which the tree belongs, height of one tree, etc.

 

Variable: A set of values that describe a specific data characteristic e.g. diameters of all trees in a group. The variable can have a name upto a maximum of 8 characters and must begin with a letter or underscore. Variables are of two types:

 

Character Variable: It is a combination of letters of alphabet, numbers and special characters or symbols.

 

Numeric Variable: It consists of numbers with or without decimal points and with + or -ve signs.

 

Observation: A set of data values for the same item i.e. all measurement on a tree. Data section starts with Data statements as

DATA NAME (it has to be supplied by the user);

 

Input Statements

Input statements are part of data section. This statement provides the SAS system the name of the variables with the format, if it is formatted.

 

List Directed Input

 Data are read in the order of variables given in input statement.

 Data values are separated by one or more spaces.

 Missing values are represented by period (.).

 Character values are followed by $ (dollar sign).

 

Example

Data A;

INPUT ID SEX $ AGE HEIGHT WEIGHT; CARDS;

1 M 23 68 155

2 F . 61 102

3. M 55 70 202

;

 

Column Input

Starting column for the variable can be indicated in the input statements for example:

 

INPUT ID 1-3 SEX $ 4 HEIGHT 5-6 WEIGHT 7-11; CARDS;

001M68155.5

2F61 99

3M53 33.5

;

 

Alternatively, starting column of the variable can be indicated along with its length as

INPUT @ 1 ID 3.

@ 4 SEX $ 1.

@ 9 AGE 2.

@ 11 HEIGHT 2.

@ 16 V_DATE MMDDYY 6.

;

Reading More than One Line Per Observation for One Record of Input Variables

 

 

INPUT # 1 ID 1-3 AGE 5-6 HEIGHT 10-11

# 2 SBP 5-7 DBP 8-10; CARDS;

001 56 72

140 80

;

 

Reading the Variable More than Once

Suppose id variable is read from six columns in which state code is given in last two columns of id variable for example:

 

INPUT @ 1 ID 6. @ 5 STATE 2.; OR

INPUT ID 1-6 STATE 5-6;

 

Formatted Lists

DATA B;

INPUT ID @1(X1-X2)(1.)

@4(Y1-Y2)(3.); CARDS;

11 563789

22 567987

;

PROC PRINT; RUN;

 

Output

Obs.

ID

x1

x2

y1

y2

1

11

1

1

563

789

2

22

2

2

567

987

 

DATA C;

INPUT X Y Z @; CARDS;

1 1 1 2 2 2 5 5 5 6 6 6

1 2 3 4 5 6 3 3 3 4 4 4

;

PROC PRINT; RUN;

 

Output

Obs. X Y Z

1 1 1 1

2 1 2 3

DATA D;

INPUT X Y Z @@;

 

CARDS;

1 1 1 2 2 2 5 5 5 6 6 6

1 2 3 4 5 6 3 3 3 4 4 4

;

PROC PRINT; RUN;

 

Output:

Obs.

X

Y

Z

1

1

1

1

2

2

2

2

3

5

5

5

4

6

6

6

5

1

2

3

6

4

5

6

7

3

3

3

8

4

4

4

 

SAS System Can Read and Write

DATA FILES

A. Simple ASCII files are read with input and infile statements

B. Output Data files

Creation of SAS Data Set

DATA EX1;

INPUT GROUP $ X Y Z; CARDS;

T1 12 17 19

T2 23 56 45

T3 19 28 12

T4 22 23 36

T5 34 23 56

;

Creation of SAS File From An External (ASCII) File

DATA EX2;

INFILE 'B:MYDATA'; INPUT GROUP $ X Y Z;

OR

DATA EX2A;

FILENAME ABC 'B:MYDATA'; INFILE ABC;

INPUT GROUP $ X Y Z;

;

 

Creation of A SAS Data Set and An Output ASCII File Using an External File

DATA EX3;

FILENAME IN 'C:MYDATA';

 

FILENAME OUT 'A:NEWDATA'; INFILE IN;

FILE OUT;

INPUT GROUP $ X Y Z; TOTAL =SUM (X+Y+Z);

PUT GROUP $ 1-10 @12 (X Y Z TOTAL)(5.); RUN;

 

This above program reads raw data file from 'C: MYDATA', and creates a new variable TOTAL

and writes output in the file 'A: NEWDATA’.

 

Creation of SAS File from an External (*.csv) File

data EX4;

infile'C:\Users\Admn\Desktop\sscnars.csv' dlm=',' ;

/*give the exact path of the file, file should not have column headings*/

input sn loc $ year season $ crop $ rep trt gyield syield return kcal; /*give the variables in ordered list in the file*/

/*if we have the first row as names of the columns then we can write in the above statement

firstobs=2 so that data is read from row 2 onwards*/ biomass=gyield+syield; /*generates a new variable*/ proc print data=EX4;

run;

 

Note: To create a SAS File from a *.txt file, only change csv to txt and define delimiter as per file created.

 

Creation of SAS File from an External (*.xls) File

Note: it is always better to copy the name of the variables as comment line before Proc Import.

/* name of the variables in Excel File provided the first row contains variable name*/

proc import datafile = 'C:\Users\Desktop\DATA_EXERCISE\descriptive_stats.xls'

/*give the exact path of the file*/

out = descriptive_stats replace; /*give output file name*/

proc print;

run;

If we want to make some transformations, then we may use the following statements:

data a1;

set descriptive_stats;

x = fs45+fw;

run;

 

Here proc import allows the SAS user to import data from an EXCEL spreadsheet into SAS. The datafile statement provides the reference location of the file. The out statement is used to name the SAS data set that has been created by the import procedure. Print procedure has been utilized to view the contents of the SAS data set descriptive_stats. When we run above codes we obtain the output which will same as shown above because we are using the same data.

 

Creating a Permanent SAS Data Set LIBNAME XYZ 'C:\SASDATA'; DATA XYZ.EXAMPLE;

INPUT GROUP $ X Y Z; CARDS;

.....

.....

..... RUN;

 

This program reads data following the cards statement and creates a permanent SAS data set in a subdirectory named \SASDATA on the C: drive.

 

Using Permanent SAS File

LIBNAME XYZ 'C:\SASDATA';

PROC MEANS DATA=XYZ.EXAMPLE; RUN;

 

TITLES

One can enter upto 10 titles at the top of output using TITLE statement in your procedure.

 

PROC PRINT;

TITLE ‘HEIGHT-DIA STUDY’; TITLE3 ‘1999 STATISTICS’; RUN;

 

Comment cards can be added to the SAS program using

/* COMMENTS */;

FOOTNOTES

One can enter upto 10 footnotes at the bottom of your output.

 

PROC PRINT DATA=DIAHT; FOOTNOTE ‘1999’;

FOOTNOTE5 ‘STUDY RESULTS’; RUN;

 

For obtaining output as RTF file, use the following statements

Ods rtf file=’xyz.rtf’ style =journal; Ods rtf close;

 

For obtaining output as PDF/HTML file, replace rtf with pdf or html in the above statements. If we want to get the output in continuos format, then we may use

Ods rtf file=’xyz.rtf’ style =journal bodytitle startpage=no;

LABELLING THE VARIABLES

Data dose;

title ‘yield with factors N P K’;

input N P K Yield;

 

Label N = “Nitrogen”; Label P = “ Phosphorus”; Label K = “ Potassium”; cards;

...

...

...

;

Proc print;

run;

We can define the linesize in the output using statement OPTIONS. For example, if we wish that the output should have the linesize (number of columns in a line) is 72 use Options linesize

=72; in the beginning.

 

2. Statistical Procedure

SAS/STAT has many capabilities using different procedures with many options. There are a total of 73 PROCS in SAS/STAT. SAS/STAT is capable of performing a wide range of statistical analysis that includes:

1. Elementary / Basic Statistics

2. Graphs/Plots

3. Regression and Correlation Analysis

4. Analysis of Variance

5. Experimental Data Analysis

6. Multivariate Analysis

7. Principal Component Analysis

8. Discriminant Analysis

9. Cluster Analysis

10. Survey Data Analysis

11. Mixed model analysis

12. Variance Components Estimation

13. Probit Analysis and many more…

A brief on SAS/STAT Procedures is available at http://support.sas.com/rnd/app/da/stat/procedures/Procedures.html

 

Example 2.1: To Calculate the Means and Standard Deviation: DATA TESTMEAN;

INPUT GROUP $ X Y Z; CARDS;

CONTROL 12 17 19

TREAT1 23 25 29

TREAT2 19 18 16

TREAT3 22 24 29

CONTROL 13 16 17

TREAT1 20 24 28

TREAT2 16 19 15

 

TREAT3 24 26 30

CONTROL 14 19 21

TREAT1 23 25 29

TREAT2 18 19 17

TREAT3 23 25 30

;

PROC MEANS; VAR X Y Z; RUN;

 

The default output displays mean, standard deviation, minimum value, maximum value of the desired variable. We can choose the required statistics from the options of PROC MEANS. For example, if we require mean, standard deviation, median, coefficient of variation, coefficient of skewness, coefficient of kurtosis, etc., then we can write

 

PROC MEANS mean std median cv skewness kurtosis; VAR X Y Z;

RUN;

The default output is 6 decimal places, desired number of decimal places can be defined by using option maxdec=…. For example, for an output with three decimal places, we may write

 

PROC MEANS mean std median cv skewness kurtosis maxdec=3; VAR X Y Z;

RUN;

 

For obtaining means group wise use, first sort the data by groups using

 

Proc sort; By group; Run;

And then make use of the following

PROC MEANS; VAR X Y Z;

by group; RUN;

Or alternatively, me may use PROC MEANS; CLASS GROUP; VAR X Y Z;

RUN;

For obtaining descriptive statistics for a given data one can use PROC SUMMARY. In the above example, if one wants to obtain mean standard deviation, coefficient of variation, coefficient of skewness and kurtosis, then one may utilize the following:

 

PROC SUMMARY PRINT MEAN STD CV SKEWNESS KURTOSIS; CLASS GROUP;

 

VAR X Y Z; RUN;

 

Most of the Statistical Procedures require that the data should be normally distributed. For testing the normality of data, PROC UNIVARIATE may be utilized.

 

PROC UNIVARIATE NORMAL; VAR X Y Z;

RUN;

 

If different plots are required then, one may use:

PROC UNIVARIATE DATA=TEST NORMAL PLOT;

/*plot option displays stem-leaf, boxplot & Normal prob plot*/ VAR X Y Z;

/*creates side by side BOX-PLOT group-wise. To use this option first sort the file on by variable*/

BY GROUP;

HISTOGRAM/KERNEL NORMAL; /*displays kernel density along with normal curve*/ PROBPLOT; /*plots probability plot*/

QQPLOT X/NORMAL SQUARE; /*plot quantile-quantile QQ-plot*/

CDFPLOT X/NORMAL; /*plots CDF plot*/

/*plots pp plot which compares the empirical cumulative distribution function (ecdf) of a variable with a specified theoretical cumulative distribution function. The beta, exponential, gamma, lognormal, normal, and Weibull distributions are available in both statements.*/

PPPLOT X/NORMAL;

RUN;

 

Example 2.2: To Create Frequency Tables

DATA TESTFREQ;

INPUT AGE $ ECG CHD $ CAT $ WT; CARDS;

<55 0 YES YES 1

<55 0 YES YES 17

<55 0 NO YES 7

<55 1 YES NO 257

<55 1 YES YES 3

<55 1 YES NO 7

<55 1 NO YES 1

55+ 0 YES YES 9

55+ 0 YES NO 15

55+ 0 NO YES 30

55+ 1 NO NO 107

55+ 1 YES YES 14

55+ 1 YES NO 5

55+ 1 NO YES 44

55+ 1 NO NO 27

;

PROC FREQ DATA=TESTFREQ;

 

TABLES AGE*ECG/MISSING CHISQ; TABLES AGE*CAT/LIST;

RUN:

 

SCATTER PLOT

PROC PLOT DATA = DIAHT; PLOT HT*DIA = ‘*’;

/*HT=VERTICAL AXIS DIA = HORIZONTAL AXIS.*/ RUN;

 

CHART

PROC CHART DATA = DIAHT; VBAR HT;

RUN;

 

PROC CHART DATA = DIAHT; HBAR DIA;

RUN;

 

PROC CHART DATA = DIAHT; PIE HT;

RUN;

 

Example 2.3: To Create A Permanent SAS DATASET and use that for Regression

LIBNAME FILEX 'C:\SAS\RPLIB'; DATA FILEX.RP;

INPUT X1-X5;

CARDS;

1 0 0 0 5.2

.75 .25 0 0 7.2

.75 0 .25 0 5.8

.5 .25 .25 0 6.3

.75 0 0 .25 5.5

.5 0 .25 .25 5.7

.5 .25 0 .25 5.8

.25 .25 .25 .25 5.7

; RUN;

 

 

 

 

LIBNAME FILEX 'C:\SAS\RPLIB'; PROC REG DATA=FILEX.RP; MODEL X5 = X1 X2/P;

MODEL X5 = X1 X2 X3 X4 / SELECTION = STEPWISE;

TEST: TEST X1-X2=0; RUN;

 

 

Various other commonly used PROC Statements are PROC ANOVA, PROC GLM; PROC CORR; PROC NESTED; PROC MIXED; PROC RSREG; PROC IML; PROC PRINCOMP; PROC VARCOMP; PROC FACTOR; PROC CANCORR; PROC DISCRIM, etc. Some of these are described in the sequel.

 

PROC TTEST is the procedure that is used for comparing the mean of a given sample. This PROC is also used for compares the means of two independent samples. The paired observations t test compares the mean of the differences in the observations to a given number. The underlying assumption of the t test in all three cases is that the observations are random samples drawn from normally distributed populations. This assumption can be checked using the UNIVARIATE procedure; if the normality assumptions for the t test are not satisfied, one should analyze the data using the NPAR1WAY procedure. PROC TTEST computes the group comparison t statistic based on the assumption that the variances of the two groups are equal. It also computes an approximate t based on the assumption that the variances are unequal (the Behrens-Fisher problem). The following statements are available in PROC TTEST.

PROC TTEST <options>; CLASS variable;

PAIRED variables; BY variables;

VAR variables; FREQ Variables; WEIGHT variable;

 

No statement can be used more than once. There is no restriction on the order of the statements after the PROC statement. The following options can appear in the PROC TTEST statement. ALPHA= p: option specifies that confidence intervals are to be 100(1-p)% confidence intervals, where 0<p<1. By default, PROC TTEST uses ALPHA=0.05. If p is 0 or less, or 1 or more, an error message is printed.

COCHRAN: option requests the Cochran and Cox approximation of the probability level of the

approximate t statistic for the unequal variances situation.

H0=m: option requests tests against m instead of 0 in all three situations (one-sample, two- sample, and paired observation t tests). By default, PROC TTEST uses H0=0.

 

A CLASS statement giving the name of the classification (or grouping) variable must accompany the PROC TTEST statement in the two independent sample cases. It should be omitted for the one sample or paired comparison situations. The class variable must have two, and only two, levels. PROC TTEST divides the observations into the two groups for the t test using the levels of this variable. One can use either a numeric or a character variable in the CLASS statement.

In the statement PAIRED PairLists, the PairLists in the PAIRED statement identifies the variables to be compared in paired comparisons. You can use one or more PairLists. Variables or lists of variables are separated by an asterisk (*) or a colon (:). Examples of the use of the asterisk and the colon are shown in the following table.

 

The PAIRED Statements

Comparisons made

PAIRED A*B;

A-B

 

 

PAIRED A*B C*D;

A-B and C-D

PAIRED (A B)*(C B);

A-C, A-B and B-C

PAIRED (A1-A2)*(B1-B2);

A1-B1, A1-B2, A2-B1 and A2-B2

PAIRED (A1-A2):(B1-B2);

A1-B1 and A2-B2

 

PROC ANOVA performs analysis of variance for balanced data only from a wide variety of experimental designs whereas PROC GLM can analyze both balanced and unbalanced data. As ANOVA takes into account the special features of a balanced design, it is faster and uses less storage than PROC GLM for balanced data. The basic syntax of the ANOVA procedure is as given:

 

PROC ANOVA < Options>; CLASS variables;

MODEL dependents = independent variables (or effects)/options;

MEANS effects/options; ABSORB variables; FREQ variables;

TEST H = effects E = effect; MANOVA H = effects E = effect;

M = equations/options; REPEATED factor - name levels / options; By variables;

 

The PROC ANOVA, CLASS and MODEL statements are must. The other statements are optional. The CLASS statement defines the variables for classification (numeric or character variables - maximum characters =16).

 

The MODEL statement names the dependent variables and independent variables or effects. If no effects are specified in the MODEL statement, ANOVA fits only the intercept. Included in the ANOVA output are F-tests of all effects in the MODEL statement. All of these F-tests use residual mean squares as the error term. The MEANS statement produces tables of the means corresponding to the list of effects. Among the options available in the MEANS statement are several multiple comparison procedures viz. Least Significant Difference (LSD), Duncan’s New multiple - range test (DUNCAN), Waller - Duncan (WALLER) test, Tukey’s Honest Significant Difference (TUKEY). The LSD, DUNCAN and TUKEY options takes level of significance ALPHA = 5% unless ALPHA = options is specified. Only ALPHA = 1%, 5% and 10% are allowed with the Duncan’s test. 95% Confidence intervals about means can be obtained using CLM option under MEANS statement.

 

The TEST statement tests for the effects where the residual mean square is not the appropriate term such as main - plot effects in split - plot experiment. There can be multiple MEANS and TEST statements (as well as in PROC GLM), but only one MODEL statement preceded by RUN statement. The ABSORB statement implements the technique of absorption, which saves time and reduces storage requirements for certain type of models. FREQ statement is used when each observation in a data set represents ‘n’ observations, where n is the value of FREQ variable. The MANOVA statement is used for implementing multivariate analysis of variance. The

 

REPEATED statement is useful for analyzing repeated measurement designs and the BY statement specifies that separate analysis are performed on observations in groups defined by the BY variables.

 

PROC GLM for analysis of variance is similar to using PROC ANOVA. The statements listed for PROC ANOVA are also used for PROC GLM. In addition; the following more statements can be used with PROC GLM:

 

CONTRAST ‘label’ effect name< ... effect coefficients > </options>; ESTIMATE ‘label’ effect name< ... effect coefficients > </options>; ID variables;

LSMEANS effects < / options >;

OUTPUT < OUT = SAS-data-set>keyword=names< ... keyword = names>; RANDOM effects < / options >;

WEIGHT variables

 

Multiple comparisons as used in the options under MEANS statement are useful when there are no particular comparisons of special interest. But there do occur situations where preplanned comparisons are required to be made. Using the CONTRAST, LSMEANS statement, we can test specific hypothesis regarding pre - planned comparisons. The basic form of the CONTRAST statement is as described above, where label is a character string used for labeling output, effect name is class variable (which is independent) and effect - coefficients is a list of numbers that specifies the linear combination parameters in the null hypothesis. The contrast is a linear function such that the elements of the coefficient vector sum to 0 for each effect. While using the CONTRAST statements, following points should be kept in mind.

 

How many levels (classes) are there for that effect. If there are more levels of that effect in the data than the number of coefficients specified in the CONTRAST statement, the PROC GLM adds trailing zeros. Suppose there are 5 treatments in a completely randomized design denoted as T1, T2, T3, T4, T5 and null hypothesis to be tested is

Ho: T2+T3 = 2T1 or 2T1+T2+T3 = 0

 

Suppose in the data treatments are classified using TRT as class variable, then effect name is TRT CONTRAST ‘TIVS 2&3’ TRT 2 1 1 0 0; Suppose last 2 zeros are not given, the trailing zeros can be added automatically. The use of this statement gives a sum of squares with

1 degree of freedom (d.f.) and F-value against error as residual mean squares until specified. The name or label of the contrast must be 20 characters or less.

 

 

 

The available CONTRAST statement options are

E: prints the entire vector of coefficients in the linear function, i.e., contrast. E = effect: specifies an effect in the model that can be used as an error term ETYPE = n: specifies the types (1, 2, 3 or 4) of the E effect.

 

Multiple degrees of freedom contrasts can be specified by repeating the effect name and coefficients as needed separated by commas. Thus the statement for the above example

CONTRAST ‘All’ TRT 2 1 1 0 0, TRT 0 1 -1 0 0;

 

 

 

This statement produces two d.f. sum of squares due to both the contrasts. This feature can be used to obtain partial sums of squares for effects through the reduction principle, using sums of squares from multiple degrees of freedom contrasts that include and exclude the desired contrasts. Although only t1 linearly independent contrasts exists for t classes, any number of contrasts can be specified.

 

The ESTIMATE statement can be used to estimate linear functions of parameters that may or may not be obtained by using CONTRAST or LSMEANS statement. For the specification of the statement only word CONTRAST is to be replaced by ESTIMATE in CONTRAST statement.

 

Fractions in effects coefficients can be avoided by using DIVISOR = Common denominator as an option. This statement provides the value of an estimate, a standard error and a t-statistic for testing whether the estimate is significantly different from zero.

 

The LSMEANS statement produces the least square estimates of CLASS variable means i.e. adjusted means. For one-way structure, there are simply the ordinary means. The least squares means for the five treatments for all dependent variables in the model statement can be obtained using the statement.

LSMEANS TRT / options;

Various options available with this statement are:

STDERR: gives the standard errors of each of the estimated least square mean and the t-statistic for a test of hypothesis that the mean is zero.

 

PDIFF: Prints the p - values for the tests of equality of all pairs of CLASS means.

 

SINGULAR: tunes the estimability checking. The options E, E=, E-TYPE = are similar as discussed under CONTRAST statement.

 

Adjust=T: gives the probabilities of significance of pairwise comparisons based on T-test. Adjust=Tukey: gives the probabilities of significance of pairwise comparisons based on Tukey's

test

Lines: gives the letters on treatments showing significant and non-significant groups

 

When the predicted values are requested as a MODEL statement option, values of variable specified in the ID statement are printed for identification besides each observed, predicted and residual value. The OUTPUT statement produces an output data set that contains the original data set values alongwith the predicted and residual values.

 

Besides other options in PROC GLM under MODEL statement we can give the option: 1. solution 2. xpx (=X`X) 3 . I (g-inverse)

 

PROC GLM recognizes different theoretical approaches to ANOVA by providing four types of sums of squares and associated statistics. The four types of sums of squares in PROC GLM are called Type I, Type II, Type III and Type IV.

 

The Type I sums of squares are the classical sequential sums of squares obtained by adding the terms to the model in some logical sequence. The sum of squares for each class of effects is adjusted for only those effects that precede it in the model. Thus the sums of squares and their expectations are dependent on the order in which the model is specified.

 

The Type II, III and IV are ‘partial sums of squares' in the sense that each is adjusted for all other classes of the effects in the model, but each is adjusted according to different rules. One general rule applies to all three types: the estimable functions that generate the sums of squares for one class of squares will not involve any other classes of effects except those that “contain” the class of effects in question.

 

For example, the estimable functions that generate SS (AB) in a three- factor factorial will have zero coefficients on main effects and the (A  C) and (B  C) interaction effects. They will contain non-zero coefficient on the (A  B  C) interaction effects, because A  B  C interaction “contains” A  B interaction.

 

Type II, III and IV sums of squares differ from each other in how the coefficients are determined for the classes of effects that do not have zero coefficients - those that contain the class of effects in question. The estimable functions for the Type II sum of squares impose no restriction on the values of the non-zero coefficients on the remaining effects; they are allowed to take whatever values result from the computations adjusting for effects that are required to have zero coefficients. Thus, the coefficients on the higher-order interaction effects and higher level nesting effects are functions of the number of observations in the data. In general, the Type II sums of squares do not possess of equitable distribution property and orthogonality characteristic of balanced data.

 

The Type III and IV sums of squares differ from the Type II sums of squares in the sense that the coefficients on the higher order interaction or nested effects that contain the effects in question are also adjusted so as to satisfy either the orthogonality condition (Type III) or the equitable distribution property (Type IV).

 

The coefficients on these effects are no longer functions of the nij and consequently, are the same for all designs with the same general form of estimable functions. If there are no empty cells (no nij = 0) both conditions can be satisfied at the same time and Type III and Type IV sums of squares are equal. The hypothesis being tested is the same as when the data is balanced.

 

When there are empty cells, the hypotheses being tested by the Type III and Type IV sums of squares may differ. The Type III criterion of orthogonality reproduces the same hypotheses one obtains if effects are assumed to add to zero. When there are empty cells this is modified to “the effects that are present are assumed to be zero”. The Type IV hypotheses utilize balanced subsets of non-empty cells and may not be unique. For a 2x3 factorial for illustration purpose adding the terms to the model in the order A, B, AB various types sums of squares can be explained as follows:

 

Effect

Type I

Type II

Type III

Type IV

General Mean

R()

R()

 

 

A

R(A/ )

R(A/ ,B)

R(A/,B,AB)

 

 

 

B

R(B/,A)

R(B/,A)

R(B/,A,AB)

 

A*B

R(A*B/ ,A,B)

R(A*B/,A,B)

R(AB/,A,B)

 

R (A/) is sum of squares adjusted for , and so on.

 

Thus in brief the four sets of sums of squares Type I, II, III & IV can be thought of respectively as sequential, each - after-all others, -restrictions and hypotheses.

 

There is a relationship between the four types of sums of squares and four types of data structures (balanced and orthogonal, unbalanced and orthogonal, unbalanced and non-orthogonal (all cells filled), unbalanced and non-orthogonal (empty cells)). For illustration, let nIJ denote the number of observations in level I of factor A and level j of factor B. Following table explains the relationship in data structures and Types of sums of squares in a two-way classified data.

 

Data Structure Type

 

Effect

1

Equal nIJ

2

Proportionate

3

Disproportionate

4

Empty Cell

 

A

 

I=II=III=IV

nIJ

I=II,III=IV

non-zero nIJ

III=IV

 

B

I=II=III=IV

I=II,III=IV

I=II,III=IV

I=II

A*B

I=II=III=IV

I=II=III=IV

I=II=III=IV

I=II=III=IV

 

In general,

 

I=II=III=IV

(balanced data); II=III=IV

(no interaction models)

I=II, III=IV

(orthogonal data); III=IV

(all cells filled data).

Proper Error terms: In general F-tests of hypotheses in ANOVA use the residual mean squares in other terms are to be used as error terms. For such situations PROC GLM provides the TEST statement which is identical to the test statement available in PROC ANOVA. PROC GLM also allows specification of appropriate error terms in MEANS LSMEANS and CONTRAST statements. To illustrate it let us use split plot experiment involving the yield of different irrigation (IRRIG) treatments applied to main plots and cultivars (CULT) applied to subplots. The data so obtained can be analysed using the following statements.

 

 

 

 

data splitplot;

input REP IRRIG CULT YIELD;

cards;

. . .

. . .

. . .

;

PROC print; run; PROC GLM;

class rep, irrig cult;

 

model yield = rep irrig rep*irrig cult irrig* cult;

test h = irrig e = rep * irrig;

contrast ‘IRRIGI Vs IRRIG2’ irrig 1 -1 / e = rep* irrig;

run;

 

As we know here that the irrigation effects are tested using error (A) which is sum of squares due to rep* irrig, as taken in test statement and contrast statement respectively.

 

 

 

It may be noted here that the PROC GLM can be used to perform analysis of covariance as well. For analysis of covariance, the covariate should be defined in the model without specifying under CLASS statement.

 

PROC RSREG fits the parameters of a complete quadratic response surface and analyses the fitted surface to determine the factor levels of optimum response and performs a ridge analysis to search for the region of optimum response.

PROC RSREG < options >;

MODEL responses = independents / <options >; RIDGE < options >;

WEIGHT variable; ID variable;

By variable;

run;

The PROC RSREG and model statements are required. The BY, ID, MODEL, RIDGE, and WEIGHT statements are described after the PROC RSREG statement below and can appear in any order.

 

The PROC RSREG statement invokes the procedure and following options are allowed with the

PROC RSREG:

DATA = SAS - data-set : specifies the data to be analysed.

NOPRINT : suppresses all printed results when only the output data set is required.

OUT : SAS-data-set: creates an output data set.

 

The model statement without any options transforms the independent variables to the coded data. By default, PROC RSREG computes the linear transformation to perform the coding of variables by subtracting average of highest and lowest values of the independent variable from the original value and dividing by half of their differences. Canonical and ridge analyses are performed to the model fit to the coded data. The important options available with the model statement are: NOCODE : Analyses the original data.

ACTUAL : specifies the actual values from the input data set.

COVAR = n : declares that the first n variables on the independent side of the model are simple linear regression (covariates) rather than factors in the quadratic response surface.

LACKFIT : Performs lack of fit test. For this the repeated observations must appear

together.

NOANOVA : suppresses the printing of the analysis of variance and parameter

 

estimates from the model fit.

NOOPTIMAL (NOOPT): suppresses the printing of canonical analysis for quadratic response surface.

NOPRINT : suppresses both ANOVA and the canonical analysis. PREDICT : specifies the values predicted by the model. RESIDUAL : specifies the residuals.

A RIDGE statement computes the ridge of the optimum response. Following important options

available with RIDGE statement are

MAX: computes the ridge of maximum response. MIN: computes the ridge of the minimum response.

 

At least one of the two options must be specified.

NOPRINT: suppresses printing the ridge analysis only when an output data set is required. OUTR = SAS-data-set: creates an output data set containing the computed optimum ridge. RADIUS = coded-radii: gives the distances from the ridge starting point at which to compute the optimum.

 

PROC REG is the primary SAS procedure for performing the computations for a statistical analysis of data based on a linear regression model. The basic statements for performing such an analysis are

 

PROC REG;

MODEL list of dependent variable = list of independent variables/ model options; RUN;

The PROC REG procedure and model statement without any option gives ANOVA, root mean square error, R-squares, Adjusted R-square, coefficient of variation etc.

The options under model statement are

 

P: It gives predicted values corresponding to each observation in the data set. The estimated standard errors are also given by using this option.

CLM: It yields upper and lower 95% confidence limits for the mean of subpopulation corresponding to specific values of the independent variables.

CLI : It yields a prediction interval for a single unit to be drawn at random from a subpopulation.

STB: Standardized regression coefficients.

XPX, I: Prints matrices used in regression computations.

NOINT: This option forces the regression response to pass through the origin. With this option total sum of squares is uncorrected and hence R-square statistic are much larger than those for the models with intercept.

However, if no intercept model is to be fitted with corrected total sum of squares and hence usual definition of various statistic viz R2, MSE etc. are to be retained then the option RESTRICT intercept = 0; may be exercised after the model statement.

For obtaining residuals and studentized residuals, the option ‘R’ may be exercised under model statement and Cook’s D statistic.

 

The ‘INFLUENCE’ option under model statement is used for detection of outliers in the data and provides residuals, studentized residuals, diagonal elements of HAT MATRIX, COVRATIO, DFFITS, DFBETAS, etc.

For detecting multicollinearity in the data, the options ‘VIF’ (variance inflation factors) and

‘COLLINOINT’ or ‘COLLIN’ may be used.

 

Besides the options for weighted regression, output data sets, specification error, heterogeneous variances etc. are available under PROC REG.

 

PROC PRINCOMP can be utilized to perform the principal component analysis.

 

Multiple model statements are permitted in PROC REG unlike PROC ANOVA and PROC GLM. A model statement can contain several dependent variables.

 

The statement model y1, y2, y3, y4=x1 x2 x3 x4 x5 ; performs four separate regression analyses of variables y1, y2, y3 and y4 on the set of variables x1, x2, x3, x4, x 5.

 

Polynomial models can be fitted by using independent variables in the model as x1=x, x2=x**2, x3=x**3, and so on depending upon the order of the polynomial to be fitted. From a variable, several other variables can be generated before the model statement and transformed variables can be used in model statement. LY and LX gives Logarithms of Y & X respectively to the base e and LogY, LogX gives logarithms of Y and X respectively to the base 10.

TEST statement after the model statement can be utilized to test hypotheses on individual or any linear function(s) of the parameters.

 

For e.g. if one wants to test the equality of coefficients of x1 and x2 in y=o+1x1+2 x2

regression model, statement

 

TEST 1: TEST x1 - x2 = 0;

Label: Test < equation ..., equation >;

 

The fitted model can be changed by using a separate model statement or by using DELETE

variables; or ADD variables; statements.

 

The PROC REG provides two types of sums of squares obtained by SS1 or SS2 options under model statement. Type I SS are sequential sum of squares and Types II sum of squares are partial SS are same for that variable which is fitted at last in the model.

 

For most applications, the desired test for a single parameter is based on the Type II sum of squares, which are equivalent to the t-tests for the parameter estimates. The Type I sum of squares, however, are useful if there is a need for a specific sequencing of tests on individual coefficients as in polynomial models.

 

PROC ANOVA and PROC GLM are general purpose procedures that can be used for a broad range of data classification. In contrast, PROC NESTED is a specialized procedure that is useful only for nested classifications. It provides estimates of the components of variance using the analysis of variance method of estimation. The CLASS statement in PROC NESTED has a

 

broader purpose then it does in PROC ANOVA and PROC GLM; it encompasses the purpose of MODEL statement as well. But the data must be sorted appropriately. For example in a laboratory microbial counts are made in a study, whose objective is to assess the source of variation in number of microbes. For this study n1 packages of the test material are purchased and n2 samples are drawn from each package i.e. samples are nested within packages. Let logarithm transformation is to be used for microbial counts. PROPER SAS statements are:

PROC SORT; By package sample; PROC NESTED;

CLASS package sample; Var logcount;

run;

Corresponding PROC GLM statements are

PROC GLM;

Class package sample;

Model Logcount= package sample (package);

 

The F-statistic in basic PROC GLM output is not necessarily correct. For this RANDOM statement with a list of all random effects in the model is used and Test option is utilized to get correct error term. However, for fixed effect models same arguments for proper error terms hold as in PROC GLM and PROC ANOVA. For the analysis of the data using linear mixed effects model, PROC MIXED of SAS should be used. The best linear unbiased predictors and solutions for random and fixed effects can be obtained by using option ‘s’ in the Random statement.

 

PROCEDURES FOR SURVEY DATA ANALYSIS

PROC SURVEYMEANS procedure produces estimates of population means and totals from sample survey data. You can use PROC SURVEYMEANS to compute the following statistics:

 estimates of population means, with corresponding standard errors and t tests

 estimates of population totals, with corresponding standard deviations and t tests

 estimates of proportions for categorical variables, with standard errors and t tests

 ratio estimates of population means and proportions, and their standard errors

 confidence limits for population means, totals, and proportions

 data summary information

PROC SURVEYFREQ procedure produces one-way to n-way frequency and crosstabulation tables from sample survey data. These tables include estimates of population totals, population proportions (overall proportions, and also row and column proportions), and corresponding standard errors. Confidence limits, coefficients of variation, and design effects are also available. The procedure also provides a variety of options to customize your table display.

PROC SURVEYREG procedure fits linear models for survey data and computes regression coefficients and their variance-covariance matrix. The procedure allows you to specify classification effects using the same syntax as in the GLM procedure. The procedure also provides hypothesis tests for the model effects, for any specified estimable linear functions of the model parameters, and for custom hypothesis tests for linear combinations of the regression parameters. The procedure also computes the confidence limits of the parameter estimates and their linear estimable functions.

 

PROC SURVEYLOGISTIC procedure investigates the relationship between discrete responses and a set of explanatory variables for survey data. The procedure fits linear logistic regression models for discrete response survey data by the method of maximum likelihood, incorporating the sample design into the analysis. The SURVEYLOGISTIC procedure enables you to use categorical classification variables (also known as CLASS variables) as explanatory variables in an explanatory model, using the familiar syntax for main effects and interactions employed in the GLM and LOGISTIC procedures.

 

The SURVEYSELECT procedure provides a variety of methods for selecting probability-based random samples. The procedure can select a simple random sample or a sample according to a complex multistage sample design that includes stratification, clustering, and unequal probabilities of selection. With probability sampling, each unit in the survey population has a known, positive probability of selection. This property of probability sampling avoids selection bias and enables you to use statistical theory to make valid inferences from the sample to the survey population.

PROC SURVEYSELECT provides methods for both equal probability sampling and sampling with probability proportional to size (PPS). In PPS sampling, a unit's selection probability is proportional to its size measure. PPS sampling is often used in cluster sampling, where you select clusters (groups of sampling units) of varying size in the first stage of selection. Available PPS methods include without replacement, with replacement, systematic, and sequential with minimum replacement. The procedure can apply these methods for stratified and replicated sample designs.

 

3. Exercises

Example 3.1: An experiment was conducted to study the hybrid seed production of bottle gourd (Lagenaria siceraria (Mol) Standl) Cv. Pusa hybrid-3 under open field conditions during Kharif-2005 at Indian Agricultural Research Institute, New Delhi. The main aim of the investigation was to compare natural pollination and hand pollination. The data were collected on 10 randomly selected plants from each of natural pollination and hand pollination on number of fruit set for the period of 45 days, fruit weight (kg), seed yield per plant (g) and seedling length (cm). The data obtained is as given below:

 

Group

No. of fruit

Fruit weight

Seed yield/plant

Seedling length

1

7.0

1.85

147.70

16.86

1

7.0

1.86

136.86

16.77

1

6.0

1.83

149.97

16.35

1

7.0

1.89

172.33

18.26

1

7.0

1.80

144.46

17.90

1

6.0

1.88

138.30

16.95

1

7.0

1.89

150.58

18.15

1

7.0

1.79

140.99

18.86

1

6.0

1.85

140.57

18.39

1

7.0

1.84

138.33

18.58

 

 

2

6.3

2.58

224.26

18.18

2

6.7

2.74

197.50

18.07

2

7.3

2.58

230.34

19.07

2

8.0

2.62

217.05

19.00

2

8.0

2.68

233.84

18.00

2

8.0

2.56

216.52

18.49

2

7.7

2.34

211.93

17.45

2

7.7

2.67

210.37

18.97

2

7.0

2.45

199.87

19.31

2

7.3

2.44

214.30

19.36

{Here 1 denotes natural pollination and 2 denotes the hand pollination}

 

1. Test whether the mean of the population of Seed yield/plant (g) is 200 or not.

2. Test whether the natural pollination and hand pollination under open field conditions are equally effective or are significantly different.

3. Test whether hand pollination is better alternative in comparison to natural pollination.

 

Procedure:

For performing analysis, input the data in the following format. {Here Number of fruit (45 days) is termed as nfs45, Fruit weight (kg) is termed as fw, seed yield/plant (g) is termed as syp and Seedling length (cm) is termed as sl. It may, however, be noted that one can retain the same name or can code in any other fashion}.

 

data ttest1; /*one can enter any other name for data*/

input group nfs45 fw syp sl;

cards;

. . . . .

. . . . .

. . . . .

;

 

*To answer the question number 1 use the following SAS statements

proc ttest H0=200;

var syp;

run;

*To answer the question number 2 use the following SAS statements;

proc ttest;

class group;

var nfs45 fw syp sl;

run;

 

To answer the question number 3 one has to perform the one tail t-test. The easiest way to convert a two-tailed test into a one-tailed test is take half of the p-value provided in the output of

2-tailed test output for drawing inferences. The other way is using the options sides in proc

 

statement. Here we are interested in testing whether hand pollination is better alternative in comparison to natural pollination, therefore, we may use Sides=L as

 

proc ttest sides=L;

class group;

var nfs45 fw syp sl;

run;

 

Similarly this option can also be used in one sample test and for right tail test Sides=U is used.

 

Exercise 3.2: A study was undertaken to find out whether the average grain yield of paddy of farmers using laser levelling is more than the farmers using traditional land levelling methods. For this study data on grain yield in tonne/hectare was collected from 59 farmers (33 using traditional land levelling methods and 26 using new land leveller) and is given as:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Test whether the traditional land levelling and laser levelling give equivalent yields or are significantly different.

 

Procedure:

For performing analysis, input the data in the following format. {Here traditional land levelling is termed as LL, laser levelling as LL, method of levelling as MLevel and grain yield in t/ha as gyld. It may, however, be noted that one can retain the same name or can code in any other fashion}.

 

data ttestL; /*one can enter any other name for data*/

input MLevel gyld;

 

cards;

. . . . .

. . . . .

. . . . .

;

 

*To answer the question number 1 use the following SAS statements

proc ttest data =ttestL;

var gyld;

run;

 

Exercise 3.3: The observations obtained from 15 experimental units before and after application of the treatment are the following:

 

 

 

 

1. Test whether the mean score before application of treatment is 65.

2. Test whether the application of treatments has resulted into some change in the score of the experimental units.

3. Test whether application of treatment has improved the scores.

 

Procedure:

data ttest;

input sn preapp postapp;

cards;

1 80 82

2 73 71

3 70 95

4 60 69

5 88 100

6 84 71

7 65 75

8 37 60

9 91 95

10 98 99

11 52 65

12 78 83

13 40 60

14 79 86

15 59 62

;

 

*For objective 1, use the following; PROC TTEST H0=65;

VAR PREAPP; RUN;

 

*For objective 2, use the following; PROC TTEST;

PAIRED PREAPP*POSTAPP; RUN;

 

*For objective 3, use the following; PROC TTEST sides=L;

PAIRED PREAPP*POSTAPP; RUN;

 

Exercise 3.4: In F2 population of a breeding trial on pea, out of a total of 556 seeds, the frequency of seeds of different shape and colour are: 315 rounds and yellow, 101 wrinkled and yellow, 108 round and green , 32 wrinkled and green. Test at 5% level of significance whether the different shape and colour of seeds are in proportion of 9:3:3:1 respectively.

 

Procedure:

/*rndyel=round and yellow, rndgrn=round and green, wrnkyel=wrinkled and yellow, wrnkgrn=wrinkled and green*/;

 

data peas;

input shape_color $ count;

cards; rndyel 315 rndgrn 108 wrnkyel 101 wrnkgrn 32

;

 

proc freq data=peas order=data;

weight count ;

tables shape_color / chisq testp=(0.5625 0.1875 0.1875 0.0625);

exact chisq;

run;

 

Exercise 3.5: The educational standard of adoptability of new innovations among 600 farmers are given as below:

 

 

 

 

 

Draw the inferences whether educational standard has any impact on their adoptability of innovation.

 

Procedure:

data innovation;

input edu $ adopt $ count;

cards;

Matric adopt 100

 

Matric Noadopt 50 grad adopt 60

grad Noadopt 20 illit adopt 80

illit Noadopt 290

;

proc freq order=data;

weight count ;

tables edu*adopt / chisq ;

run;

 

Exercise 3.6: An Experiment was conducted using a Randomized complete block design in 5 treatments a, b, c, d & e with three replications. The data (yield) obtained is given below:

 

 

 

 

 

 

 

1. Perform the analysis of variance of the data.

2. Test the equality of treatment means.

3. Test H0: 2T1=T2+T3, where as T1, T2, T3, T4 and T5 are treatment effects.

Procedure:

Prepare a SAS data file using

DATA Name;

INPUT REP TRT $ yield; Cards;

. . .

. . .

. . .

;

 

Print data using PROC PRINT. Perform analysis using PROC ANOVA, obtain means of treatments and obtain pairwise comparisons using least square differences, Duncan’s New Multiple range tests and Tukey’s Honest Significant difference tests. Make use of the following statements:

 

PROC Print; PROC ANOVA; Class REP TRT;

Model Yield = REP TRT; Means TRT/lsd;

Means TRT/duncan;

Means TRT/tukey; Run;

 

Perform contrast analysis using PROC GLM. Proc glm;

 

Class rep trt;

Model yld = rep trt; Means TRT/lsd; Means TRT/duncan; Means TRT/tukey

Contrast ‘1 Vs 2&3’ trt 2 -1 -1; Run;

 

Exercise 3.7: In order to select suitable tree species for Fuel, Fodder and Timber an experiment was conducted in a randomized complete block design with ten different trees and four replications. The plant height was recorded in cm. The details of the experiment are given below: Plant Height (Cms): Place – Kanpur

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Analyze the data and draw your conclusions.

 

Exercise 3.8: An experiment was conducted with 49 crop varieties (TRT) using a simple lattice design. The layout and data obtained (Yld) is as given below:

REPLICATION (REP)-I

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

REPLICATION (REP)-II

 

1(24)

11(51)

48(121)

37(85)

40(33)

10(30)

42(23)

36(58)

4(39)

41(22)

9(10)

12(48)

31(50)

35(54)

29(97)

39(67)

6(75)

30(65)

33(73)

38(30)

28(54)

15(47)

32(93)

34(44)

44(5)

26(56)

45(103)

7(85)

1. Perform the analysis of variance of the data. Also obtain Type II SS.

2. Obtain adjusted treatment means with their standard errors.

3. Test the equality of all adjusted treatment means.

4. Test whether the sum of 1 to 3 treatment means is equal to the sum of 4 to 6 treatments.

5. Estimate difference between average treatment 1 average of 2 to 4 treatment means.

6. Divide the between block sum of squares into between replication sum of squares and between blocks within replications sum of squares.

7. Assuming that the varieties are a random selection from a population, obtain the genotypic variance.

8. Analyze the data using block effects as random.

 

PROCEDURE

Prepare the DATA file. DATA Name;

INPUT REP BLK TRT yield;

Cards;

. . . .

. . . .

. . . .

;

Print data using PROC PRINT. Perform analysis of 1 to 5 objectives using PROC GLM. The statements are as follows:

Proc print; Proc glm;

Class rep blk trt;

Model yld= blk trt/ss2; Contrast ‘A’ trt 1 1 1 -1 -1 -1; Estimate ‘A’ trt 3 -1 -1 -1/divisor=3; Run;

 

The objective 6 can be achieved by another model statement. Proc glm;

Class rep blk trt;

Model yield= rep blk (rep) trt/ss2;

run;

The objective 7 can be achieved by using the another PROC statement

 

Proc Varcomp Method=type1; Class blk trt;

Model yield = blk trt/fixed = 1; Run;

 

The above obtains the variance components using Hemderson’s method. The methods of maximum likelihood, restricted maximum likelihood, minimum quadratic unbiased estimation can also be used by specifying method =ML, REML, MIVQE respectively.

Objective 8 can be achieved by using PROCMIXED.

 

Proc Mixed ratio covtest; Class blk trt;

Model yield = trt;

Random blk/s; Lsmeans trt/pdiff; Store lattice;

Run;

PROC PLM SOURCE = lattice; LSMEANS trt /pdiff lines; RUN;

 

Exercise 3.9: Analyze the data obtained through a Split-plot experiment involving the yield of 3

Irrigation (IRRIG) treatments applied to main plots and two Cultivars (CULT) applied to subplots in three Replications (REP). The layout and data (YLD) is given below:

Replication-I Replication -II Replication-III

I1 I2 I3 I1 I2 I3 I1 I2 I3

C1 (1.6) C2 (3.3)

C1 (2.6) C2 (5.1)

C1 (4.7) C2 (6.8)

C1 (3.4) C2 (4.7)

C1 (4.6) C2 (1.1)

C1 (5.5) C2 (6.6)

C1 (3.2) C2 (5.6)

C1 (5.1) C2 (6.2)

C1 (5.7) C2 (4.5)

Perform the analysis of the data. (HINT: Steps are given in text).

Remark 3.9.1: Another way proposed for analysis of split plot designs is using replication as random effect and analyse the data using PROC MIXED of SAS. For the above case, the steps for using PROC MIXED are:

PROC MIXED COVTEST; CLASS rep irrig cult;

MODEL yield = irrig cult irrig*cult / DDFM=KR; RANDOM rep rep*irrig;

LSMEANS irrig cult irrig*cult / PDIFF; STORE spd;

run;

/* An item store is a special SAS-defined binary file format used to store and restore information with a hierarchical structure*/

 

/* The PLM procedure performs post fitting statistical analyses for the contents of a SAS item store that was previously created with the STORE statement in some other SAS/STAT procedure*/

PROC PLM SOURCE = SPD;

LSMEANS irrig cult irrig*cult /pdiff lines; RUN;

 

Remark 3.9.2: In Many experimental situations, the split plot designs are conducted across environments and a pooled is required. One way of analysing data of split plot designs with two factors A and B conducted across environment is

PROC MIXED COVTEST; CLASS year rep a b;

MODEL yield = a b a*b / DDFM=KR;

/* DDFM specifies the method for computing the denominator degrees of freedom for the tests of fixed effects resulting from the MODEL*/

RANDOM year rep(year) year*a year*rep*a year*a*b; LSMEANS a b a*b / PDIFF;

STORE spd1;

run;

PROC PLM SOURCE = SPD1; LSMEANS a b a*b/pdiff lines; RUN;

 

Exercise 3.10: An agricultural field experiment was conducted in 9 treatments using 36 plots arranged in 4 complete blocks and a sample of harvested output from all the 36 plots are to be analysed blockwise by three technicians using three different operations. The data collected is given below:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1. Perform the analysis of the data considering that technicians and operations are crossed with each other and nested in the blocking factor.

2. Perform the analysis by considering the effects of technicians as negligible.

3. Perform the analysis by ignoring the effects of the operations and technicians.

 

 

 

Procedure:

Prepare the data file.

 

DATA Name;

INPUT BLK TECH OPER TRT OBS; Cards;

. . . .

 

. . . .

. . . .

;

Perform analysis of objective 1 using PROC GLM. The statements are as follows: Proc glm;

Class blk tech oper trt;

Model obs= blk tech (blk) oper(blk) trt/ss2; Lsmeans trt oper(blk)/pdiff;

Run;

 

Perform analysis of objective 2 using PROC GLM with the additional statements as follows: Proc glm;

Class blk tech oper trt;

Model obs= blk oper(blk) trt/ss2;

run;

 

Perform analysis of objective 3 using PROC GLM with the additional statements as follows: Proc glm;

Class blk tech oper trt; Model obs = blk trt/ss2; run;

 

Exercise 3.11: A greenhouse experiment on tobacco mossaic virus was conducted. The experimental unit was a single leaf. Individual plants were found to be contributing significantly to error and hence were taken as one source causing heterogeneity in the experimental material. The position of the leaf within plants was also found to be contributing significantly to the error. Therefore, the three positions of the leaves viz. top, middle and bottom were identified as levels of second factor causing heterogeneity. 7 solutions were applied to leaves of 7 plants and number of lesions produced per leaf was counted. Analyze the data of this experiment.

 

 

 

 

 

 

 

 

The figures at the intersections of the plants and leaf position are the solution numbers and the figures in the parenthesis are number of lesions produced per leaf.

 

 

Procedure:

Prepare the data file. DATA Name;

INPUT plant posi $ trt count; Cards;

. . . .

. . . .

. . . .

 

;

Perform analysis using PROC GLM. The statements are as follows:

 

Proc glm;

Class plant posi trt count;

Model count= plant posi trt/ss2; Lsmeans trt/pdiff; Run;

Exercise 3.12: The following data was collected through a pilot sample survey on Hybrid Jowar crop on yield and biometrical characters. The biometrical characters were average Plant Population (PP), average Plant Height (PH), average Number of Green Leaves (NGL) and Yield (kg/plot).

1. Obtain correlation coefficient between each pair of the variables PP, PH, NGL and yield.

2. Fit a multiple linear regression equation by taking yield as dependent variable and biometrical characters as explanatory variables. Print the matrices used in the regression computations.

3. Test the significance of the regression coefficients and also equality of regression coefficients of a) PP and PH b) PH and NGL

4. Obtain the predicted values corresponding to each observation in the data set.

5. Identify the outliers in the data set.

6. Check for the linear relationship among the biometrical characters.

7. Fit the model without intercept.

8. Perform principal component analysis.

 

25

88.44

0.9800

5.00

4.080

26

99.55

0.6450

9.60

2.830

27

63.99

0.6350

5.60

2.570

28

101.77

0.2900

8.20

7.420

29

138.66

0.7200

9.90

2.620

30

90.22

0.6300

8.40

2.000

31

76.92

1.2500

7.30

1.990

32

126.22

0.5800

6.90

1.360

33

80.36

0.6050

6.80

0.680

34

150.23

1.1900

8.80

5.360

35

56.50

0.3550

9.70

2.120

36

136.00

0.5900

10.20

4.160

37

144.50

0.6100

9.80

3.120

38

157.33

0.6050

8.80

2.070

39

91.99

0.3800

7.70

1.170

40

121.50

0.5500

7.70

3.620

41

64.50

0.3200

5.70

0.670

42

116.00

0.4550

6.80

3.050

43

77.50

0.7200

11.80

1.700

44

70.43

0.6250

10.00

1.550

45

133.77

0.5350

9.30

3.280

46

89.99

0.4900

9.80

2.690

Procedure: Prepare a data file Data mlr;

Input PP PH NGL Yield; Cards;

. . . .

. . . .

;

For obtaining correlation coefficient, Use PROC CORR; Proc Corr;

Var PP PH NGL Yield;

run;

 

For fitting of multiple linear regression equation, use PROC REG

 

 

 

Proc Reg;

Model Yield = PP PH NGL/ p r influence vif collin xpx i; Test 1: Test PP =0; Test 2: Test PH=0;

Test 3: Test NGL=0;

Test 4: Test PP-PH=0; Test 4a: Test PP=PH=0; Test 5: Test PH-NGL=0; Test 5a: Test PH=NGL=0;

 

Model Yield = PP PH NGL/noint;

run;

Proc reg;

Model Yield = PP PH NGL; Restrict intercept =0;

Run;

For diagnostic plots

Proc Reg plots(unpack)=diagnostics; Model Yield = PP PH NGL;

run;

For variable selection, one can use the following option in model statement:

Selection=stepwise sls=0.10;

For performing principal component analysis, use the following: PROC PRINCOMP;

VAR PP PH NGL YIELD;

run;

Example 3.13: An experiment was conducted at Division of Agricultural Engineering, IARI, New Delhi for studying the capacity of a grader in number of hours when used with three different speeds and two processor settings. The experiment was conducted using a factorial completely randomised design in 3 replications. The treatment combinations and data obtained on capacity of grader in hours given as below:

 

3

2

1

4

2265

3

2

2

5

2280

3

2

3

6

2278

3

3

1

7

3040

3

3

2

8

3028

3

3

3

9

3040

 

Experimenter was interested in identifying the best combination of speed and processor setting that gives maximum capacity of the grader in hours.

 

Solution: This data can be analysed as per procedure of factorial CRD and one can use the following SAS steps for performing the nalysis:

 

Data ex1a;

Input rep speed proset cgrader;

/*here rep: replication; proset: processor setting and cgrader: capacity of the grader in hours*/ Cards;

1 1 1 1852

1 1 2 1848

1 1 3 1855

. . . .

. . . .

. . . .

3 3 1 3040

3 3 2 3028

3 3 3 3040

;

Proc glm data=ex1; Class speed prost;

Model cgrader=speed post speed*post;

Lsmeans speed post speed*post/pdiff adjust=tukey lines; Run;

 

The above analysis would identify test the significance of main effects of speed and processor setting and their interaction. Through this analysis one can also identify the speed level (averaged over processor setting) {Processor Setting (averaged over speed levels)} at which the capacity of the grader is maximum. The multiple comparisons between means of combinations of speed and processor setting would help in identifying the combination at which capacity of the grader is maximum.

 

Exercise 3.14: An experiment was conducted with five levels of each of the four fertilizer treatments nitrogen, Phosphorus, Potassium and Zinc. The levels of each of the four factors and yield obtained are as given below. Fit a second order response surface design using the original data. Test the lack of fit of the model. Compute the ridge of maximum and minimum responses. Obtain predicted residual Sum of squares.

 

40

30

25

20

11.28

40

30

25

60

8.44

40

30

75

20

13.29

40

90

25

20

7.71

120

30

25

20

8.94

40

30

75

60

10.9

40

90

25

60

11.85

120

30

25

60

11.03

120

30

75

20

8.26

120

90

25

20

7.87

40

90

75

20

12.08

40

90

75

60

11.06

120

30

75

60

7.98

120

90

75

60

10.43

120

90

75

20

9.78

120

90

75

60

12.59

160

60

50

40

8.57

0

60

50

40

9.38

80

120

50

40

9.47

80

0

50

40

7.71

80

60

100

40

8.89

80

60

0

40

9.18

80

60

50

80

10.79

80

60

50

0

8.11

80

60

50

40

10.14

80

60

50

40

10.22

80

60

50

40

10.53

80

60

50

40

9.5

80

60

50

40

11.53

80

60

50

40

11.02

 

Procedure:

Prepare a data file.

/* yield at different levels of several factors */

title 'yield with factors N P K Zn';

data dose;

input n p k Zn y ; label y = "yield" ;

cards;

. . . . .

. . . . .

. . . . .

;

*Use PROC RSREG.

ods graphics on;

proc rsreg data=dose plots(unpack)=surface(3d);

model y= n p k Zn/ nocode lackfit press;

 

run;

ods graphics off; *If we do not want surface plots, then we may proc rsreg;

model y= n p k Zn/ nocode lackfit press; Ridge min max;

run;

 

Exercise 3.15: Fit a second order response surface design to the following data. Take replications as covariate.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Procedure:

Prepare a data file.

/* yield at different levels of several factors */

title 'yield with factors x1 x2';

data respcov;

input fert1 fert2 x1 x2 yield ;

cards;

. . . . .

. . . . .

. . . . .

;

/*Use PROC RSREG.*/ ODS Graphics on;

proc rsreg plots(unpack)=surface(3d);

model yield = rep fert1 fert2/ covar=1 nocode lackfit ; Ridge min max;

run;

ods graphics off;

Exercise 3.16: Following data is related to the length(in cm) of the ear-head of a wheat variety

9.3, 18.8, 10.7, 11.5, 8.2, 9.7, 10.3, 8.6, 11.3, 10.7, 11.2, 9.0, 9.8, 9.3, 10.3, 10, 10.1 9.6, 10.4. Test the data that the median length of ear-head is 9.9 cm.

 

Procedure:

This may be tested using any of the three tests for location available in Proc Univariate viz. Student’s test, the sign test, and the Wilcoxon signed rank test. All three tests produce a test statistic for the null hypothesis that the mean or median is equal to a given value 0 against the

 

two-sided alternative that the mean or median is not equal to 0. By default, PROC UNIVARIATE sets the value of 0 to zero. You can use the MU0= option in the PROC UNIVARIATE statement to specify the value of 0. If the data is from a normal population, then we can infer using t-test otherwise non-parametric tests sign test, and the Wilcoxon signed rank test may be used for drawing inferences.

 

Procedure: data npsign; input length; cards;

9.3

18.8

10.7

11.5

8.2

9.7

10.3

8.6

11.3

10.7

11.2

9.0

9.8

9.3

10.3

10.0

10.1

9.6

10.4

;

PROC UNIVARIATE DATA=npsign MU0=9.9; VAR length;

HISTOGRAM / NOPLOT ;

RUN;

QUIT;

 

Exercise 3.17: An experiment was conducted with 21 animals to determine if the four different feeds have the same distribution of Weight gains on experimental animals. The feeds 1, 3 and 4 were given to 5 randomly selected animals and feed 2 was given to 6 randomly selected animals. The data obtained is presented in the following table.

 

 

 

 

 

 

 

Procedure:

 

data np;

input feed wt;

datalines;

1

3.35

1

3.80

1

3.55

1

3.36

1

3.81

2

3.79

2

4.10

2

4.11

2

3.95

2

4.25

2

4.40

3

4.00

3

4.50

3

4.51

3

4.75

3

5.00

4

3.57

4

3.82

4

4.09

4

3.96

4

3.82

;

 

PROC NPAR1WAY DATA=np WILCOXON; /*for performing Kruskal-Walis test*/;

VAR wt; CLASS feed; RUN;

 

Example 3.18: Finney (1971) gave a data representing the effect of a series of doses of carotene (an insecticide) when sprayed on Macrosiphoniella sanborni (some obscure insects). The Table below contains the concentration, the number of insects tested at each dose, the proportion dying and the probit transformation (probit+5) of each of the observed proportions.

 

Concentratio n (mg/1)

No. of insects (n)

No. of affected (r)

%kill (P)

Log concentration (x)

Empirical probit

10.2

50

44

88

1.01

6.18

7.7

49

42

86

0.89

6.08

5.1

46

24

52

0.71

5.05

3.8

48

16

33

0.58

4.56

2.6

50

6

12

0.41

3.82

0

49

0

0

-

-

 

Perform the probit analysis on the above data.

Procedure data probit; input con n r; datalines;

10.2 50 44

7.7 49 42

5.1 46 24

3.8 48 16

2.6 50 6

0 49 0

;

ods html;

Proc Probit log10 ;

Model r/n=con/lackfit inversecl; title ('output of probit analysis'); run;

ods html close;

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Model Information

Data Set WORK.PROBIT Events Variable r Trials Variable n Number of Observations 5

Number of Events 132

Number of Trials 243

Name of Distribution Normal

Log Likelihood -120.0516414

Number of Observations Read

6

Number of Observations Used

5

Number of Events

132

Number of Trials

243

Algorithm converged.

 

Goodness-of-Fit Tests

Statistic

Value

DF

Pr > ChiSq

Pearson Chi-Square

1.7289

3

0.6305

L.R. Chi-Square

1.7390

3

0.6283

 

Response-Covariate Profile

Response Levels 2

Number of Covariate Values 5

 

Since the chi-square is small (p > 0.1000), fiducial limits will be calculated using a t value of 1.96

 

Type III Analysis of Effects

Wald

Effect DF

Chi-Square Pr > ChiSq

Log10(con) 1 77.5920 <.0001

 

Analysis of Parameter Estimates

 

Parameter

 

DF

 

Estimate

Standard

Error

95% Confidence

Limits

 

Chi-Square

 

Pr > ChiSq

Intercept

1

-2.8875

0.3501

-3.5737 -2.2012

68.01

<.0001

Log10(con)

1

4.2132

0.4783

3.2757 5.1507

77.59

<.0001

 

Probit Model in Terms of

Tolerance Distribution

MU SIGMA

0.68533786 0.23734947

 

Estimated Covariance Matrix for

Tolerance Parameters

MU SIGMA

 

MU

0.000488

-0.000063

SIGMA

-0.000063

0.000726

 

Probit Analysis on Log10(con) Probability Log10(con) 95% Fiducial Limits

0.01

0.13318

-0.03783

0.24452

0.02

0.19788

0.04453

0.29830

0.03

0.23893

0.09668

0.33253

0.04

0.26981

0.13584

0.35834

0.05

0.29493

0.16764

0.37940

0.06

0.31631

0.19466

0.39737

0.07

0.33506

0.21832

0.41316

0.08

0.35184

0.23946

0.42733

0.09

0.36711

0.25866

0.44026

0.10

0.38116

0.27631

0.45218

0.15

0.43934

0.34898

0.50192

0.20

0.48558

0.40618

0.54202

0.25

0.52525

0.45467

0.57700

0.30

0.56087

0.49759

0.60904

0.35

0.59388

0.53666

0.63942

0.40

0.62521

0.57295

0.66905

0.45

0.65551

0.60716

0.69861

0.50

0.68534

0.63983

0.72870

0.55

0.71516

0.67142

0.75986

0.60

0.74547

0.70240

0.79265

0.65

0.77679

0.73330

0.82766

0.70

0.80980

0.76480

0.86563

0.75

0.84543

0.79777

0.90761

0.80

0.88510

0.83352

0.95533

0.85

0.93133

0.87427

1.01188

0.90

0.98951

0.92456

1.08401

0.91

1.00357

0.93658

1.10155

0.92

1.01883

0.94960

1.12065

0.93

1.03562

0.96387

1.14170

0.94

1.05436

0.97976

1.16526

0.95

1.07574

0.99783

1.19218

0.96

1.10086

1.01898

1.22388

0.97

1.13174

1.04490

1.26294

0.98

1.17279

1.07924

1.31498

0.99

1.23750

1.13315

1.39721

 

 

 

Probit Analysis on con

Probability con 95% Fiducial Limits

 

0.01

1.35888

0.91657

1.75599

0.02

1.57718

1.10799

1.98745

0.03

1.73353

1.24935

2.15043

0.04

1.86129

1.36724

2.28215

0.05

1.97212

1.47110

2.39553

0.06

2.07163

1.56554

2.49671

0.07

2.16302

1.65317

2.58917

0.08

2.24825

1.73565

2.67506

0.09

2.32868

1.81410

2.75586

0.10

2.40526

1.88932

2.83257

0.15

2.75005

2.23349

3.17629

0.20

3.05900

2.54788

3.48353

0.25

3.35157

2.84884

3.77571

0.30

3.63808

3.14478

4.06477

0.35

3.92538

3.44084

4.35935

0.40

4.21897

3.74068

4.66710

0.45

4.52389

4.04724

4.99582

0.50

4.84549

4.36343

5.35423

0.55

5.18995

4.69265

5.75260

0.60

5.56506

5.03963

6.20374

0.65

5.98127

5.41132

6.72450

0.70

6.45363

5.81830

7.33883

0.75

7.00531

6.27722

8.08377

0.80

7.67532

6.81590

9.02252

0.85

8.53758

7.48633

10.27723

0.90

9.76143

8.40534

12.13411

0.91

10.08243

8.64132

12.63428

0.92

10.44313

8.90434

13.20233

0.93

10.85466

9.20181

13.85792

0.94

11.33346

9.54469

14.63036

0.95

11.90537

9.95006

15.56609

0.96

12.61427

10.44674

16.74479

0.97

13.54388

11.08927

18.32046

0.98

14.88655

12.00168

20.65263

0.99

17.27807

13.58779

24.95808

 

Interpretation: The goodness-of-fit tests (p-values = 0.6305, 0.6283) suggest that the distribution and the model fits the data adequately. In this case, the fitting is done on normal equivalent deviate only without adding 5. Therefore, log LD50 or lof ED50 corresponds to the value of Probit=0. Log LD50 is obtained as 0.685338. Therefore, the stress level at which the

50% of the insects will be killed is (100.685338=4.845 mg/l). Similarly the stress level at which

65% of the insects will be killed is (100.776793 = 5.981 mg/l). Although both values are given in the table above.

 

4. Discussion

We have initiated a link “Analysis of Data” at Design Resources Server (www.iasri.res.in/design) to provide steps of analysis of data generated from designed experiments by using statistical packages like SAS, SPSS, MINITAB, and SYSTAT, MS-

 

EXCEL etc. For details and live examples one may refer to the link Analysis of data at http://www.iasri.res.in/design/Analysis%20of%20data/Analysis%20of%20Data.html.

 

How to see SAS/STAT Examples?

One can learn from the examples available at http://support.sas.com/rnd/app/examples/STATexamples.html

 

How to use HELP?

Help  SAS help and Documentation  Contents  Learning to use SAS  Sample SAS Programs  SAS/STAT …

 

5. Strengthening Statistical Computing for NARS

NAIP Consortium on Strengthening Statistical Computing for NARS (www.iasri.res.in/sscnars)

targets at providing

‐ research guidance in statistical computing and computational statistics and creating sound and healthy statistical computing environment

‐ Providing advanced, versatile, and innovative and state-of the art high end statistical packages to enable them to draw meaningful and valid inferences from their research.

 

The efforts also involve designing of intelligent algorithms for implementing statistical techniques particularly for analysing massive data sets, simulation, bootstrap, etc.

 

The objectives of the consortium are:

‐ To strengthen the high end statistical computing environment for the scientists in NARS;

‐ To organize customized training programmes and also to develop training modules and

manuals for the trainers at various hubs; and

‐ To sensitize the scientists in NARS with the statistical computing capabilities available for enhancing their computing and research analytics skills.

 

This consortium has provided the platform for closer interactions among all NARS

organizations.

 

Capacity Building

For capacity building of researchers in the usage of high end statistical computing facility and statistical techniques,

‐ 209 trainers have been trained through 30 working days training programmes;

‐ 2166 researchers have been trained through 104 training programmes of one week duration

each in the usage.

The capacity building efforts have paved the way for publishing research papers in the high impact factor journals.

 

Indian NARS Statistical Computing Portal

For providing service oriented computing, developed and established Indian NARS Statistical Computing portal, which is available to NARS users through IP authentication at http://stat.iasri.res.in/sscnarsportal. Any researcher from Indian NARS may obtain User name and password from Nodal Officers of their respective NARS organizations, list available at

 

www.iasri.res.in/sscnars. It is a paradigm of computing techniques that operate on software-as- a-service). There is no need of installation of statistical package at client side. Following 24 different modules of analysis of data are available on this portal, which have been classified into four broad categories as

Basic Statistics

• Descriptive Statistics

• Univariate Distribution Fitting

• Test of Significance based on t-test

• Test of Significance based on Chi-square test

• Correlation Analysis

• Regression Analysis

Designs of Experiments

• Completely randomized designs

• Block Designs (includes both complete and incomplete block designs)

• Combined Block Designs

• Augmented Block Designs

• Resolvable Block Designs

• Nested Block Designs

• Row-Column Designs

• Cross Over Designs

• Split Plot Designs

• Split-Split-Plot Designs

• Split Factorial (main A, sub B  C) designs

• Split Factorial (main AB, sub CD) designs

• Strip Plot Designs

• Response Surface Designs

Multivariate Analysis

• Principal Component Analysis

• Linear Discriminant Analysis

Statistical Genetics

• Estimation of Heritability from half- sib data

• Estimation of variance-Covariance matrix from Block Designs

 

 

The above modules can be used by uploading *.xlsx, *.csv and *.txt files and results can be saved as *.RTF or *.pdf files. This has helped them in analyzing their data in an efficient manner without losing any time.

 

 

Requirements of Excel Files during analysis over Indian NARS Statistical Computing

Portal

1. Excel file must have the .xls, .xlsx, .csv or .txt extensions

2. This system will only consider the first sheet of the excel file which has name appearing first in lexicographic order. It will not analyze the data which lies in subsequent sheets in excel file.

3. Do not put period (.) or Zero (0) to display missing values in the treatment. It will not

consider as missing. Please leave the missing observations as blank cells.

4. If you are getting some wrong analysis then kindly check your excel file. Go to First Column, first cell and then press Ctrl+Shift+End. It will select all the filled rows and columns. If it selects some missing rows and columns then kindly delete those rows and columns otherwise it will give wrong analysis result.

5. Do not use special characters in the variable/column names. Also variable names should not start with spaces.

6. Do not use any formatting to the Excel sheet including formats or expressions to the cell values. It should be data value.

7. If the First row cells has been merged then it will not detect as Column/Variable names.

8. If any rows or columns are hidden then it will be displayed during the analysis.

Basic Statistics

 

9. Descriptive Statistics: The data file should contain at least one quantitative analysis variable.

10. Univariate Distribution Fitting: The data file should contain at least one quantitative numeric variable.

11. Test of Significance based on t-distribution: The data file should contain at least one quantitative variable name and one classificatory variable.

12. Chi-Square Test: The data file should contain at least one categorical variable and weights or frequency counts variable if frequencies are entered in a separate column. Data may also have classificatory in it.

13. Correlation: The data file should contain at least two quantitative variables.

14. Regression Analysis: The data file should contain at least one Dependent and one

Independent variable.

Design of Experiments

15. Unblock Design: Prepare a data file containing one variable to describe the Treatment details and at least one response/ dependent variable in the experimental data to be analyzed. Also, the treatment details may be coded or may have actual names (i.e. data values, for variable describing treatment column may be in numeric or character). The maximum length of treatment value is 20 characters. The variables can be entered in any order.

16. Block Design: Prepare a data file containing two variables to describe the block and treatment details. There should be at least one response/ dependent variable in the experimental data to be analyzed. Also, the block/treatment details may be coded or may have actual names (i.e. data values, for variables describing block and treatment column may be in numeric or character). The maximum length of treatment value is 20 character. The variables can be entered in any order. (These conditions are applicable to other similar experimental designs also)

17. Combined Block Design: The data file should contain three variables to describe

Environment, Block, Treatment variables and at least one Dependent variable.

18. Augmented Block Design: The data file should contain two variables to describe Block

& Treatment variables and at least one Dependent variable. At present, Portal supports only numeric treatment and block variables for augmented designs. An augmented block design involves two sets of treatments known as check or control and test treatments. The treatments should be numbered in such a fashion that the check or control treatments are numbered first followed by test treatments. For example, if there are 4 control treatments and 8 test treatments, then the control treatments are renumbered as 1, 2, 3, 4 and tests are renumbered as 5, 6, 7, 8, 9, 10, 11, 12.

19. Resolvable Block Design: The data file should contain three variables to describe the

Replication, Block, Treatment variables and at least one Dependent/ response variable.

20. Nested Block Design: The data file should contain three variables to describe Block, SubBlock, Treatment variables and at least one Dependent variable.

21. Row Column Design: The data file should contain three variables to describe Row, Column, Treatment variables and at least one Dependent variable.

22. Crossover Design: Create a data file with at least 5 variables, one for units, one for periods, one treatments, one for residual, and one for the dependent or analysis variable. For performing analysis using the portal, please rearrange the data in the following order: animal numbers as units; periods can be coded as 1, 2, 3, and so on, treatments as

 

alphabets or numbers (coding could be done as follows: for every first period the number one has assigned (fixed) and for other periods code 1 to 3 are given according to the treatment received by the unit in the previous period) and residual effect as residual. It may, however, be noted that one can retain the same name or can code in any other fashion. A carry-over or residual term has the special property as a factor, or class variate, of having no level in the first period because the treatment in the first period is not affected by any residual or carry over effect of any treatment. When we consider the residual or carryover effect in practice the fact that carry-over or residual effects will be adjusted for period effects (by default all effects are adjusted for all others in these analysis). As a consequence, any level can be assigned to the residual variate in the first period, provided the same level is always used. An adjustment for periods then removes this part of the residual term. (For details a reference may made to Jones, B. and Kenward,M.G. 2003. Design and Analysis of Cross Over Trials. Chapman and Hall/CRC. New York . Pp: 212)

23. Split Plot Design: The data file should contain three variables to describe Replication, Main Plot, Sub Plot variables and at least one Dependent variable.

24. Split Split Plot Design: The data file should contain four variables to describe Replication, Main Plot, Sub Plot, and Sub-Sub Plot Treatment variables and at least one Dependent variable.

25. Split Factorial (Main A, Sub B×C) Plot Design The data file should contain four variables to describe Replication, Main Plot, Sub Plot(1){levels of factor 1 in sub plot} , and Sub Plot(2) ){levels of factor 21 in sub plot} Treatment variables and at least one Dependent variable.

26. Split Factorial (Main A×B, Sub C×D) Plot Design: Create a data file with at least 6 variables, one for block or replication, one for main plot- treatment factor 1, one main plot- treatment factor 2, one for subplot- treatment factor 1, one for subplot- treatment factor 2 and at least one for the dependent or analysis variable. If the data on more than one dependent variable is collected in the same experiment, the data on all variables may be entered in additional columns. One may give actual levels used for different factors applied in main plot-treatment factor 1, main plot- treatment factor 2, subplot- treatment factor 1 and subplot- treatment factor 2. Please remember that there should not be any space between a single data value. Main plot- treatment factor 1, main plot- treatment factor 2, subplot- treatment factor 1, subplot- treatment factor 2 treatments and block numbers may be coded as 1, 2, 3 and so on. One can have character values also.

27. Strip Plot Design: The data file should contain at least 4 variables to describe Replication, Horizontal Strip, Vertical Strip variables and at least one Dependent variable.

28. Response Surface Design: The data file should contain at least one treatment factor variable and at least one dependent variable

Multivariate Analysis

29. Principal Component Analysis: The data file should contain at least one quantitative analysis variable.

30. Discriminant Analysis: The data file should contain at least one quantitative analysis variable and a classificatory variable.

Statistical Genetics

 

31. Genetic Variance Covariance: Create a data file with at least 4 variables, one for blocking variable, one for treatments and at least two analysis variable.

32. Heritability Estimation from Half-Sib Data: The data file should contain at least one quantitative analysis variable and a classificatory variable.

 

Other IP Authenticated Services

Following can also be accessed through IP authenticated networks:

 Web Report Studio: http://stat.iasri.res.in/sscnarswebreportstudio

 BI DashBoard: http://stat.iasri.res.in/sscnarsbidashboard

 Web OLAP Viewer: http://sas.iasri.res.in:8080/sscnarswebolapviewer

 E-Miner 6.1: http://sas.iasri.res.in:6401/AnalyticsPlatform

 E-Miner 7.1: http://stat.iasri.res.in/SASEnterpriseMinerJWS/Status

 

Accessing SAS E-Miner through URL (IP Authenticated Services)

For Accessing E-miner 6.1 and 7.1 through URLs, following ports should be open

Server

Ports

1) Metadata server

8561

2) Object spawner

8581

3) Table Server

2171

4) Remote Server

5091

5) SAS App. Olap Server

5451

6) SAS Deployment Tester Server

10021

7) Analytics Platform Server

6411

8) Framework Server

22031

However, if you are accessing only E-miner 6.1, then following port need not be opened.

Framework Server 22031

 

Steps for accessing SAS Enterprise Miner 6.1 and SAS Enterprise Miner 7.1 separately

SAS Enterprise Miner 6.1

 

Pre-requisite:

‐ JRE 1.5 Update 15

‐ If Firewall and proxy has been implemented then kindly open following ports:

 

 

 

 

 

 

 

 

 

 

 

 

 

Steps to be followed:

‐ If you have installed multiple Java Runtime Environment then

 

Go to Control Panel  Java Java tab  View  Keep check on JRE 1.5.0_15 and

Uncheck all others

‐ Check the entry of the sas.iasri.res.in in the host file, if not then open host file C:\Windows\System32\drivers\etc and edit the host file by entering the IP as shown below or specify the internal/external IP given by IASRI. Internal IP is to be specified only at IASRI, New Delhi. All other NARS organizations should specify external IP only which is: 203.197.217.209 sas.iasri.res.in sas as shown below

‐

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‐ Now Go to URL: http://sas.iasri.res.in:6401/AnalyticsPlatform

‐ Click on Launch and then Run

 

SAS Enterprise Miner 7.1

Pre-requisite:

‐ JRE 1.6 Update 16 or higher

‐ If Firewall and/or proxy has been implemented then kindly open the following ports:

 

 

 

 

 

 

 

 

 

 

 

Steps to be followed:

‐ If you have installed multiple Java Runtime Environment then

Go to Control Panel  Java Java tab  View  Keep check on JRE 1.6.0_16 or

higher available version and Uncheck all other

 Check the entry of the stat.iasri.res.in in the host file, if not then open host file C:\Windows\System32\drivers\etc and edit the host file by entering the IP as shown below or specify the internal/external IP given by IASRI, New Delhi. Internal IP is to be

 

specified only at IASRI, New Delhi. All other NARS organizations should specify external

IP only which is: 14.139.56.156 stat.iasri.res.in stat (earlier 203.197.217.221 stat.iasri.res.in stat) as shown below stat.iasri.res.in stat as shown below

 

 

‐ Now Go to URL: http://stat.iasri.res.in/SASEnterpriseMinerJWS/Status

‐ Click on Launch and then Run

 

Please note: You cannot run both E-Miner 6.1 and E-Miner 7.1 together. If you want to run JMP 6.1 then JAVA 1.5.0_15 should be available and for running JMP 7.1, JAVA version 1.6 onwards should be available on your system.

 

Indian NARS Statistical Computing Portal and other IP authenticated services are best viewed in

Internet Explorer 6 to 8 and Firefox 2.0.0.11 and 3.0.6

 

 

 

Macros Developed

Macros have been developed for some commonly used statistical analysis and made available at

Project Website www.iasri.res.in/sscnars. Following macros have been developed:

1. Analysis of data from Augmented Block designs http://www.iasri.res.in/sscnars/augblkdsgn.aspx

2. Analysis of data from Split Factorial ( main A, Sub B  C) designs http://www.iasri.res.in/sscnars/spltfctdsgn.aspx

3. Analysis of data from Split Factorial (Main AB, Sub C) designs http://www.iasri.res.in/sscnars/spltfctdsgnm2s1.aspx

4. Analysis of data from Split Factorial ( main AB, Sub C  D) designs http://www.iasri.res.in/sscnars/spltfactm2s2.aspx

5. Analysis of data from Split Split Plot designs http://www.iasri.res.in/sscnars/spltpltdsgn.aspx

6. Analysis of data from Strip Plot designs

http://www.iasri.res.in/sscnars/StripPlot.aspx

7. Analysis of data from Strip-Split Plot designs

 

http://www.iasri.res.in/sscnars/stripsplit.aspx.

8. Econometric Analysis ((diversity indices, instability index, compound growth rate, Garret scoring technique and Demand analysis using LA-AIDS model) and available at http://www.iasri.res.in/sscnars/ecoanlysis.aspx

9. Estimation of heritability along with its standard error from half sib data

http://www.iasri.res.in/sscnars/heritability.aspx

10. Generation of Polycross designs http://www.iasri.res.in/sscnars/polycrossdesign.aspx

11. Generation of TFNBCB designs http://www.iasri.res.in/sscnars/TFNBCBdesigns.aspx

 

How to see updated version of reference manual?

Reference manual is updated regularly and updated version may be downloaded from

http://www.iasri.res.in/sscnars/contentmain.htm

 

How to Renew License Files for SAS 9.2M2?

1. Go to http://stat.iasri.res.in/sscnarsportal/public

2. Click on SAS License Downloads 2011-12. It will redirect to New Page. It will start the Download of the SAS_Licenses11-12.zip. If it does not start automatically, then it would show Yellow Bar below the URL bar. Click on the Yellow Bar and Select Download File. Dialog box showing Open/Save/Cancel would appear. Click on Save and Browse the desired Location for saving the file.

3. Click on Portal Page link which is on top of the Page to go back to the main page.

4. Click on How to apply License Files?. Again it will redirect to the New Page and will start the Download Renew_the_licenses_for_SAS92_JMP8_JMPGenomics4.doc If it does not start automatically, then it would show Yellow Bar below the URL bar. Click on the Yellow Bar and Select Download File. Dialog box showing Open/Save/Cancel would appear. Click on Save and Browse the desired Location for saving the file.

You can also follow the following links for renewal of SAS Licenses:

http://support.sas.com/kb/31/187.html

 

Following link is only for Windows 7 and Windows Vista:

http://support.sas.com/kb/31/290.html

 

 

 

SAS 9.3

In SAS 9.3, the default destination in the SAS windowing environment is HTML, and ODS Graphics is enabled by default. These new defaults have several advantages. Graphs are integrated with tables, and all output is displayed in the same HTML file using a new style. This new style, HTML Blue, is an all-color style that is designed to integrate tables and modern statistical graphics. The default settings in the Results tab are as follows:

 The Create listing check box is not selected, so LISTING output is not created.

 The Create HTML check box is selected, so HTML output is created.

 The Use WORK folder check box is selected, so both HTML and graph image files are saved in the WORK folder (and not your current directory).

 The default style, HTMLBlue, is selected from the Style drop-down list.

 

 The Use ODS Graphics check box is selected, so ODS Graphics is enabled.

 Internal browser is selected so results are viewed in an internal SAS browser

 

We can view and modify the default settings by selecting ToolsOptionsPreferencesResult Tab from the menu at the top of the SAS window usually known as TOPR pronounced "topper". Snap shot is as under.

 

 

 

 To get SAS listing instead of HTML, Select check box Create listing option and deselect

Create HTML check box.

 Once HTML checkbox is deselected "Use work folder " get deselected automatically.

 Select View results as they are generated , if ODS Graphics is not required as default output. In many cases, graphs are an integral part of a data analysis. If we do not need graphics, ODS Graphics should be disabled, which will improve the performance of our program in terms of time and memory. One can disable and re-enable ODS Graphics in our SAS programs with the ODS GRAPHICS OFF and ODS GRAPHICS ON statements.

 

References

Littel, R.C., Freund, R.J. and Spector, P.C. (1991). SAS System for Linear Models, Third

Edition. SAS Institute Inc.

Searle, S.R. (1971). Linear Models. John Wiley & Sons, New York.

Searle, S.R., Casella, G and McCulloch, C.E. (1992). Analysis of Variance Components. John

Wiley & Sons, New York.

 

Impressum

Tag der Veröffentlichung: 04.02.2015

Alle Rechte vorbehalten