NCSS 2000 Number Cruncher

Comprehensive, Easy to use, Statistical Software

NCSS 2000 is a comprehensive, easyto-use, statistical program. It is Comprehensive and accurate. Includes over 200 procedures and graphics. Easy to learn and use. Windows 9x/NT compatible. Imports/ exports major spreadsheet, database, and statistical file formats. Year 2000 Compliant . Output mixes text and graphics and is easily transferred to word processors. Processes large data files (over 1000 variables 200,000 rows). Trial Copyand Available
You can try out the copy of NCSS 2000 on the enclosed CD or download it from our website. This trial copy allows the analysis of up to one hundred rows of data. It is good for 30 days from the time you install the software.

Built-In Word Processor

All reports are displayed using our built-in word processor. You can quickly view, edit, save, and print your output. Reports are stored in the rich-text format that can be read by most word processors, so you can easily save NCSS reports for further use in Word . The output document mixes text and graphs together. The text portions of the reports are formatted using tabs (not spaces), so they are easily reformatted.

Built-In Spreadsheet
Entering data is simple using our Microsoft Excel compatible spreadsheet. We import (and export) Lotus , Excel , dBase , Access , SAS , SPSS , Paradox , and ASCII formats. So if your data is in another format, chances are you can easily import it into NCSS 2000.

System Requirements
Runs on Windows 9x, NT, and compatible computers with at least 8 megs of RAM and 30 megs of hard disk space.
Box Plots

Height vs Weight
80.0

Normal Probability Plot of SepalLength

Iris 1 3 2

Histogram with Everything On It

60.0

80.0

70.0

250.0

45.0

SepalLength

200.0

SepalLength
50.0 57.5 65.0 72.5 80.0

Count

60.0

Weight

30.0

15.0

50.0

150.0

0.0

40.0

50.0

100.0

40.0

50.0

60.0

70.0

80.0

40.0

50.0 -3.0

60.0

70.0

Iris

Height

-1.5

0.0

1.5

3.0

Normal Distribution

SepalLength

Analysis of Variance and T-Tests Analysis of Covariance Analysis of Variance Barlett Variance Test Factorial Design Analysis Friedman Test Geiser-Greenhouse Correction General Linear Models Mann-Whitney Test MANOVA Multiple Comparison Tests One-Sample T-Tests One-Way ANOVA Paired T-Tests Power Calculations Repeated Measures ANOVA Two-Sample T-Tests Time Series Analysis Wilcoxon Test ARIMA / Box - Jenkins Decomposition Exponential Smoothing Harmonic Analysis Holt - Winters Seasonal Analysis Spectral Analysis Trend Analysis

Plots and Graphs Bar Charts Box Plots Contour Plot Dot Plots Error Bar Charts Histograms Percentile Plots Pie Charts Probability Plots ROC Curves Scatter Plots Scatter Plot Matrix Surface Plots Violin Plots Designs Experimental Balanced Inc. Block Box-Behnken Central Composite D-Optimal Designs Fractional Factorial Latin Squares Placket-Burman Response Surface Screening Taguchi

Regression / Correlation All-Possible Search Canonical Correlation Correlation Matrices Kendalls Tau Correlation Logistic Regression Multiple Regression Nonlinear Regression PC Regression Proportional-Hazards Response-Surface Ridge Regression Robust Regression Stepwise Regression Spearman Correlation Variable Selection Quality Control Xbar-R Chart C, P, NP, U Charts Capability Analysis Cusum, EWMA Chart Individuals Chart Moving Average Chart Pareto Chart R & R Studies

NCSS Statistical Software 329 North 1000 East Kaysville, Utah 84037 Internet (download free demo version): http://www.ncss.com Email: sales@ncss.com Toll Free: (800) 898-6109 Tel: (801) 546-0445 Fax: (801) 546-3907

Survival / Reliability Accelerated Life Tests Censoring - All Types Exponential Fitting Extreme-Value Fitting Hazard Rates Kaplan-Meier Lognormal Fitting Log-Rank Tests Probit Analysis Proportional-Hazards Survival Distributions Weibull Analysis Weibull Regression Multivariate Analysis Clustering - Kmeans Clustering - Hierarchical Correspondence Analysis Discriminant Analysis Factor Analysis Item Analysis Item Response Analysis Loglinear Models MANOVA Multi-Way Tables Multidimensional Scaling Principal Components

Curve Fitting Built-In Models Model Searching Nonlinear Regression Ratio of Polynomials User-Specified Models Miscellaneous Appraisal Ratio Studies Area Under Curve Chi-Square Test Confidence Limits Cross Tabulation Data Screening Fishers Exact Test Frequency Distributions Mantel-Haenszel Test Nonparametric Tests Normality Tests Probability Calculator Proportion Tests Tables of Means, Etc. Trimmed Means Univariate Statistics

Charts and Graphs

NCSS provides and complete set of statistical charts and graphs. You control the labels, lines, and colors of the graphs. The graphs may be copied to your favorite word processor or graphics editing package.
Sample Plots
Box Plots
80.0

Contour Plot of Test3

100.0

SepalLength

SepalLength

Test2

50.0

Test3 40.0 50.0 60.0 70.0 80.0 90.0

Dot Plots
80.0 40.0 50.0 60.0 70.0

70.0

50.0

60.0

40.0

0.0

25.0 30.0

75.0

47.5

65.0

82.5

100.0

Iris

Test1

Iris

Grid Plot of PetalLength

45.0

22.5

15.0

PetalLength A 3.5 B 10.5 C 17.5 D 24.5 E 31.5 F 38.5 G 45.5 H 52.5 I 59.5 J 66.5

Histogram with Everything On It

Density Trace at 20% Data Range
60.0

37.5

SepalWidth

45.0

Count

30.0

Count

30.0

0.0 40.0
50.0 60.0 70.0 80.0

15.0

30.0

45.0

60.0

15.0

40.0

50.0

60.0

70.0

80.0
SepalLength

SepalLength

0.0

50.0

60.0

70.0

80.0

40.0

SepalLength

Height vs Weight
Weibull Probability Plot of FailTime
80.0

1000

70.0

FailTime

SepalLength

60.0

50.0

10

40.0

50.0 50.0

100.0
0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90

150.0

100

Weight

200.0

250.0

Normal Probability Plot of SepalLength

-3.0

-1.5

0.0

1.5

3.0

57.5

65.0

72.5

80.0

Normal Distribution

Weibull Distribution

Height

Violin Plots at 30% Data

80.0
250.0

Height vs Weight
Height vs Weight
250.0

60.0

Amount

40.0

200.0

Weight

20.0

150.0

Weight
57.5 65.0 72.5 80.0

0.0

100.0

50.0

50.0

50.0 50.0

100.0

150.0

200.0

SepalLength

SepalWidth

Variables

PetalLength

PetalWidth

57.5

65.0

72.5

80.0

Height

Height

Box plot Chi-square probability plot Density trace Dot plot Exponential probability plot Frequency polygon Gamma probability plot

Grid plot Half normal probability plot Histogram Normal probability plot Percentile plot Scatter plot Scatter plot matrix

Sunflower plot Uniform probability plot Violin plot Weibull probability plot

Descriptive Statistics and Cross Tabulation

Several procedures are available for tabulating and summarizing your data. The frequency table procedure provides tabulation of single variables. The cross tabulation procedure provides tabulation of two variables into two-way tables. The descriptive tables procedure computes summary statistics (means, median, standard deviations, etc.) according to up to eight breakdown variables. All of these procedures provide numeric and graphic reports. A specialized appraisal ratio module provides reports for mass appraisal.
Frequency Table Report Row Percentages Section Values Variables 0 Garage Size 4.7 Fireplaces 26.0 Brick Ratio 34.0 Total 21.6

0.5 0.0 0.0 31.3 10.4

1 65.3 52.0 34.7 50.7

2 28.7 22.0 0.0 16.9

3 1.3 0.0 0.0 0.4

Total 100.0 100.0 100.0 100.0

Column Percentages Section Values Variables 0 0.5 Garage Size 7.2 0.0 Fireplaces 40.2 0.0 Brick Ratio 52.6 100.0 Total 100.0 100.0
Row Pct of Values by Variables
80.0

1 43.0 34.2 22.8 100.0

2 56.6 43.4 0.0 100.0

3 100.0 0.0 0.0 100.0

Total 33.3 33.3 33.3 100.0

Column Pct of Values by Variables

55.0

Variables Garage Size Fireplaces Brick Ratio

Column Pct

120.0

85.0

Variables Garage Size Fireplaces Brick Ratio

Row Pct

30.0

50.0

5.0

15.0

-20.0 0 0.5 1 2 3

-20.0 0 0.5 1 2 3

Values

Values

Cross Tabulation Armitage test for trend Break Variables (up to five) Chi-square test Contingency analysis Continuous variables allowed Cramers V Expected values Fishers exact test Gamma Kendalls tau - B & C Lambda A & B McNemar Test Percentages Phi coefficient Text variables allowed Tschuprows T Value labels

Descriptive Statistics Average absolute deviation Coef. of dispersion - COD Coefficient of variation Confidence limits Density trace Fishers g1 and g2 Geometric mean Harmonic mean Histogram Kurtosis measures Means, medians, modes Mininmum and maximum Normal probability plot Normality tests Percentiles Price-related diff. - PRD Skewness measures Standard deviation / error Stem-leaf plot Trimmed mean

Descriptive Tables Numeric break variables One-way and Two-way tables Plots of tabulated statistics Tables of medians, means, standard deviations, counts, COVs, and CODs Text break variables Value labels Frequency Tables Chi-square test Counts Cumulative statistics Frequency distribution Multinomial test Percents

Curve Fitting
Curve fitting refers to nonlinear-regression techniques for fitting curved lines to X-Y data. You can select a standard model from the large list of precoded models or you can enter your own. If you do not have a specific model in mind, the program can search through hundreds of possible models looking for the one that best fits your data. The program will fit models with up to four independent variables.
Model Estimation Section Parameter Name A B Parameter Estimate 0.3901402 0.101633 Asymptotic Standard Error 5.033759E-03 1.336168E-02 Lower 95% C.L. 0.3799816 7.466801E-02 Upper 95% C.L. 0.4002987 0.1285979

Dependent Y Independent X Model Y = A+(0.49-A)*EXP(-B*(X-8)) R-Squared 0.873375 Iterations 13 Estimated Model (.3901402)+(0.49-(.3901402))*EXP(-(.101633)*((X)-8)) Predicted Values and Residuals Section Row No. 1 2 3 4 5 Predicted Value 0.49 0.49 0.4716319 0.4716319 0.4716319 Lower 95.0% Value 0.4679772 0.4679772 0.4494232 0.4494232 0.4494232 Upper 95.0% Value 0.5120228 0.5120228 0.4938406 0.4938406 0.4938406

X 8 8 10 10 10

Y 0.49 0.49 0.48 0.47 0.48

Residual 0 0 0.0083681 -0.0016319 0.0083681

Y = PolyRatio(X, 3, 4)

Plot of Y = A+(0.49-A)*EXP(-B*(X-8))
70.0 5.0 15.0 25.0 35.0 45.0 10.0 0.0 25.0 40.0 0.6

0.5

0.4

0.4

55.0

0.5

25.0

50.0

75.0

100.0

Bleasdale-Nelder Model Double-Exponential Model Exponential Model Farazdaghi Model Gompertz Models Goodness of Fit Measures Holliday Model Logistic Model Lognormal Model

Marquart algorithm Monomolecular Model Morgan-Mercer Model Nonlinear Regression Normal Model Piecewise-Polynomials Probability Plots Ratio of Polynomials Reciprocal Model

Richards Model Scatter-Plot Matrix Search Routines Sum of Functions Regression Transformation Bias Correction User-Supplied Models Weibull Model

Regression Analysis
Regression analysis refers to a group of techniques for studying the relationships among two or more variables. NCSS makes it easy to run either a simple linear regression analysis or a complex multiple regression analysis. You can perform a regression analysis with modern graphical and numeric residual analysis. Major options include multiple regression, stepwise regression, correlation matrix, residual analysis, robust regression, all-possible regressions, response surface regression, proportional hazards regression, ridge regression, and logistic regression.
Multiple Regression Report Regression Equation Section Independent Variable Intercept Test1 Test2 Test3 R-Squared Regression Coefficient 85.24039 -1.933571 -1.659881 .1049543 .399068 Standard Error 23.69514 1.029096 .872896 .219902 T-Value (Ho: B=0) 3.5974 -1.8789 -1.9016 .4773 Prob Level .005772 .092969 .089661 .644541 Decision (5%) Reject Ho Accept Ho Accept Ho Accept Ho Power (5%) .891498 .389629 .397357 .071290

Normality Tests Section Assumption Value Skewness 2.0329 Kurtosis 1.5798 Omnibus 6.6285
Histogram

Probability .042064 .114144 .036361

Decision(5%) Rejected Accepted Rejected

Normal Probability Plot of Residuals of IQ
25.0

6.0

8.0

Residuals of IQ
-5.0 5.0 15.0 25.0

Count

4.0

2.0

0.0

-15.0

-15.0

-5.0 -2.0

5.0

15.0

-1.0

0.0

1.0

2.0

Residuals of IQ

Normal Distribution

All-possible regressions Alpha level is flexible Beta coefficients Biweight regression Coefficient of variation Condition number Confidence limits Cooks D Correlations CovRatio Cp Statistic DFBetas Dffits Durbin Watson statistic Eigenvalues Eigenvectors F-ratios Hat diagonal Inverse diagonal LAV regression

Logistic regression Mean squares Multiple regression Multiple regression Normal probability plot Normality tests of residuals Partial correlation Pearsons correlation Power calculations Predicted values Prediction limits Predicts new observations Press statistic P.C. Regression Proportional-Hazards R-bar squared R-squared Regression coefficient Residual analysis and plots Response surface analysis

Ridge Regression Robust regression Rstudent Serial correlation plot Simple linear regression Spearmans rank correlation Standard error of beta Standardized coefficient Stepwise regression Sum of squares T-test for beta=0 Through the origin Tolerance level Variable selection Variance inflation factor Weighted regression

T-Tests
The t-test procedures test the difference between two averages. The two sample t-test is used when the means come from two independent samples. The paired t-test is used when the observations are obtained in pairs. The one sample t-test is used to compare a single mean with a target value. NCSS offers a comprehensive t-test analysis. It automatically checks all model assumptions, performs analogous nonparametric tests, displays appropriate plots, and computes confidence limits.
Two-Sample T-Test
Descriptive Statistics Section Variable YldA YldB Count 13 16 Mean 549.3846 557.5 Standard Deviation 168.7629 104.6219 Standard Error 46.80641 26.15546 95% LCL of Mean 447.4022 501.7509 95% UCL of Mean 651.367 613.249

Confidence-Limits of Difference Section Variance Mean Assumption DF Difference Equal 27 -8.115385 Unequal 19.1690 -8.115385 Equal-Variance T-Test Section Alternative Hypothesis T-Value (YldA)-(YldB)<>0 -0.1588 (YldA)-(YldB)<0 -0.1588 (YldA)-(YldB)>0 -0.1588 Tests of Assumptions Section Assumption Skewness Normality (YldA) Kurtosis Normality (YldA) Omnibus Normality (YldA) Variance-Ratio Equal-Variance Test Modified-Levene Equal-Variance Test
Box Plots
1000.0

Standard Deviation 136.891 198.5615

Standard Error 51.11428 53.61855

95% LCL of Mean -112.9932 -120.2734

95% UCL of Mean 96.76247 104.0426

Prob Level 0.875032 0.437516 0.562484

Decision (5%) Accept Ho Accept Ho Accept Ho

Power (Alpha=.05) 0.052693 0.068110 0.035954

Power (Alpha=.01) 0.010837 0.014804 0.006616

Value 0.2691 0.3081 0.1673 2.6020 13.9235

Probability 0.787854 0.758028 0.919743 0.092546 0.002234

Decision(5%) Cannot reject normality Cannot reject normality Cannot reject normality Cannot reject equal variances Reject equal variances
Normal Probability Plot of YldB

800.0

Amount

600.0

400.0

200.0

YldA

YldB

300.0

425.0 -2.0

550.0

YldB

675.0

800.0

-1.0

0.0

1.0

2.0

Variables

Normal Distribution

Aspin-Welch unequal-variance test Box plots Confidence limits Descriptive statistics Kolmogorov-Smirnov test Levene equal-variance test

Mann-Whitney U test Normal probability plots Normality tests One and two tailed tests One sample t-test Pair comparisons t-test

Power calculations Probability levels Quantile test Two sample t-test Unequal variance case Wilcoxon rank sum test

Analysis of Variance
Analysis of variance is the statistical technique for testing differences among group means. NCSS contains several analysis of variance procedures, including general linear models (GLM), one-way analysis of variance, unweighted means analysis of variance, and MANOVA. You can perform a simple one-way analysis of variance, a complex repeated-measures analysis of variance, a factorial analysis of variance, an analysis of covariance, and a host of multiple-comparison tests. All procedures work with balanced and unbalanced experimental designs.
Analysis of Variance Table Source Term A (Treatment) B(A) C (Time) AC BC(A) Sum of Squares 486.7488 4064.833 277.7778 352.0833 82.5 Mean Square 165.5833 270.9889 277.7778 117.3611 5.5 Prob Level .555761 .000004* .000041* Power (Alpha=0.05) .089834 .995231 .968851

DF 3 15 1 3 15

F-Ratio .61 50.51 21.34

Tukey's HSD Multiple-Comparison Test Term A: Exercise Alpha=0.050 Error Term=B(A) DF=15 MSE=270.9889 Critical Value=4.895637 Different Group Count Mean From Groups 1 3 210.6667 4 3 211.3333 3 3 238 2 3 257.6667

Means of Response
300.0
300.0

Means of Response
Treatment 1 2 3 4

250.0

Response

Response
1 2 3 4

200.0

150.0

100.0

100.0

150.0

200.0

250.0

Treatment

Blocks

Analysis of covariance ANOVA table Area Under Curve Assumptions testing Automatic F-ratios Bonferroni tests Box plots Boxs M test Comparisons Contrasts Covariance analysis Duncans test Expected mean squares F-ratios Factorial designs Fishers LSD test Fixed terms Fractional factorial designs Geisser-Greenhouse GLM solution

Hotellings trace Huynh-Feldt Correction Interation plots Kruskal Wallis z tests Latin square designs Least-squares means Levene Variance test MANOVA (with up to 10 factors) Mauchleys Test Mean Squares Means plots Multiple comparisons of factors Multiple comparisons of interactions Nested terms Newman-Keuls test Normality tests Pillais Trace

Planned comparisions Post-hoc tests Power calculations Random terms Randomized complete block Repeated measures R and R Study Roys Root Scheffes test Split-plot designs Tukey-Kramer HSD tests Unbalanced data Wilks Lambda

Repeated Measures ANOVA

The repeated measures procedure performs an analysis of variance on within-subject designs using the general linear models approach. The experimental design may include up to three betweensubject factors as well as three within-subject factors. Boxs M test and Mauchleys test of the assumptions about the within-subject covariance matrices are provided. Geisser-Greenhouse, Box, and Huynh-Feldt corrected probability levels on the within-subject F tests are given along with the associated test power.
Analysis of Variance Table Source Term A: Exercise B(A): Subject C: Time AC BC(A) Sum of Squares 427.4445 5234.556 547.4445 191.4444 222.4444 Mean Square 213.7222 348.9704 273.7222 47.86111 7.414815 Prob Level 0.555040 0.000000* 0.000716* Power (Alpha=0.05) 0.076179 0.999974 0.874052

DF 2 15 2 4 30

F-Ratio 0.61 36.92 6.45

Probability Levels for F-Tests with Geisser-Greenhouse Adjustments Geisser Greenhouse Box Regular Epsilon Epsilon Source Prob Prob Prob Term DF F-Ratio Level Level Level A: Exercise 2 0.61 0.555040 B(A): Subject 15 C: Time 2 36.92 0.000000* 0.000021* 0.000000* AC 4 6.45 0.000716* 0.009496* 0.000755* BC(A) 30 Power Values for F-Tests with Geisser-Greenhouse Adjustments Geisser Greenhouse Box Regular Epsilon Epsilon Source Power Power Power Term DF F-Ratio (Alpha=0.05) (Alpha=0.05) (Alpha=0.05) A: Exercise 2 0.61 0.076179 B(A): Subject 15 C: Time 2 36.92 0.999974 0.889537 0.999966 AC 4 6.45 0.874052 0.371111 0.867399 BC(A) 30
Means of HeartRate
90.00

Huynh Feldt Epsilon Prob Level

0.000000* 0.000716*

Huynh Feldt Epsilon Power (Alpha=0.05)

0.999974 0.874052

HeartRate vs Time by Subject

Time 0 10 20
90.00

78.75

78.75

HeartRate

67.50

HeartRate
0 - None 1 - Week 2 - Dail

67.50

56.25

56.25

45.00

45.00 0 10 20

Exercise

Time

Analysis of variance Boxs M test Circularity test Compound symmetry test Contrasts Covariance matrix tests Epsilon values F ratios

Fixed or random factors Geisser-Greenhouse Huynh-Feldt correction Mauchleys test Means plots Multiple comparisons Planned comparisons Power values

Probability levels Randomized block designs Repeated measures Subject plots Within-subject designs

Design of Experiments
These programs can generate 2k factorial designs, B.I.B. designs, d-optimal designs, fractional factorial designs, Latin square designs, screening designs, Taguchi designs, and response surface designs. NCSS also provides for the analysis of these designs using either the general purpose ANOVA and regression procedures or specialized procedures for designs with factors at only two levels.
Means and Effects Section Term No. 0 1 2 3 4 5 6 7 Term Symbol Grand Mean A (Temp) B (Concentration) AB C (Catalyst) AC BC ABC Estimated Effect 64.25 23.00 -5.00 1.50 1.50 10.00 0.00 0.50 Standard Error 0.71 1.41 1.41 1.41 1.41 1.41 1.41 1.41

Mean 52.75 66.75 63.50 63.50 59.25 64.25 64.00

Mean + 75.75 61.75 65.00 65.00 69.25 64.25 64.50

Analysis of Variance Section Term No. 1 2 3 4 5 6 7 Term Symbol A (Temp) B (Concentration) AB C (Catalyst) AC BC ABC Error Total Mean Square 2116.0000 100.0000 9.0000 9.0000 400.0000 0.0000 1.0000 8.0000 2699.0000 Prob Level 0.000000 0.007670 0.319813 0.319813 0.000105 1.000000 0.732810 Statistically Significant Yes Yes No No Yes No No

DF 1 1 1 1 1 1 1 8 15

F-Ratio 264.50 12.50 1.13 1.13 50.00 0.00 0.13

Normal Probability Plot of Effects

25.0
1.5

Contour Plot of Odor

Odor -40.0 -20.0 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0

16.3

Effects

Temp

-1.3

7.5

-10.0

-1.5

-0.8

0.0

0.8

1.5

-1.5 -2

-0.8

0.0

0.8

-1

Normal Distribution

0 Ratio

Aliasing reports Analysis of variance tables Automatic design detection Balanced incomplete block Block confounding Blocking Box-Behnken designs Box-Hunter-Hunter designs Central composite designs Contour plots

D-optimal designs Effect estimation Fractional factorials Latin Squares Factorial designs Means Plots Placket-burman designs Probability plots of residuals and effects Repeated Measures designs

Residual analysis Response-surface designs Screening designs (up to 31 Factors) Taguchi designs

Survival Analysis
Survival analysis includes several techniques to study data in which the response variable is elapsed time. Procedures include survival distribution analysis (including the Kaplan-Meier survival distribution estimate), log rank tests, and proportional hazards regression.
Kaplan-Meier Product-Limit Survival Distribution Sample Size 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 Survivorship S(t) .947368 .891641 .835913 .780186 .724458 .664087 .603715 .536636 .469556 Std Error of S(t) .051228 .072440 .086738 .097223 .105043 .112306 .117205 .121875 .123732 Hazard Fn H(t)=-Log(S(t)) .054067 .114692 .179230 .248223 .322331 .409343 .504653 .622436 .755967 Std Error of H(t) .238910 .301855 .352326 .399657 .447373 .504634 .567076 .650547 .749122

Rank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Time 3.0 3.0+ 4.0 6.0 8.0 8.0 8.0+ 10.0 12.0 13.0+ 16.0 17.0 21.0+ 26.0+ 30.0 33.0 35.0+ 44.0+ 45.0+

.375645 .281734

.129821 .126865

.979111 1.266793

.959169 1.264245

S(t) Exponential Plot

1.000 1.400

Hazard Function Plot

0.750

Hazard Function
0.0 12.5 25.0 37.5 50.0

S(t) Exponential

0.500

0.250

0.000

0.000

0.350 0.0

0.700

1.050

12.5

25.0

37.5

50.0

TIME3

TIME3

Accelerated life testing Beta distribution Censoring All Types Coxs regression Cox-Mantel test Exponential distribution Exponential regression Extreme Value distribution Extreme Value regression Gamma distribution Gehans Wilcoxon test Goodness of fit

Hazard rate plots Hazard functions & plots Kaplan-Meier product limit Least squares estimates Log-logistic distribution Log-logistic regression Log rank test Lognormal distribution Lognormal regression Maximum likelihood Peto / Wilcoxon test Probit Analysis

Proportional hazards regression Rayleigh distribution Reliability statistics Stress variables Survival functions & plots Survival percentiles Variable selection options Weibull distribution Weibull regression

Reliability Analysis
The reliability procedure estimates the parameters of the exponential, extreme value, logistic, loglogistic, lognormal, normal, and Weibull probability distributions by maximum likelihood and least squares. It can fit complete, right censored, left censored, interval censored (readout), and grouped data values. It also computes the nonparametric Kaplan-Meier and Nelson-Aalen estimates of reliability and associated hazard rates. Another module fits the regression relationship between time-to-failure and one or more independent variables. The distribution of the residuals (errors) can follow the exponential, extreme value, logistic, log-logistic, lognormal, lognormal10, normal, or Weibull distributions. The data may include failed, left censored, right censored, and interval observations. This type of data often arises in the area of accelerated life testing.
Weibull Parameter Estimation Section Probability Maximum Plot Likelihood Parameter Estimate Estimate Shape 1.26829 1.511543 Scale 279.7478 238.3481 Log Likelihood -80.05649 Reliability Section Prob Prob Failure Estimate of Time Survival 40.0 0.9186 80.0 0.8151 120.0 0.7105 160.0 0.6112 Percentile Section Prob Prob Failure Estimate of Time Failure Percentage Time 25.0000 104.7 50.0000 209.5 75.0000 361.9 95.0000 664.5
Weibull Hazard Rate Plot
0.010

MLE Standard Error 0.4128574 57.21616

MLE 95% Lower Conf. Limit 0.8849655 148.8944

MLE 95% Upper Conf. Limit 2.581753 381.5444

Max Like Estimate of Survival 0.9349 0.8253 0.7016 0.5784

95% Lower Conf. Limit of Survival 0.8065 0.6728 0.5312 0.3775

95% Upper Conf. Limit of Survival 0.9791 0.9112 0.8199 0.7352

Max Like Estimate of Failure Time 104.5 187.0 295.8 492.6

95% Lower Conf. Limit of Failure Time 69.7 124.7 170.9 227.8

95% Upper Conf. Limit of Failure Time 156.8 280.5 512.1 1065.0
Weibull Probability Plot

5.50

4.75
0.008

Hazard Rate

Ln(Time)

4.00

0.005

0.003

3.25

0.000

0.0

50.0

100.0

150.0

200.0

2.50 -4.00

-3.13

-2.25

-1.38

-0.50

Time

Weibull Quantile

Accelerated life testing Arrhenius transformation Beta distribution Censoring All Types Confidence Limits Cox-Snell Residuals Distribution selection Exponential distribution Exponential regression Extreme-value distribution Extreme-value regression Failure time percentiles

Gamma distribution Goodness of fit Hazard rate plots Hazard functions & plots Information Matrix Kaplan-Meier product limit Least squares estimates Log-logistic distribution Log-logistic regression Lognormal distribution Lognormal regression Maximum likelihood

Probability plots Reliability statistics Residual Analysis Residual Life Report Stress variables Stress plots Threshold Estimate Weibull distribution Weibull regression

Quality Control Charts and SPC

NCSS provides a complete set of the latest variables and attribute control charts. It includes popular features such as runs tests, individuals charts, automatic outlier removal, capability analysis, usersupplied sigma, and comprehensive chart labeling. Since the quality control programs are completely integrated with the rest of the system, once you detect an out-of-control signal, you can investigate with the other procedures in NCSS.
Xbar and R Charts
Xbar Chart
2015.0

Range Chart
60.0

13 22

36
42.5

11

2006.3

Range

1997.5

Xbar

1988.8

1980.0

-10.0

7.5 0.0

25.0

0.0

30.0

60.0

90.0

120.0

30.0

60.0

90.0

120.0

Row

Row

Out-of-Control List Row Mean 4 6.8 5 12.2 6 9.8 7 5.6

Range 5 48 27 4

Row Label 4 5 6 7

Reason Range: 4 of 5 in zone B or beyond Xbar: beyond control limits Xbar: 2 of 3 in zone A Xbar: 2 of 3 in zone A

Capability Analysis Section Parameter 3-Sigma Limits 4-Sigma Limits Specification Limits Specification z-Values Percent Outside Specification Capacities Cp Index Cpk Index Count = 246 Sigma = 3.809114

Lower Center -5.390758 6.036585 -9.199872 6.036585 1 -1.322246 3.252033 0.440749 0.518452 0.568811 0.383108 0.440749 Alpha Level = 0.050000

Upper 17.46393 21.27304 14 2.090621 2.439024 0.696874 0.619112 0.498389

Attribute charts C chart Capability analysis Cp index Cpk Index Cusum chart EWMA chart Exception report Frequency distribution Histogram

Individuals chart Moving average chart Moving range chart Normality test NP chart Out-of-control list P chart Pareto charts Primary control limits R and R Study

R chart Robust chart analysis Runs tests S (sigma) chart Secondary control limits U chart Variables charts Xbar chart Zone display

Multivariate Analysis
Multivariate analysis refers to a group of statistical techniques that analyze two or more variables at a time, including: discriminant analysis, factor analysis, cluster analysis, logistic regression, MANOVA, and principal component analysis.
Factor Analysis Report Eigenvalues after Varimax Rotation Individual No. Eigenvalue Percent 1 3.288191 54.89 2 2.701207 45.09 3 0.001207 0.02 4 -.000099 0.00 5 -.000121 0.00 6 -.000295 0.00 Factor Loadings after Varimax Variables Factor1 X1 -.019936 X2 -.967470 X3 -.994037 X4 -.478418 X5 -.594943 X6 -.883654 Rotation Factor2 -.998792 -.252572 -.107126 -.873578 -.803812 -.468080
Factor Loadings
0.0

Cumulative Percent 54.89 99.99 100.00 100.00 100.00 100.00

Scree Plot ||||||||||| |||||||||| | | | |

Factor Scores
3.0

16
Score1

19 21 5 15 17 11

18 30 26 12 3 1
0.0

13

10

-0.5

0.5

24 28 29 22

Loading1

20

2 27

-0.3

23 14
1.8

4 5

-0.8

-2.0

-1.0

4 8 25
1.0 2.0

-0.8

6 2
-1.0 -0.8 -0.5 -0.3

3
0.0

-2.0

-1.0

Score2

Loading2

Discriminant Analysis Canonical coefficient reports Classification reports Influential variable reports Linear discriminant functions, scores, and coefficients Prior probabilities Stepwise variable selection Cluster Algorithms Complete Linkage Dendrograms Fuzzy clustering Hierarchical K-means algorithm Mediod partitioning Nearest neighbor Regression clustering Single Linkage

Factor and Principal Component Analysis Bartletts sphericity test Communality estimates Correlation matrix input Eigenvalue/vector analysis Factor loadings and scores Gleason-Staelin redundancy Missing value estimation Outlier detection Principal axis method Quartimax rotation Robust estimation Score and Loading Plots Scree plot T2 analysis Varimax rotation Equality of Covariance Bartletts variance test Boxs M test Cronbachs alpha Eigenvalue analysis

MANOVA Approximate F-test Boxs M test Canonical analysis Homogeneity of variance test Hotellings T2 Lawley-Hotelling trace Pillais trace Roys largest root Wilks lambda Logistic Regression Beta estimates Chi-square tests Classification table Normal probability plots Odds ratios Predict new observations Step-down variable selection Step-up variable selection

More Multivariate Procedures

Loglinear models techniques are used to analyze the relationships among the factors in a multi-way contingency table. Several model selection techniques, goodness of fit tests, and parameter estimation reports are provided. Correspondence analysis is a technique for graphically displaying a two-way contingency table by calculating coordinates representing its rows and columns. These coordinates are analogous to factors in a factor analysis. Multidimensional scaling is a technique that creates a map displaying the relative positions of a number of objects, given only a table of the distances between them. The program calculates either the metric or the non-metric solution.
Loglinear Models Multiple-Term Test Section Like. Ratio Prob K-Terms DF Chi-Square Level 1WAY & Higher 31 2666.19 0.0000 2WAY & Higher 26 596.84 0.0000 3WAY & Higher 16 19.56 0.2406 4WAY & Higher 6 3.23 0.7791 5WAY & Higher 1 1.02 0.3116 Like. Ratio Chi-Square 2069.35 577.28 16.33 2.21 Prob Level 0.0000 0.0000 0.0906 0.8196

Pearson Chi-Square 3811.81 751.31 21.21 3.32 1.01

Prob Level 0.0000 0.0000 0.1705 0.7680 0.3157

K-Terms 1WAY Only 2WAY Only 3WAY Only 4WAY Only

DF 5 10 10 5

Correspondence Plot
0.15 Light

MDS Map
2.50

SC

Golf
1.50

Basketball

JE Factor2 (12%) Medium

Dim2

-0.50

0.50

SE None

Heavy -0.25 JM

SM

Croquet
-1.50

Hockey Tennis Football

0.00 2.00 4.00

-0.40 Factor1 (88%)

0.50

-4.00

-2.00

Dim1

Loglinear Models Automatic model selection Delta value for zero cells Estimated effects Goodness of fit statistics Hierarchical models Likelihood ratio tests Pearson chi-square tests Residual analysis Up to seven factors

Correspondence Analysis Chi-square contributions Column/row profiles Coordinates report Eigenvalues Mass values Point quality index Supplementary rows Variable/profile correlation

Multidimensional Scaling Classical (metric) scaling Dissimilarity and correlation input Eigenvalue report Kruskals stress statistic Non-metric scaling

Forecasting and Time Series Analysis

NCSS provides several methods of forecasting and time series analysis. For forecasting, it provides classical methods based on exponential smoothing, trend-season-cycle decomposition (like X11), and ARIMA (Box-Jenkins). For time series analysis, it provides spectral analysis, ARIMA, and autocorrelation analysis. The program contains a theoretical procedure that generates univariate time series from a specified model. This is useful for improving your forecasting skills.
Model Estimation Section Parameter Parameter Name Estimate AR(1) 0.5778422 AR(2) 0.3962256 MA(1) 0.7604249 Observations Iterations Pseudo R-Squared

Standard Error 0.1725721 0.1596349 0.1349649 40 20 21.586142

Prob T-Value Level 3.3484 0.000813 2.4821 0.013062 5.6342 0.000000 Residual Sum of Squares Mean Square Error Root Mean Square

28.84119 0.7794916 0.8828882

Autocorrelations of Residuals of Series1-MEAN Lag Correlation Lag Correlation Lag 1 -0.091919 11 0.003063 21 2 -0.089101 12 0.000842 22 3 -0.006732 13 -0.037963 23 4 -0.091508 14 0.038344 24 . . . . . . . . . . Significant if |Correlation|> 0.316228

Correlation 0.090220 -0.036684 0.105105 0.048037 . .

Lag 31 32 33 34 . .

Correlation 0.057736 -0.069254 0.112294 0.049954 . .

Series1-MEAN Chart
1.000 2.0

Autocorrelations of Residuals

Autocorrelations
2006 2022 2038 2054

Series1-MEAN

1990

-1.000

-4.0

-0.500 0

-2.5

0.000

-1.0

0.500

0.5

10

19

29

38

Time

Lag

ARIMA ARMA Autocorrelations Automatic Box-Jenkins Box-Jenkins Method Census X11 Cycle Analysis Decomposition Methods Double-Exponential Smoothing

Easy To Use Error Plots Exponential Smoothing Forecast Plots Generate Series Harmonic Analysis Log Transformation Modified Yule-Walker Equations Partial Autocorrelations

Periodogram Portmanteau Test Prediction Limits - ARIMA Residual Analysis Seasonal Adjustment Seasonal Analysis Seasonal ARIMA Models Spectral Analysis Trend Analysis

NOW YOU CAN

Be up and running in less than 30 minutes with our Quick Start booklet. Complete your first statistical analysis in less than an hour. Have your technical questions answered free by a PhD statistician. Have confidence when ordering because of our 30 day unconditional money back guarantee.

Easily mix graphs and reports together in a format recognized by word processors. Enter your data into a spreadsheet. Have confidence in the accuracy of your results since they are backed by our 18-year-old company. Spend your time doing interpretation rather than programming. Analyze large (200,000+ rows)

NCSS LETS YOU

WHEN YOU BUY NCSS, YOU WILL HAVE A PROGRAM THAT

Is comprehensive and accurate Is easy to learn and use. Imports and Exports all major spreadsheet, database, and statistical file formats (including Excel and dBase). Outputs reports in a tabbed format that is ready for Word or Word Perfect. Costs least than half what our competitors charge.

OUR GUARANTEE
If you are not completely satisfied with an NCSS product during the first 30 days for any reason, return the program for a full, prompt refund (excluding shipping)no questions asked. HERES WHAT DR. B ILL B IGG
OF HUMBOLDT STATE

UNIVERSITY F ORESTRY DEPARTMENT SAYS In the fifteen years I have been teaching I have used several different statistical software packages. None have been easier to learn and operate than NCSS for Windows. The help file is particularly complete and makes interpretation of the output straight forwardeven when you cant remember what a particular item means.

HERES WHAT DR. CHET MCCALL OF HERES WHOS P UBLISHING USING NCSS TO DO THEIR ANALYSIS EDUCATION AND P SYCHOLOGY AT Am Heart Journal N.S. Kleiman 1994 P EPPERDINE UNIVERSITY SAYS Am J. of Physical Anthropology C. Tardieu 1994
THE G RADUATE SCHOOL OF

For more than seven years, we have watched NCSS evolve into a most user friendly and powerful statistical system. We find that NCSS is easy to use and, for those with a math phobia, it does the job with minimal instruction. Tutorials are outstanding. Anybody in higher education who needs statistical software could do well to view NCSS. I personally feel it is the best buy on the market.

Am J. of Physiology Analytical Biochemistry Annals of Human Biology Annals of Oncology Annals of Surgery Bone Marrow Transplantation Hypertension J. Am College of Nutrition J. Applied Physiology J. of Family Practice J. Pharmacy and Pharmacology Oecologia Preventive Veterinary Medicine

M. Okada S. Butenas J.H. Relethford L. Vanwarmerdam J.A. Morris E. Conde P. Boutouyrie M.J. Glade J. Qvist R. Sheff P. Bustamante C.J. Bilbrough P.D. Mansell

1994 1995 1994 1994 1994 1994 1995 1993 1993 1994 1993 1995 1993

NCSS S ACCURACYWe at NCSS have put a great deal of effort into finding the most accurate algorithms possible. The programs have been tested and verified over and over, both by us and by our customers. Each routine has been verified against textbooks, journal articles, and, where possible, other software. NCSS is one of the most accurate statistical analysis programs available. NCSS calculates with seventeen-digit, double-precision accuracy. Yes! I want to obtain the latest version of these NCSS products. My Payment Option: Please rush me my own personal license of NCSS 2000 and/or ___ Check enclosed ___ Please charge my: __VISA __MasterCard __Amex PASS 2000 at their amazingly low prices.
Qty ___ NCSS 2000 CD and printed documentation: $399.95.......... $ ____ ___ NCSS 2000 CD (electronic documentation): $299.95 ......... $ ____ ___ PASS 2000 CD and Printed Manual: $249.95...................... $ ____ ___ PASS 2000 Upgrade CD for PASS 6.0 users: $99.95............ $ ____ ___ PASS 2000 Manual for PASS 6.0 users: $49.95 ................... $ ____
NAME

___ Purchase order enclosed

VISA/MC/AMEX Expiration

Signature_______________________________________________

Telephone Number (Required) _________________________

Ship my NCSS Products to:

_________________________________________________
COMPANY _______________________________________________ ADDRESS (1) ADDRESS (2)

Shipping & Handling: USA: $10 regular, $17 2-day, $30 overnight. Canada: $17 Mail. Europe: $40. International: $70 .................... $ ____ Total: ................................................................................... $ ____ FOR FASTEST DELIVERY, CALL

_____________________________________________ _____________________________________________

1-800-898-6109
Email your order to sales@ncss.com or fax your order to 1-801-546-3907 NCSS, 329 North 1000 East, Kaysville, UT 84037

CITY/STATE/ZIP ___________________________________________ COUNTRY _______________________________________________ EMAIL _____________________________________________________________________