Demonstration of Statistical Software

Eugene TseytlinDepartment of BioMedical Informatics,

University of Pittsburgh

Overview

• Dataset Overview• Descriptive Statistics Using Calc• Descriptive Statistics Using PSPP• Descriptive Statistics Using R• EpiInfo Demonstration Video

Brain Size and IntelligenceAre the size and weight of your brain indicators of your mental capacity? In this

study by Willerman et al. (1991) the researchers use Magnetic Resonance Imaging (MRI) to determine the brain size of the subjects. The researchers take into account gender and body size to draw conclusions about the connection between brain size and intelligence.

http://lib.stat.cmu.edu/DASL/Stories/BrainSizeandIntelligence.html

Methods Correlation

Regression

Scatterplot

Brain Size and Intelligence

Description: Willerman et al. (1991) collected a sample of 40 right-handed Anglo introductory psychology students at a large southwestern university. Subjects took four subtests (Vocabulary, Similarities, Block Design, and Picture Completion) of the Wechsler (1981) Adult Intelligence Scale-Revised. The researchers used Magnetic Resonance Imaging (MRI) to determine the brain size of the subjects. Information about gender and body size (height and weight) are also included. The researchers withheld the weights of two subjects and the height of one subject for reasons of confidentiality.

Data

Gender: Male or FemaleFSIQ: Full Scale IQ scores based on the four Wechsler (1981) subtestsVIQ: Verbal IQ scores based on the four Wechsler (1981) subtestsPIQ: Performance IQ scores based on the four Wechsler (1981) subtestsWeight: body weight in poundsHeight: height in inchesMRI_Count: total pixel Count from the 18 MRI scans

Number of cases: 40

Gender FSIQ VIQ PIQ Weight Height MRI_CountFemale 133 132 124 118 64.5 816932Male 140 150 124 ¥ 72.5 1001121Male 139 123 150 143 73.3 1038437Male 133 129 128 172 68.8 965353Female 137 132 134 147 65.0 951545Female 99 90 110 146 69.0 928799Female 138 136 131 138 64.5 991305Female 92 90 98 175 66.0 854258Male 89 93 84 134 66.3 904858Male 133 114 147 172 68.8 955466Female 132 129 124 118 64.5 833868Male 141 150 128 151 70.0 1079549Male 135 129 124 155 69.0 924059Female 140 120 147 155 70.5 856472Female 96 100 90 146 66.0 878897Female 83 71 96 135 68.0 865363Female 132 132 120 127 68.5 852244Male 100 96 102 178 73.5 945088Female 101 112 84 136 66.3 808020Male 80 77 86 180 70.0 889083Male 83 83 86 ¥ ¥ 892420Male 97 107 84 186 76.5 905940Female 135 129 134 122 62.0 790619Male 139 145 128 132 68.0 955003Female 91 86 102 114 63.0 831772Male 141 145 131 171 72.0 935494Female 85 90 84 140 68.0 798612Male 103 96 110 187 77.0 1062462Female 77 83 72 106 63.0 793549Female 130 126 124 159 66.5 866662Female 133 126 132 127 62.5 857782Male 144 145 137 191 67.0 949589Male 103 96 110 192 75.5 997925Male 90 96 86 181 69.0 879987Female 83 90 81 143 66.5 834344Female 133 129 128 153 66.5 948066Male 140 150 124 144 70.5 949395Female 88 86 94 139 64.5 893983Male 81 90 74 148 74.0 930016Male 89 91 89 179 75.5 935863

Descriptive Statistics in Calc

Load Dataset into Calc

Import DatasetCreate BMI Column

=(weight/(height^2))*703

Categorize BMIUnderweight <18.5Normal < 25Overweight < 30Obese > 30

=IF(BMI<18.5,”Underweight”, IF(BMI<25,”Normal”, IF(BMI<30,”Overweight”,”Obese”)))

Use DataPilot Feature

Data → DataPilot → Start..

Charts and Graphs

• Bar chart of Male IQ vs Female IQ

• XY Scatter Plot of IQ vs MRI pixel count

• XY Scatter Plot of Weight vs Height

Female Male110

111

112

113

114

115

116

70 80 90 100 110 120 130 140 1500

200000

400000

600000

800000

1000000

1200000

100 110 120 130 140 150 160 170 180 190 20050

55

60

65

70

75

80

Descriptive Statistics Using R

Processing Data in R

Import Data> data =read.table("brain-size.csv",1,"\t");

Add BMI Data Columns> data$bmi=data$Weight/(data$Height^2)*703;

Descriptive Statistics in R

Mean IQ> mean(data$FSIQ)

[1] 113.45

Standard Deviation of IQ> sd(data$FSIQ)

[1] 24.08207

Summary> summary(data$FSIQ)

Min. 1st Qu. Median Mean 3rd Qu. Max.

77.00 89.75 116.50 113.40 135.50 144.00

T-Test in R> t.test(data$FSIQ[data$Gender=="Female"],data$FSIQ[data$Gender=="Male"])

Welch Two Sample t-test

data: data$FSIQ[data$Gender == "Female"] and data$FSIQ[data$Gender == "Male"]

t = -0.4027, df = 37.892, p-value = 0.6895

alternative hypothesis: true difference in means is not equal to 0

95 percent confidence interval:

-18.68639 12.48639

sample estimates:

mean of x mean of y

111.9 115.0

Correlation in R> cor.test(data$Weight,data$Height)

Pearson's product-moment correlation

data: data$Weight and data$Height

t = 5.8748, df = 36, p-value = 1.021e-06

alternative hypothesis: true correlation is not equal to 0

95 percent confidence interval:

0.4893837 0.8329941

sample estimates:

cor

0.699614

Plots in R

XY Scatter Plot between Verbal IQ and Total IQ> plot(data$FSIQ,data$VIQ)

Charts in R

Bar Graph of Means of Male vs Female Iqs>barplot(c(mean(data$FSIQ[data$Gender=="Female"]),mean(data$FSIQ[da

ta$Gender=="Male"])),names.arg=levels(data$Gender))

PSPP

Descriptive Statistics in PSPP

Analyze → Descriptive Statistics → Descriptives

Tests in PSPP

Independent Sample T-TestAnalyze → Compare Means → Independent Sample T Test

Conclusion

There are many open source software packages for statistical analysis

While some packages are completely analogous to their respective non-free alternatives, others are still work in progress

The important thing is to know what is out thereFor rest there is always Google.