GAP Toolkit 5 Training in basic drug abuse data management and analysis

GAP Toolkit 5 Training in basic drug abuse data management and analysis

Training session 7

SPSS: Recode and Compute

Objectives

• To introduce and demonstrate SPSS tools for recoding variables and creating new variables

• To introduce and demonstrate the use of SPSS command syntax

• To introduce and demonstrate Help features in SPSS from a dialogue box

Types of recode

• A categorical variable into a new set of categories• A continuous variable into a categorical variable

Recode approaches

• Into pre-defined categories of interest• Into categories of equal size

Example: pre-defined categories

• Recode weekly expenditure on drugs into the following categories:– Low expenditure <= £100 a week– High expenditure >= £101 a week

select variable for recoding and click

click here to map old new values

Frequency Percent Valid percent

Cumulative percent

Valid Low 31 62.0 62.0 62.0

High 19 38.0 38.0 100.0

Total 50 100.0 100.0

Weekly drug expenditure

Example: quantiles

• Recode weekly expenditure into the following:– 4 categories, each with approximately the same number of

cases (quartiles)– 10 categories, each with approximately the same number of

cases (deciles)

Frequency Percent Valid percent Cumulative percent

Valid <=£90 12 24.0 24.0 24.0

£91 <= X <= £98 12 24.0 24.0 48.0

£99 <= X <= £105 13 26.0 26.0 74.0

£106 <= X >= £116 13 26.0 26.0 100.0

Total 50 100.0 100.0

NTILES of Expend

Frequency Percentage Valid percentage Cumulative percentage

Valid 1 5 10.0 10.0 10.0

2 4 8.0 8.0 18.0

3 7 14.0 14.0 32.0

4 4 8.0 8.0 40.0

5 4 8.0 8.0 48.0

6 6 12.0 12.0 60.0

7 5 10.0 10.0 70.0

8 5 10.0 10.0 80.0

9 6 12.0 12.0 92.0

10 4 8.0 8.0 100.0

Total 50 100.0 100.0

NTILES of EXPEND

Compute

• Creates new variables using:– Standard mathematical operators– A full range of functions– Conditional statements

Transform/Computeclick here to define variable type

Type and Label

Yexpend=Expend*52

Explore: Yexpend

Statistic Std. error

Yearly drug expenditure Mean 5062.72 77.468

95% Confidence Interval for Mean

Lower Bound 4907.04

Upper Bound 5218.40

5% Trimmed Mean 5072.89

Median 5148.00

Variance 300062.3

Std. Deviation 547.779

Minimum 3848

Maximum 6032

Range 2184

Interquartile Range 754.00

Skewness -.333 .337

Kurtosis -.446 .662

Descriptives

Example 2: Age

• Convert date of birth to age in years

(DATE.DMY(01,01,2002) - dob) / (365.25*24*60*60)

Explore: Age

Statistic Std. error

AGE Mean 25.43 1.219

95% Confidence Interval for Mean

Lower Bound 22.97

Upper Bound 27.88

5% Trimmed Mean 25.21

Median 22.84

Variance 72.867

Std. Deviation 8.536

Minimum 11

Maximum 44

Range 32

Interquartile Range 12.77

Skewness .579 .340

Kurtosis -.518 .668

Descriptives

Help

• Context-sensitive Help:– Right click on any part of a dialogue box to obtain help

• Dialogue box Help button

Command language

• Command-line rather than Windows interface• More flexible• Reusable• Saved as text files

Syntax window

• Enter SPSS commands:– Type directly– Paste from a dialogue box

• Run SPSS commands• Save as a text file

The Paste button

click to copy command into Syntax Window

The Run menu

Syntax for Age

Summary

• Recode:– Into same variable– Into different variable

• Compute:– Mathematical operators– Functions

• Dialogue box Help:– Context-sensitive Help– Dialogue box Help button

• Command language:– Syntax Window– Dialogue box Paste button