ARL Workshop for Safety Sciences Doctoral Program 2015-07-16

Downloading SPSS to Home Computer

Website:

You need to know which version to install.

PC: 32 bit vs. 64 bit version

StartRight click “Computer”Select “Properties”Locate“System” Read “System type”

Mac: “bitness-agnostic;” processer and year of creation -newer than 2008 = 64-bit, but not important for downloading SPSS from IUP IT. If you want to check, however, click the Apple logo in the upper left hand corner, select “About this Mac.” Or, if you know the processor name, then refer to this chart.

The software expires every August, so you will need to renew your license.

SPSS Hands-On Practice

Exercise Manually Creating Data Set:

Worker respirable silica (quartz) exposures in µg/m3 as an 8-hour TWA are measured in stone-working shops. Ten shops use equipment withwet methods for dust suppression and 10 shops use equipment withoutwet methods for dust suppression. Time weight average (TWA) is calculated based on 8 hour exposure (work day).The calculated exposure level wasalso compared to OSHA PEL (permissible exposure limit calculated based on 8hr TWA)of 50 µg/m3 .

Method / Exposure Level / OSHA PEL
Dry / 12 / yes
90 / no
25 / yes
18 / yes
55 / no
50 / yes
35 / yes
77 / no
145 / no
130 / no
Wet / missing
36 / Yes
13 / Yes
10 / yes
24 / yes
22 / yes
17 / yes
32 / yes
53 / no
49 / yes

Options

EditOptionsOutput TabVariable values in labels are shown asValues and Labels

Data View vs. Variable View

Important Variable Properties

  • Variable Name (rules)
  • Unique
  • Begin with letter
  • No stops, spaces, or symbols
  • No commands
  • ≤ 64 characters
  • Variable Type (common types)
  • Numeric
  • String
  • Date
  • Variable Label
  • Describe variable
  • will be shown in output tables b/c of “options” set-up
  • Variable Values
  • Define meaning of variable
  • “operationalize”
  • Missing Values
  • Logically different value
  • E.g. -99=not applicable
  • Go back to Variable Values and add missing value for specific variable
  • Variable Measure
  • Level of Measurement
  • Nominal = Nominal
  • Ordinal = Ordinal (but can be treated as continuous = Scale)
  • Interval = Scale
  • Ratio = Scale

Entering Data

Each row = “case” and the values of each variable for that case

Move from left to right

Trick: Utilize Value Labels Icon

Enter -99 for missing data

Create ID Variable for each case

Introducing Syntax:

FileNewSyntax

Copy/Paste into Syntax window:

COMPUTE id=$CASENUM.

FORMAT id (F8.0).

EXECUTE.

Highlight SyntaxClick Green “play” button aka “run selection”

Save Syntax

Check Data/Variable Viewmove ID variable to first position

Codebook

AnalyzeReportsCodebook

Or Copy/Paste/Run in Syntax:

CODEBOOK id [n] Method [n] ExposureLevel [s] OSHAPEL [n]

/VARINFO LABEL TYPE MEASURE VALUELABELS MISSING

/OPTIONS VARORDER=VARLIST SORT=ASCENDING MAXCATS=200

/STATISTICS NONE.

Save Syntax

Save & export output as PDF

Preliminary Analysis or “Getting to know your data”

Case Summaries

AnalyzeReportsCase Summaries

Or Copy/Paste/Run in Syntax:

SUMMARIZE

/TABLES=id Method ExposureLevel OSHAPEL

/FORMAT=VALIDLIST NOCASENUM NOTOTAL

/TITLE='Case Summaries'

/MISSING=VARIABLE

/CELLS=NONE.

Save Syntax

Save output.

Frequencies (categorical variables)

AnalyzeDescriptive StatisticsFrequencies

Move categorical variables into Variables box

Click Paste

Go to SyntaxHighlightRun Selection

Save Syntax

Save Output

Descriptives (continuous variables)

AnalyzeDescriptive StatisticsDescriptives

Move continuous variables into Variables box

OptionsSelect Mean, Standard Deviation, Minimum, Maximum

Click Paste

Go to SyntaxHighlightRun Selection

Save Syntax

Save Output

Graphs (visualizing your data)

  • Histograms (distribution of scores on a continuous variable)
  • Bar Chart (cases/score per categorical variable)
  • Box plots (dispersion & differences b/w distribution of variable scores/outliers)
  • Scatter plot (relationship b/w two continuous variables/outliers/linearity)
  • qq plot (checking for normality)

Analysis

Parametric

Independent Samples T-Test

Non-parametric

Chi-square test for independence

APA Table Set-up

Step 1: Dr. JJ Roth’s procedures (see workshop):

Step 2: Change SPSS default tables to APA: EditOptionsPivot TablesBrowse

Open Saved TablelookApplyOK

Assignment 14:Exercise importing SPSS data set

Worker respirable silica (quartz) exposures in µg/m3 as an 8-hour TWA are measured in three stone-working shops that perform similar work using similar equipment both with and without wet methods for dust suppression. Differences in the shops’ exhaust ventilation systems suggest there might be differences in worker exposures across shops.
TWA: time weight average. It is calculated based on 8 hr exposure (work day).
/ The 10 readings are taking at different times during both processes; wet or dry.
For each time, the shops are doing the same job (using the same equipment and methods), but have differences in the ventilation rate/system. Ventilation will affect the level of silica the workers maybe exposed to.

Reading In Data Set: exercise2.sav

Method / Shop A / Shop B / Shop C
Dry / 12 / 198 / 25
90 / 60 / 15
25 / 51 / 45
18 / 116 / 70
55 / 135 / 10
50 / 59 / 18
35 / 69 / 20
77 / 74 / 30
145 / 28 / 85
130 / 161 / 90
Wet / 7 / 19 / 13
36 / 31 / 8
13 / 36 / 21
10 / 26 / 29
24 / 49 / 6
22 / 42 / 10
17 / 34 / 10
32 / 35 / 15
53 / 37 / 34
49 / 55 / 36

Analysis

  1. Identify the independent and dependent variables and their format.
  2. Complete a descriptive analysis to show means and Standard deviation for all variables.
  3. Conduct an independent t-test to compare the 2 methods “Dry” and “Wet”

3.1Write the Null and Alternative hypotheses for the test.

  1. Conduct an ANOVA test (Univariate) to show which independent variables; method, shop and their interaction (method x shop) are significant.
  2. Write the Null and Alternative hypotheses for the test.
  3. Provide the analysis table
  4. Conduct a post-hoc analysis using Scheffe's AND Tukeys HSD.
  5. State your inferences and conclusions.
  6. Compare the AVONA results for the Method part with the t-test done earlier (for methods); are the results the same?

Page 1 of 7