Unit 4: Data preparation and analysis using SPSS
Marketing Research notes · PTU syllabus (MBA 302-18)
On this page
Unit summary
SPSS turns survey responses into evidence for decisions. This unit covers understanding SPSS, creating SPSS data sheets and entering data, basic descriptive statistics, assessing reliability and coefficient alpha, scale refinement and item analysis, correlation analysis, factor analysis and regression analysis.
After this unit you can
- Set up an SPSS data file and enter survey data
- Run descriptive statistics and reliability analysis
- Refine scales through item analysis
- Apply correlation, factor and regression analysis
PTU syllabus topics
- Understanding SPSS
- creating SPSS sheets and data entry
- basic descriptive statistics
- assessing reliability and coefficient alpha
- scale refinement and item analysis
- correlation analysis
- factor analysis
- regression analysis
Topic 1
Understanding SPSS
SPSS has two main views: the Data View (rows are cases, columns are variables) and the Variable View (where each variable's name, type, label, values and measure are defined). Output appears in a separate Output Viewer. IBM SPSS Statistics is menu-driven software for managing and analysing data, widely used in marketing research.
- Windows: Data Editor (Data View and Variable View), Output Viewer, Syntax Editor, Chart Editor.
- Menus: File, Edit, View, Data, Transform, Analyze, Graphs, Utilities.
Example
A typical use: a researcher enters 300 survey responses in SPSS, runs frequencies for demographics, a reliability test on the satisfaction scale and a correlation between satisfaction and loyalty — all through menus, without writing code.
Topic 2
Creating SPSS data sheets and entering data
- 1
Prepare a codebook
Variable names, codes, missing-value codes
- 2
Define variables in Variable View
Name, type, label, values, missing, measure
- 3
Enter data in Data View
One row per respondent
- 4
Verify entries
Frequencies to spot invalid codes
- 5
Recode and compute
Reverse-code negative items; compute scale totals or means
- 6
Save the .sav file
- Value labels: 1 = Strongly disagree … 5 = Strongly agree; 1 = Male, 2 = Female.
- Reverse coding: Transform → Recode into Different Variables (1→5, 2→4, 3→3, 4→2, 5→1) for negatively worded items.
Example
Coding a question: for "Gender: Male / Female / Other", define a variable gender, type Numeric, with value labels 1 = Male, 2 = Female, 3 = Other, measure Nominal. Each respondent's answer is then entered as a number in Data View.
Exam tip
Prepare the codebook before entering data, and set a missing-value code (e.g. 99) for unanswered questions.
Topic 3
Basic descriptive statistics in SPSS
| Analysis | SPSS menu | Output used |
|---|---|---|
| Frequencies | Analyze → Descriptive Statistics → Frequencies | Counts, percentages, bar charts |
| Descriptives | Analyze → Descriptive Statistics → Descriptives | Mean, SD, minimum, maximum |
| Cross-tabulation | Analyze → Descriptive Statistics → Crosstabs | Row and column percentages, chi-square |
| Explore | Analyze → Descriptive Statistics → Explore | Box plots, normality tests |
Example
Frequencies show 62% of 400 respondents shop online at least weekly; Crosstabs reveal 78% of those aged 18–25 do so against 41% of those over 40.
Topic 4
Assessing reliability in SPSS
- Menu: Analyze → Scale → Reliability Analysis; move the scale items; Model = Alpha; under Statistics tick Item, Scale, Scale if item deleted, Inter-item correlations.
- Output: Cronbach's alpha; Item-Total Statistics — corrected item-total correlation and alpha if item deleted.
Example
Reading the output: for a six-item trust scale, SPSS shows Cronbach's alpha = 0.78. The "Alpha if Item Deleted" column shows that removing item 4 would raise alpha to 0.84, so item 4 may be poorly worded.
Exam tip
Reverse-code negatively worded items before running reliability analysis.
Topic 5
Scale refinement and item analysis
- Item analysis checks how well each item contributes to the scale.
- Rules of thumb: drop items with corrected item-total correlation below 0.3, or whose deletion substantially raises alpha; check inter-item correlations (0.3 to 0.8 is healthy).
- Iterate: remove one weak item at a time, rerun reliability, then confirm with factor analysis.
Example
A 6-item service-quality scale has α = 0.68; item 4 has item-total correlation 0.12 and "alpha if deleted" 0.79. Dropping item 4 raises reliability to an acceptable level.
Topic 6
Correlation analysis in SPSS
- Menu: Analyze → Correlate → Bivariate — choose Pearson (interval or ratio data) or Spearman (ordinal data); two-tailed test.
- Read: correlation coefficient (r), Sig. and N; flagged correlations are significant.
- Partial correlation: Analyze → Correlate → Partial — relationship controlling for a third variable.
Example
Interpreting r: SPSS reports r = 0.62, p < 0.01 between service quality and customer loyalty scores. This is a moderately strong, positive and significant relationship — but it does not prove that quality causes loyalty.
Exam tip
Rule of thumb: |r| < 0.3 weak, 0.3–0.7 moderate, > 0.7 strong.
Topic 7
Factor analysis
Two tests check whether the data suit factor analysis: the KMO measure (should be above 0.6) and Bartlett's test of sphericity (should be significant). Factors are usually extracted by principal component analysis, kept if their eigenvalue is above 1, and rotated (varimax) for clearer interpretation. Factor analysis reduces many correlated items into a few underlying factors (e.g., 20 attitude statements into four dimensions).
- 1Check suitability
KMO above 0.6; Bartlett's test significant
- 2Extract factors
Principal components; eigenvalue above 1 or scree plot
- 3Rotate
Varimax (orthogonal) for clearer loadings
- 4Interpret loadings
Items loading above 0.5 on a factor
- 5Name factors and compute factor scores
- Menu: Analyze → Dimension Reduction → Factor.
- Output: communalities, total variance explained, rotated component matrix.
- Exploratory vs confirmatory: EFA discovers structure; CFA (in AMOS or similar) tests a hypothesised structure.
Example
Reducing 20 statements: factor analysis of 20 statements about a mobile network groups them into four factors — call quality, data speed, price and customer care. These four factors are then used in a regression on overall satisfaction.
Exam tip
Remember the checks in order: KMO, Bartlett's test, eigenvalue > 1, rotation, naming factors.
Topic 8
Regression analysis in SPSS
- Menu: Analyze → Regression → Linear — dependent variable (e.g., overall satisfaction), independent variables (factor scores for product, service, price).
- Read: R² and adjusted R² (Model Summary), F and Sig. (ANOVA), unstandardised B and standardised Beta with Sig. (Coefficients); VIF for multicollinearity.
Example
Regression of satisfaction on three factors gives R² = 0.54; Beta: service 0.46, product 0.31, price 0.12 (not significant). Service quality is the strongest driver of satisfaction.
Key terms
- Variable View
- SPSS window for defining variables
- Reverse coding
- Recoding negatively worded items so high scores mean the same direction
- Corrected item-total correlation
- Correlation of an item with the rest of the scale
- KMO
- Measure of sampling adequacy for factor analysis
- Standardised Beta
- Regression coefficient comparable across variables
Quick revision
- SPSS windows; codebook; Variable View and Data View.
- Frequencies, Descriptives, Crosstabs, Explore.
- Reliability: Analyze → Scale → Reliability Analysis; alpha if item deleted.
- Item analysis: drop items with item-total correlation below 0.3.
- Correlation (Pearson, Spearman); factor analysis (KMO, eigenvalue, varimax); regression (R², Beta, VIF).
Important exam questions
Practice questions written to the PTU exam pattern for this unit's syllabus: short answers (Section A style) and long answers (Sections B and C style).
Short-answer questions
- Q1.What is the Variable View in SPSS?
- Q2.Why are negatively worded items reverse-coded?
- Q3.Name the SPSS menu for reliability analysis.
- Q4.What is "alpha if item deleted"?
- Q5.What does a KMO value indicate?
- Q6.What does standardised Beta show in regression?
Long-answer questions
- Q1.Explain how to create an SPSS data file and enter survey data.
- Q2.Explain reliability assessment and scale refinement in SPSS.
- Q3.Explain the steps of factor analysis and interpretation of its output.
- Q4.Explain correlation and regression analysis in SPSS with an example.
Stuck on this unit?
Message SBS on WhatsApp for help with Marketing Research, or to ask about studying MBA at Synetic.
