Data Analysis

(with SPSS)

UiTM Cawangan Sarawak, Kampus Samarahan

Contents
 Data preparation (processing)
 The process
 Quantitative data analysis
 Descriptive statistical analysis
 Inferential statistical analysis
 Hypothesis testing process
• Qualitative data analysis
- Content analysis
The Research Process

(Sekaran, 2013)
Data Preparation
 Preliminary plan of data analysis
 Checking & Editing
 Structured/unstructured data
 Coding & Code book
 Qualitative/quantitative
 Transcribe/Data entry
 Data cleaning (consistency & missing)
 Data adjustment (weight/re-specification)
 Select the analysis strategy
 Non-statistical/Statistical
Treating Missing Responses
2-3 from >30 items are missing/left blank!

 To assign the midpoint of the item

 Allow computer to ignore the blank response
 Mean of the item from all respondents
 Mean of the related responses (of that respondent) for
items measuring the variable
 To give the missing response a random number within
the range for that scale
N.B.:
SE=Service Experience, US=User Satisfaction, RI=Repurchase Intention
PE=Previous Experience
 Service Environment

Figure 2: Relationships among Servicescape, Human Service,

Food Quality, Price, Youth Satisfaction and Youth Loyalty
Service Orientation in Design

Customer
Orientation
(CUS)

Competitor Service Orientation Client

Orientation of Designers Satisfaction
(CUS)

Inter-functional Demographic
Coordination Charateristics
(IC)
Goodness of Measure
Reliability Analysis with ITC, Split Samples (Voon, 2006)
Exploratory Factor Analysis (Kim et al., 2008)
Descriptive and Validity Analyses (Kim et al., 2008)
Example

SERVQUAL (Parasuraman et al., 1988)

Step 1: Definition of service quality (literature review and qualitative research)

Step 2: Identification of 10 dimensions making up the domains of the
service quality construct.
Step 3: Generation of 97 items for the 10 dimensions.
Step 4: Collections of expectations and perceptions data (survey)
Step 5: Scale purification:
i. Computation of coefficient alpha and item-to-total correlations for each
dimension.
ii. Deletion of items with low item-to-total correlations
iii. Factor analysis to verify dimensionality
iv. Reassignment of items and restructuring of dimensions
Step 6: Identification of 34 items representing 7 dimensions.
Step 7: Collections of expectations and perceptions data (using the 34 items)
Step 8: Evaluation and further purification (see step 5)
Step 9: Identification of more parsimonious, 22-item scale
Step 10: Evaluation of the scale reliability and factor structure pertaining to the
22 items, to verify the internal consistency and dimensionality.
Step 11: Assessment of the scale’s validity.
Multi-item SERVQUAL (Parasuraman et al., 1988)
SPSS for reliability analysis (LibQual 25 items)
SPSS for reliability analysis (LibQual 25 items)
SPSS for reliability analysis (LibQual 25 items)
SPSS for reliability analysis (LibQual 25 items)
SPSS for reliability analysis (LibQual 25 items)

To retain items
with Corrected
ITC >0.40

All the 25 items

of LibQual are
reliable.

They will be used

in the final
questionnaire
(if no EFA)
Data Analysis
 Descriptive
 Inferential
 Univariate or multivariate
 Nature of the data
 Metric (interval/ratio)
 Non-metric (nominal/ordinal)
 Number of sample groups
 Dependent or independent groups
Descriptive Data Analysis
1. Frequency distribution
2. Statistics associated with frequency distribution
i. Measures of location
ii. Measures of variability
iii. Measures of shape
Measure of location
• Mean
• Mode
• Median
Measure of variability
i. Variance
ii. Standard deviation
iii. Coefficient of variation
(ratio of standard deviation to mean expressed as %)
 Measure of shape
i. Skewness
– The tendency of the deviations from the mean to be larger in one
direction than in the other. It can be thought of as the tendency for
one tail of the distribution to be heavier than the other.
ii. Kurtosis
– A measure of the relative peakedness or flatness of the curve
defined by the frequency distribution. The kurtosis of a normal
distribution is zero. If the kurtosis is positive, then the distribution
is more peaked than a normal distribution. A negative value means
that the distribution is flatter than a normal distribution.

Symmetric
distribution

Mean
Median
Mode
 Respondent Sex Familiarity Internet Attitude Toward Usage of Internet
 Number Usage Internet Technology Shopping Banking
 1 1.00 7.00 14.00 7.00 6.00 1.00 1.00
 2 2.00 2.00 2.00 3.00 3.00 2.00 2.00
 3 2.00 3.00 3.00 4.00 3.00 1.00 2.00
 4 2.00 3.00 3.00 7.00 5.00 1.00 2.00
5 1.00 7.00 13.00 7.00 7.00 1.00 1.00
 6 2.00 4.00 6.00 5.00 4.00 1.00 2.00
 7 2.00 2.00 2.00 4.00 5.00 2.00 2.00
 8 2.00 3.00 6.00 5.00 4.00 2.00 2.00
 9 2.00 3.00 6.00 6.00 4.00 1.00 2.00
 10 1.00 9.00 15.00 7.00 6.00 1.00 2.00
 11 2.00 4.00 3.00 4.00 3.00 2.00 2.00
 12 2.00 5.00 4.00 6.00 4.00 2.00 2.00
 13 1.00 6.00 9.00 6.00 5.00 2.00 1.00
 14 1.00 6.00 8.00 3.00 2.00 2.00 2.00
 15 1.00 6.00 5.00 5.00 4.00 1.00 2.00
 16 2.00 4.00 3.00 4.00 3.00 2.00 2.00
 17 1.00 6.00 9.00 5.00 3.00 1.00 1.00
 18 1.00 4.00 4.00 5.00 4.00 1.00 2.00
 19 1.00 7.00 14.00 6.00 6.00 1.00 1.00
 20 2.00 6.00 6.00 6.00 4.00 2.00 2.00
 .
 30 2.00 6.00 6.00 6.00 4.00 2.00 2.00
Cross-tabulation
(for non-metric/categorical data)

 It describes two or more

variables simultaneously. Internet Gender
 It results in tables that reflect usage Total
the joint distribution of two Male Female
or more variables with a Light 5 10 15
limited number of categories Heavy 10 5 15
or distinct values. Total 15 15 30
Hypothesis Testing
 Hypothesis is a declarative statement that is not tested yet.
 Null hypothesis (Ho)
 A statement in which no difference or effect is expected.
Example, Ho : µ ≤ 4.0
 Alternate hypothesis (H1)
 A statement that some difference or effect is expected.
Example, H1: µ > 4.0
 In marketing, the Ho is formulated in such a way that its rejection
leads to the acceptance of the desired conclusion.
 One-tailed test (directionally).
Example, Ho : µ ≤ 4.0 and H1: µ > 4.0
 Two-tailed test (not directionally)
Example, Ho : µ = 4.0 and H1: ≠ 4.0
 Hypothesis Testing Process
Formulate Ho and H1

Select appropriate test

Choose level of significance (α)

Determine critical value of test statistics *

Determine whether it falls into rejection area **

Determine whether Ho is rejected

Draw conclusion
* Determine probability
** Compare probability with α
 Univariate Technique

Metric data Non-metric data

1 Sample ≥2 Samples 1 Sample ≥2 Samples

• t test Unrelated samples • Frequency Unrelated samples
• Z test • 2-Group t test • Chi-square • Chi-square
2-Group F test • K-S • Man-Whitney
• Z test • Runs • Median, K-S
• One-way ANOVA • Binomial Related samples
Related samples • Sign, Wilcoxon
• Paired t test McNemar
Chi-square
 Multivariate Technique

Dependence techniques Interdependence techniques

1DV ≥2 DVs Variable Interobject

Interdependence Similarity
• Cross-tab.(>2Vs) • Multivariate Analysis • Factor • Cluster analysis
• Analysis of Variance of Variance &
Covariance
analysis • Multidimensional
& Covariance
• Multiple regression • Canonical correlation scaling
• 2-Group Discriminant • Multiple Discriminant
Analysis Analysis
• Conjoint Analysis
(Malhotra, 2006)
Example 1 Compare a mean
Suppose we wanted to test the hypothesis that the mean
familiarity rating exceeds 4.0 with α= 0.05.
Ho : µ ≤ 4.0
H1 : µ > 4.0
t = (X - )/sX
sX = s/ n
t = (4.724-4.0)/0.293 = 2.471

The degrees of freedom for the t statistic (n - 1 =28). From Statistical Table,
the probability of getting a more extreme value than 2.471 is less than 0.05
(Alternatively, the critical t value for 28 degrees of freedom and a
significance level of 0.05 is 1.7011, which is less than the calculated value).
Therefore, the null hypothesis is rejected.
The conclusion is, familiarity level exceeds 4.0.
Example 2 Testing association
 The chi-square statistic (χ²) is used to test the statistical
significance of the observed association in a cross-tabulation. The
expected frequency for each cell can be calculated by using a simple
formula:
n rn
fe = n c
where nr = total number in the row
nc = total number in the column
n = total sample size
df = (r - 1) x (c -1)
H0: No association between the two variables
H1: There is an association between the two variables
Example 3 Compare means

The hypothesis testing is to find out whether there is a

significant difference between the mean satisfaction of male
(µ1) and female (µ2) customers.

Step 1: Ho : µ1 = µ2
H1 : µ1 ≠ µ2

Step 2: To use t test

Step 3: To use α = 0.05
Step 3: P = 0.002 (obtained from SPSS output)
Step 4: Reject Ho
Step 6: There is a significant difference between the
mean satisfaction of male and female customers.
 Example 4 Regressions

Interpretations of Results
Table 1: User perception of the Library Service Quality
Dimension Mean Score Std. Dev. t-value (compare with 4.0

Service 4.7302 1.2183 13.350*

Personal Control 4.6718 1.0043 17.082*
Info. Access 4.6279 0.9900 14.183*

Place 5.1572 1.0200 25.216*

* Significant at α=0.05

The user perceptions of all the service quality dimensions are

significantly higher than the midpoint (i.e. 4.0). These imply that the
Users were satisfied with the service performance of the Library. The
Information Access dimension was perceived to be the least favourable
whereas the most favourable was Place.
Interpretations of Results
Table 1: Regression Model (Overall Quality & Library SQ)
Library quality dimensions Standardized
Coefficients t Sig.
Constant 2.113 .036
Information Access .110 1.487 .138
Service .266 3.971 .000
Place .239 3.130 .002
Personal control .127 1.591 .113
Note: Dependent variable = Overall Quality, F = 42.83, R² = 0.41

The F-value for the omnibus test is 42.83 (p = 0.001) indicates that the linear
regression model indeed an appropriate model to explain the association between
the evaluation on overall quality of the library services and the four dimensions
of library service quality. R² of 0.41 indicates that 41% of respondents’ variation
in assessing the overall quality of the Library can be statistically attributed to
their perception on Service and Place.
Respondent profile
Measurement Reliability Analysis
Descriptive Analysis
Qualitative Data Analysis
Content Analysis

- Read the content of the writings

- Summarize the writings to form key words/phrases
- Sort and categorise the words/phrased into groups or
dimensions/components
- Assign code to each key word/phrase and also code
each of the categories
- Frequency distribution and % can be done
Category and Item
Items
Item 1.1
Item 1.2
Item 1.3 Dimension 1
Item 1.4
Variable 1
Item 2.1
Item 2.2
Dimension 2
Item 2.3
Item 2.4

Item 2.5
Let us read more refereed journal
articles and do more research
for research excellence.

FREE
FREE journal articles from

www.ijsmssarawak.com