Basic Data Analysis
Tabulation
• Frequency table
• Percentages
A Typical Table
Gender
Female
Male
Missing
Total
Frequency Percentage
Valid %
100
= 100/150
= 100/145
45
= 45/150
= 45/145
5
= 5/150
150
= = (100+45)
(100+45+5)
/ 145
/150
Type of
Measurement
Type of
descriptive analysis
Nominal
Cross Tabs
Mode
Type of
Measurement
Type of
descriptive analysis
Ordinal
Rank order
Median
Type of
Measurement
Type of
descriptive analysis
Interval
Arithmetic mean
CROSS-TABULATION
• Analyze data by groups or categories
• Compare differences
• Percentage cross-tabulations
A Typical Cross-Tab Table
Gender X
ECommerce
Customer
Female
Customer
NonCustomer
Totals
100
50
150
Male
75
60
135
Totals
175
110
285
Data Transformation
• A.K.A data conversion
• Changing the original form of the data
to a new format
• More appropriate data analysis
• New variables
– Summated
– Standardized
Degrees of Significance
• Mathematical differences
• Statistically significant differences
• Managerially significant differences
Testing the Hypotheses
• The key question is whether we reject or
fail to reject the hypothesis.
• Depends on the results of the hypothesis
test
– If testing differences between groups, was the
difference statistically significant
– If testing impact of independent variable on
dependent variable, was the impact
statistically significant
• How the hypothesis was worded
Differences Between Groups
•
•
•
•
Primary tests used are ANOVA and MANOVA
ANOVA = Analysis of Variance
MANOVA = Multiple Analysis of Variance
Significance Standard:
– Churchill (1978) Alpha or Sig. less than or equal to
0.05
• If Sig. is less than or equal to 0.05, then a
statistically significant difference exists between
the groups.
Example
• Hypothesis: No difference exists
between females and males on
technophobia.
• If a statistically significant difference
exists, we reject the hypothesis.
• If no s.s. difference exists, we fail to
reject.
Example
• Hypothesis: Males are more technophobic
then females (i.e., a difference does exist)
• If a statistically significant difference
exists, and it is in the direction predicted,
we fail to reject the hypothesis.
• If no s.s. difference exists, or if females are
statistically more likely to be technophobic,
we reject the hypothesis.
Testing for Significant Causality
• Simple regression or Multiple regression
• Same standard of significance (Churchill 1978)
• Adj. R2 = percentage of the variance in the
dependent variable explained by the regression
model.
• If Sig. is less than or equal to 0.05, then the
independent variable IS having a statistically
significant impact on the dependent variable.
• Note: must take into account whether the impact
is positive or negative.
Example
• Hypothesis: Technophobia positively
influences mental intangibility.
• If a technophobia is shown to
statistically impact mental intangibility
(Sig. is less than or equal to 0.05),
AND.
• The impact is positive, we fail to reject
the hypothesis.
• Otherwise, we reject the hypothesis.