6. RUNNING STATISTICAL ANALYSIS
1. DESCRIPTIVE STATISTICS
To run the descriptive statistics in the IBM SPSS, these choices were made:
Analyze > Descriptive Statistics > Descriptives
From the test, the output test is generated
2. SCATTER PLOT
A simple scatter plot was generated from the data in order to test the assumptions of' Linearity' and 'Homoscedasticity'
The next test to be run is the Simple Linear Regression, which is the main inferential test.
The steps taken are: Analyze > Regression > Linear...
The variables are CIUS_total scores as the Independent Variable and OPLIS_total as the Dependent Variable. This is to see whether 'compulsive internet use' scores will predict the 'online privacy literacy' scores. Below are the steps:
Output:
From the output, a few assumptions could be checked which are the:
a) Outliers -
Based on the residual statistics table, the minimum value for the standardized
residual is -1.905 and the maximum value is 1.541 which did not exceed -3.29 or 3.29
respectively. This shows that there are no outliers.
b) Independent errors –
Using the Durbin-Watson test, independence of error was examined. In the Model Summary
table, the Durbin Watson test indicated 1.573 which indicated that the assumption of
independence of observation has been met.
c) Normality -
This could be checked by examining the P-P Plot. The dots generally lined up close to the line
which indicated normality. Next, the histogram also show a bell shape and the graph of
'Standardized Residual' against the 'Predicted Value' also showed no pattern which suggest
normality.
Analyzing the Linear Regression
A few information could be obtained from the output tables which are:
a) Variables Entered/Removed box:
The only predictor model is the total score of CIUS.
b) Model Summary table:
- R-value = 0.095. This suggests a strong correlation between the variables.
- R-square = 0.009. This means that 0.9% of the variance in the total score of OPLIS was
predicted by the total score of CIUS.
c) ANOVA
The result from the ANOVA output table is p = 0.559 which indicated that our result is not
statistically significant. Thus, the regression model showed that the compulsive internet use
did not predict the online privacy literacy.
d) Coefficients
- Unstandardized Coefficients (B) = 0.018
- Standardized Coefficients (β) = 0.095
- Sig. = 0.559
Compulsive internet use (total score of CIUS) did not contribute statistically significantly to
online privacy literacy (total score of OPLIS).
To analyze the realiblity test which is the Cronbach's Alpha, the steps below are taken for each scale:
a) The Cronbach Alpha for this scale is 0.933 indicating an excellent internal consistency.
b) The Cronbach Alpha for this scale is 0.623 indicating a questionable reliability.
However, the Cronbach Alpha's value could be increased or decreased if any of the items is deleted. An example would be if the the first item in the Compulsive Internet Use Scale (CIUS) is deleted, then the Cronbach Alpha's value would decrease to 0.928.
To run the descriptive statistics in the IBM SPSS, these choices were made:
Analyze > Descriptive Statistics > Descriptives
From the test, the output test is generated
2. SCATTER PLOT
A simple scatter plot was generated from the data in order to test the assumptions of' Linearity' and 'Homoscedasticity'
The results from the simple scatter plot is as below:
To find the linearity of the data, the graph was double-clicked in the SPSS output and 'Add Fit Line at Total' was clicked.
The results from the simple scatter plot showed that the data was linear as it had a straight line. Thus, assumption of linearity has been met. Based on this graph, the homoscedasticity assumption is also met as the variance of errors in the relationship between the variables are constant and produces no pattern.
3. SIMPLE LINEAR REGRESSION
The next test to be run is the Simple Linear Regression, which is the main inferential test.
The steps taken are: Analyze > Regression > Linear...
The variables are CIUS_total scores as the Independent Variable and OPLIS_total as the Dependent Variable. This is to see whether 'compulsive internet use' scores will predict the 'online privacy literacy' scores. Below are the steps:
Output:
The graph was double-clicked in the SPSS output and 'Add Fit Line at Total' was clicked.
From the output, a few assumptions could be checked which are the:
a) Outliers -
Based on the residual statistics table, the minimum value for the standardized
residual is -1.905 and the maximum value is 1.541 which did not exceed -3.29 or 3.29
respectively. This shows that there are no outliers.
b) Independent errors –
Using the Durbin-Watson test, independence of error was examined. In the Model Summary
table, the Durbin Watson test indicated 1.573 which indicated that the assumption of
independence of observation has been met.
c) Normality -
This could be checked by examining the P-P Plot. The dots generally lined up close to the line
which indicated normality. Next, the histogram also show a bell shape and the graph of
'Standardized Residual' against the 'Predicted Value' also showed no pattern which suggest
normality.
Analyzing the Linear Regression
A few information could be obtained from the output tables which are:
a) Variables Entered/Removed box:
The only predictor model is the total score of CIUS.
b) Model Summary table:
- R-value = 0.095. This suggests a strong correlation between the variables.
- R-square = 0.009. This means that 0.9% of the variance in the total score of OPLIS was
predicted by the total score of CIUS.
c) ANOVA
The result from the ANOVA output table is p = 0.559 which indicated that our result is not
statistically significant. Thus, the regression model showed that the compulsive internet use
did not predict the online privacy literacy.
d) Coefficients
- Unstandardized Coefficients (B) = 0.018
- Standardized Coefficients (β) = 0.095
- Sig. = 0.559
Compulsive internet use (total score of CIUS) did not contribute statistically significantly to
online privacy literacy (total score of OPLIS).
3. Reliability Test (Cronbach's Alpha)
To analyze the realiblity test which is the Cronbach's Alpha, the steps below are taken for each scale:
Analyze > Scale > Reliability Analysis
a) The Compulsive Internet Use Scale (CIUS)
b) The Online Privacy Literacy Scale (OPLIS)
a) The Compulsive Internet Use Scale (CIUS)
b) The Online Privacy Literacy Scale (OPLIS)
From the output, we can extract the following data:
a) The Cronbach Alpha for this scale is 0.933 indicating an excellent internal consistency.
b) The Cronbach Alpha for this scale is 0.623 indicating a questionable reliability.
However, the Cronbach Alpha's value could be increased or decreased if any of the items is deleted. An example would be if the the first item in the Compulsive Internet Use Scale (CIUS) is deleted, then the Cronbach Alpha's value would decrease to 0.928.
Comments
Post a Comment