Professional Documents
Culture Documents
Reliability-Measures
Cronbach’s alpha
Apha=(k/k-1)(1-((sum of item var)/summated scale var))
Alpha=k*AV(ri)/(1+(k-1)*(AV(ri))
SPSS: Analyze->scaling->alpha
Data screening
Case Screening
Missing data in rows
use count-blank function in Excel, if blank in a questionnaire is more than
or equal to 15% (Hair et al., 2014) , discard the questionnaire.
Else replace with
Mean or median
Expectation maximization likelihood approach
regression imputation techniques
Graphical method-
Normal Q-Q graph- mapping the variable with the equivalent normal variable with same
parameters.
Histogram
Tests in SPSS: Kolmogorov-Smirnov and Shapiro-Wilk
H0::data is normal
P value must be greater than 0.05
Goto Analyse -> Descriptive -> frequency-> charts-> histogram (check show normal
curve on histogram)
Goto Analyse -> Descriptive -> Explore-> Plots-> check Histogram and normality plots
with tests
Cross Tabulation
Used to find relationship between two categorical
variables
Excel:
Using Pivot Table
SPSS:
Analyse-> descriptive->crosstab->select row and column variables
and press ok
Chi-Square Test
Test of goodness of Fit:
Null hypothesis: no significant diff. Between observed and
expected result: Ef=Of
P value >0.05
Test of Independence
H0: Two variables are independent and has no relation with
respect to impact on each other.
22–13
Factor Analysis
Communality: The extent to which an item correlates with all other
items. Higher communalities are better. If communalities for a particular
variable are low (between 0.0-0.4), then that variable may struggle to load
significantly on any factor.
Convergent validity: means that the variables within a single factor are
highly correlated. Regardless of sample size, it is best to have loadings
greater than 0.500 and averaging out to greater than 0.700 for each factor.
Discriminant validity: refers to the extent to which factors are distinct and
uncorrelated. The rule is that variables should relate more strongly to their own
factor than to another factor. If "cross-loadings" do exist (variable loads on
multiple factors), then the cross-loadings should differ by more than 0.2.
Correlations between factors should not exceed 0.7. A correlation greater than
0.7 indicates a majority of shared variance (0.7 * 0.7 = 49% shared variance)
Condition of best fit model
See chi square= must be insignificant
Cmin==chi square=p value should not be significant
Gfi (goodness of fit index)value >.9
Fit indexes (NFI(normed), RFI(relative),
IFI(incremental), CFI(bentlers comparative))>.9
RMSEA<.05
Go to estimates : Look for significant path coefficient
(P values<.05)
Look for model modification
The thresholds listed in the table below are from Hu
and Bentler (1999).
CFA
Reliability
CR (composite reliability) > 0.7
Convergent Validity
AVE > 0.5, CR>AVE
Discriminant Validity
MSV (max shared variance)< AVE; AVE>ASV (average shared
variance)
Square root of AVE greater than inter-construct correlations
A significant standardized residual covariance is one with an
absolute value greater than 2.58. Significant residual covariances
significantly decrease your model fit. Fixing model fit per the
residuals matrix is similar to fixing model fit per the modification
indices.
Convergent validity
Average variance extracted (AVE)
Where:
L = the standardized factor loading,
i = the number of the item.
convergent validity by examining the loading paths of all
items, which should be statistically significant and exceed
0.50 (Hair et al., 2010, p. 709). It should be noted that
reliability is also considered an indicator of good
convergent validity if CR is 0.7 or more.
Common Method Bias (CMB)