Structure of Questionnaire : measure 5 attributes, each with 4-5 factors having 3-5 item statements on Likert scale . Have 757 responses. No missing data.
Prime questions :
1. Can there be outliers for likert-type scales. (individual box plots show outliers, removing them is not solving problem of Normality)
Method best suited - Mahalanobis / Cook's distance or box plots
2. Request to advise if normality needs to be checked
a. item-wise (88 statements),
b. factor-wise (as developed with EFA) - 5 constructs with 4-5 factors each or
c. overall constructs (5 constructs).
Dear Hemalatha,
First of all, you have to determine what statistical method you have to apply to analyze your data in a correct manner (in a valid and reliable way). I mean that if you want to compare many groups for many dependent variables the options that you have to do that and the requirements for all this method it's not the same that those you need for a Factorial Analysis, for example.
Generally, when the researcher wont to apply parametric statistical methods they test multivariate normality with Mardia Kurtosis and Mahalanobis distances to detect outliers. In any case, when you see that neither multivariate nor univariate normality meet, ore that the item cannot be considered numerical ore continues variables, to mention typical situations, you have two main options: a) doing something to meet this criteria with your data (typically the data were transformed, or atypical case eliminated); b) ore apply another statistical method, often called robust methods that be in line with your data (I recommend this last option).
Thanks Mr. Kanter. Can you also suggest how I go about this (my 2nd questions)
Request to advise if normality needs to be checked for liker-scale type data (1 to 5) a. item-wise (88 statements), b. factor-wise (as developed with EFA) - 5 constructs with 4-5 factors each or c. overall constructs (5 constructs).
. The overall data shows outliers (in box plot and also using skewness & kurtosis scores) , when scanned for (a) above.
. With grouping based on dependant variable, normality is met for a few groups. .
. One very large group (with almost 50% respondents) has high skewness and kurtosis value - statistic /SE is outside the range of +or - 1.96
I am afraid transforming will affect my findings.
Planned statistical analysis are :-
Factor analysis
develop a SEM equation for entire set, groups based on dependent variable
compare different group for many dependent variables - ANOVA, T-tests
Requesting guidance, please
- Badges
- Science topic
- Similar topics
- Agricultural Science
- Horticulture