I need to know from where we can get summation and how to get the Likert scale value?Please explain with an example .
∑-X
N
Where: X = mean, Σ = Summation, X = Value likert scale, n = Number of item or variables.
Decision point:
Using 3.0 as the mean decision point, any item that has its means score (XS) up to 3.0 and above is
regarded as a strong factor. In computing the means score (XS) of each item, the frequency was multiplied with its appropriate likert value and the sum deviated weight with the items, as computed with the below equation.
∑X
2. X = -----------------------
nr
Where xs = mean score
Σ = summation
nr = number of respondents to an item.
Thank you
Hello Beg,
First I think you have to explore the nature of your scale:
If the scale is symmetric you can use the average as a measure of location, otherwise must use robust measures such as the median.
To split the scale then you can use for example the method of extreme groups: case symmetrical distribution, cuts are given as mean value + -0.25 standard deviation; otherwise median + -º0,25 inter quartile range.
To "get the summation" is equal to make an index: ex use FA
First you need to test whether the items in your scale are correlated highly enough to justify adding them together. The standard statistic for assessing that is Cronbach's alpha. If your alpha is sufficiently high, then you can either add the items together or create a simple average by dividing the total score by the number of items.
I Think David is right but instead of a simple average, for me it is better to save the scores in FA and use it as an index of that scale.
The main difference between using Cronbach's alpha and using scores from an Exploratory Factor Analysis is that Cronbach's alpha does not weight the variables (i.e., adding them together effectively weights each item at 1.0). In contrast, factor scores attempt to take into account the variation in the factor loadings, so that items with higher loadings get higher weights.
In practice, factor scores are less straightforward than that. So, unless there is a lot of variation in your loadings, the score that you get from a simple alpha will be highly correlated with any factor score you calculate.
- Badges
- Science topic