Hi, I get the idea that loading values>.70 indicate that the construct explains more than 50 per cent of the indicator’s variance, thus providing acceptable item reliability. My question, What contributing to the lower loading values? how to relate the collected survey data with a lower loading value? Please help me to understand this?
Smaller standardized loadings often indicate that an item does not reflect the intended factor well (i.e., pointing to a validity problem rather than a reliability problem). The item may measure other/additional factors or contain method or item-specific variance.
Also, small loadings can indicate a lack of reliability (a lot of item variance may be due to random error), but this usually does not explain the entire issue. Usually, small loadings indicate that the inter-item correlations between the items pertaining to the factor are rather modest. Once again, this usually indicates multidimensionality of the items (that the items are not homogeneous in measuring the same dimension).
Practical consequences may be that a researcher has to delete or replace items or revise the scale(s) in some other way (e.g., by adding new, more homogeneous items or by rewording/rephrasing items).
Loading values in factor analysis represent the strength of the relationship between an indicator (survey question) and the underlying latent construct. Higher loading values indicate a stronger association, suggesting that the indicator is a good representation of the construct. A loading value above 0.70 is often considered acceptable for good item reliability, indicating that the indicator explains more than 50% of the variance in the construct.
When you encounter lower loading values, it suggests that the indicator may not be strongly related to the construct. There are several reasons why this could happen:
To relate the collected survey data with lower loading values, it is crucial to assess the quality of the indicators and the construct being measured. Review the wording, clarity, and relevance of the questions. Consider whether any modifications or additional items might be necessary to improve the measurement of the construct. Additionally, explore the correlations between the indicators and other variables to identify potential patterns or inconsistencies.
It is also recommended to perform additional analyses such as examining the factor structure, assessing internal consistency (e.g., Cronbach's alpha), and conducting validity checks to further understand the relationships between the variables and the construct of interest.
Overall, lower loading values indicate that the indicator may not be as reliable or valid in measuring the intended construct. It is important to critically evaluate and refine your measurement instrument to ensure accurate and meaningful results.
Thank you so much Sir Christian Geiser ,and Sir Ma'Mon Abu Hammad for your explanation, I really appreciate.
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude