Understanding Indicator and Construct Reliability.​

In the realm of construct reliability assessment, the initial step involves scrutinizing the extent to which each indicator’s variance is elucidated by its corresponding construct. This serves as a pivotal aspect of reflective measurement model evaluation. To gauge the reliability of each indicator, we examine the extent to which its variance is explained by the construct. This is a manifestation of indicator reliability. 

The initial step in reflective measurement model assessment involves scrutinizing indicator reliability. Within reflective models, indicators point toward the latent construct, as depicted by the arrows. For a reflective model, characterized by arrows pointing towards the indicators, and reflective constructs, denoted by arrows aligning with the indicators, understanding the reliability of indicators is paramount.

To compute this, we square the loadings obtained from earlier discussions, which represent the bivariate correlation between the indicator and the construct. An indicator loading exceeding 0.708 is recommended, indicating that over 50% of the indicator’s variance is explained by the construct, thus establishing acceptable indicator reliability.

Indicator reliability, a concept rooted in the evaluation of how much variance in each indicator is explicated by the latent construct, necessitates the computation of an indicator’s explained variance. Indicator reliability is the square of a standardized indicator’s indicator loading. It represents how much of the variation in an item is explained by the construct and is referred to as the variance extracted from the item. The loading represents the bivariate correlation between the indicator and the construct, underscoring the essence of the indicator’s reliability.

Indicator loadings above 0.708 are recommended, since they indicate that the construct explains more than 50 percent of the indicator’s variance, thus providing acceptable indicator reliability. The calculation of indicator reliability can be performed by squaring the values in the indicator/factor loading. The squared value >= 0.50 indicates the indicator reliability is established. To summarize, if the indicator loading attains or exceeds 0.708, it implies that 50 percent of the indicator’s variance is accounted for by the construct, thereby affirming acceptable indicator reliability.

In case you obtain weaker indicator loadings (< 0.708) . Do not simply delete the indicators when their loading is below 0.70, researchers should carefully examine the effects of indicator removal on other reliability and validity measures. Generally, indicators with loadings between 0.40 and 0.708 should be considered for removal only when deleting the indicator leads to an increase in the internal consistency reliability or convergent validity above the suggested threshold value. 

Another consideration in the decision of whether to delete an indicator is the extent to which its removal affects content validity. Moreover, content validity assumes significance in deciding whether to retain or eliminate an indicator. Items affecting content validity warrant careful scrutiny, and indicators with very low loadings (below 0.40) merit exclusion from the measurement model.

In conclusion, a meticulous evaluation of indicator reliability forms an integral part of reflective measurement model assessment. It involves a nuanced understanding of loadings, thresholds, and the interplay between constructs and indicators. For further insights into reliability considerations, additional resources can be explored.