Assessing Reliability and Convergent Validity in IBM SPSS

This section discusses the assessment of construct reliability and validity in IBM SPSS. It explains the concept of construct reliability and how it is measured using composite reliability and Cronbach's alpha. The guidelines for interpreting reliability values are also mentioned.

Construct Reliability Assessment

• Construct reliability refers to the extent to which a variable or set of variables is consistent in measuring what it intends to measure.

• Reliability is assessed through composite reliability and Cronbach's alpha.

• Composite reliability is calculated based on factor loadings, where the standardized vector loading for each item is summed, squared, and divided by the sum of error variance.

• Cronbach's alpha can be easily calculated using SPSS.

Interpreting Construct Reliability

• Nanali and Bernstein suggest 0.7 as a benchmark for modest or acceptable reliability.

• If all constructs have reliabilities greater than 0.7, it indicates that the measures used in the study are reliable and consistent.

Calculating Convergent Validity

This section focuses on calculating convergent validity, which is one form of construct validity. It explains the concept of construct validity and how it relates to measuring the intended construct. The importance of convergent validity in assessing multiple indicators measuring the same construct is discussed.

Construct Validity

• Construct validity measures how well selected items actually measure the intended construct.

• It helps determine if the items selected for measurement are accurately capturing the underlying construct.

Convergent Validity

• Convergent validity assesses whether multiple measures that should theoretically be related to each other actually converge or come together to measure the underlying construct.

• Indicators measuring the same concept should have a significant correlation with each other, ensuring unidimensionality of the construct.

• Convergent validity is assessed using Average Variance Extracted (AVE), which indicates how much of the indicator's variance can be explained by the latent variable.

The transcript does not provide specific steps or calculations for calculating convergent validity.

Formula for Convergent Validity

In this section, the speaker explains the formula for calculating convergent validity using the Average Variance Extracted (AVE).

Calculation of AVE

• The AVE is calculated by summing the squares of the factor loadings and dividing it by the number of items in the unobserved latent variable.

• The formula involves squaring all the factor loadings, summing them, and then dividing by n (the number of items in the latent variable).

• Lambda represents the factor loadings.

How to Calculate AVE in AMOS

• The speaker demonstrates a simple calculator to calculate AVE in AMOS.

• The factor loadings remain unchanged.

• Copying and pasting the loadings into the calculator makes them easily accessible.

Calculating AVE for Authentic Leadership Behavior

• The speaker copies and pastes the factor loadings for authentic leadership behavior into the calculator.

• The resulting AVE is slightly less than 0.5 (0.47).

• There is a suggestion to delete indicators if desired.

Composite Reliability as an Alternative

• If composite reliability (CR) for a particular indicator is greater than 0.7, deletion may not be necessary.

• If AVE is close to 0.5 and CR is greater than 0.7, it indicates validity.

Calculating AVE for Behaving Ethically and Life Satisfaction

• The speaker repeats the process for behaving ethically and life satisfaction constructs.

• Behaving ethically has an AVE of 0.65.

• Life satisfaction has an AVE of 0.67.

Overall Validity of Constructs

• AVE and composite reliability values have been calculated for all constructs.

• All constructs are considered reliable and valid.

Next Session: Discriminant Validity

The next session will focus on discriminant validity.

The summary has been written in English as per the given instructions.