Online Archive: Measurement Model Validation for Attitudes and Values
Before the development of the integrated choice and latent variable (ICLV) model, a confirmatory factor analysis was carried out in order to test the reliability and validity of the measurement models for both the attitude and the value constructs. The indicators of the final model for the attitude and value constructs are presented in Table 1, and the construct reliability measures are reported in Table 2.
Construct Measurements / Mean / Std. Dev. / ReliabilityAttitudes (new scales based on Johansson et al. 2006)
Flexibility (3 measures)
I1 - That a means of transportation is available right away is … / 4.3 / 0.77 / 0.49
I2 - That a means of transportation can be used spontaneously and without planning is … / 4.3 / 0.82 / 0.42
I3 - That a means of transportation reaches its final destination without a detour or change is … / 4.2 / 0.86 / 0.32
Convenience and Comfort (3 measures)
I4 - That a means of transportation is exceedingly convenient and comfortable is … / 3.6 / 0.97 / 0.38
I5 - That using a means of transportation is stress free and relaxed is … / 3.9 / 0.86 / 0.23
I6 - That you don’t have to worry about anything while using a means of transportation is … / 3.5 / 1.07 / 0.22
Ownership (1 measure)
I7 - That you own the means of transportation is … / 3.6 / 1.22 / 0.80
Personal Values (based on the Portraits Value Questionnaire by Schwartz et al. 2001)
Power (2 measures)
I8 - She/he wants to be the one who makes decisions.
She/he likes to be the leader. / 3.6 / 1.33 / 0.63
I9 - It is important to her/him to be in charge.
She/he wants people to do what she/he says. / 3.3 / 1.32 / 0.53
Hedonism (3 measures)
I10 - She/he seeks every chance she/he can to have fun.
It is important to her/him to do things that give pleasure. / 4.6 / 1.09 / 0.52
I11 - She/he really wants to enjoy life.
Having a good time is very important to her/him. / 4.6 / 1.13 / 0.46
I12 - Enjoying life’s pleasures is important to her/him.
She/he likes to spoil herself/himself. / 4.5 / 1.25 / 0.37
Security (4 measures)
I13 - It is very important to her/him that her/his country be safe.
She/he thinks the state must be on watch against threats. / 4.5 / 1.24 / 0.39
I14 - It is important to her/him to live in secure surroundings.
She/he avoids anything that might endanger her/his life. / 4.3 / 1.22 / 0.28
I15 - Having a stable government is important to her/him.
She/he is concerned that the social order be protected. / 4.5 / 1.15 / 0.27
I16 - It is important to her/him that things be organized and clean.
She/he really does not like things to be a mess. / 4.5 / 1.35 / 0.25
Table 1: Measurement indicators for different attitudinal and value constructs
Building on the work of Vredin Johansson et al. (2006), we developed measurement indicators for the three attitude dimensions: flexibility, convenience/comfort, and ownership/safety. Initially, the survey included the three indicators for attitude dimension. Unfortunately, the measurement model for attitudes toward ownership/safety did not perform as expected, perhaps because the three original Likert-scale questions were a mixture of personal safety and traffic safety attitudes in mode choice. Since Johansson et al. (2006) also reported that the differences in traffic safety across modes are negligible, we decided to employ only the question about ownership and labeled the attitude dimension “ownership.” In contrast to the abstract concepts of flexibility and convenience/comfort, attitude toward ownership of a means of transport can be considered concrete from a commuter’s perspective. Therefore, measuring this construct with a single, content-valid item is acceptable, although not optimal (e.g., Rossiter 2002). As a safeguard against moderate measurement errors, we assumed a reliability of 0.8 and fixed the error variance of the item accordingly.
Our final confirmatory factor model for attitudes toward mode choice and values was estimated with the robust, weighted least squares means and variance adjusted (WLSMV) estimator using Mplus software (Muthén and Muthén 1998–2007, Version 5.1). Goodness-of-fit statistics for this model indicate an acceptable overall fit to the data (=131.28, df=57, NNFI=0.91, CFI=0.90, RMSEA=0.05, weighted RMSR=0.90). Convergent validity is established by statistically significant factor loadings with t-statistics ranging from 5.07 to 10.14. Completely standardized factor loadings range from 0.47 to 0.70 for the attitude measures and from 0.50 to 0.79 for the value measures. Except for the factor of convenience/comfort, all construct reliabilities (see Table 2) are above the recommended threshold of 0.60 (Bagozzi and Yi 1988). The squared correlation between the attitude constructs of flexibility and convenience/comfort is larger than the average variance extracted for both factors, indicating a possible violation of discriminant validity (Fornell and Larcker 1981; see Table 2): However, a chi-square difference test proved that the correlation between flexibility and convenience/comfort is unequal to one (=54.31, df=5, p=0.000), thus providing support for the discriminant validity of both constructs.
Construct / 1 / 2 / 3 / 4 / 5 / 61. Flexibility / 0.67, 0.41
2. Convenience/Comfort / 0.45 / 0.53, 0.27
3. Safety/Ownership / 0.13 / 0.09 / 0.80, 0.80a
4. Power / 0.06 / 0.07 / 0.01 / 0.78, 0.58
5. Hedonism / 0.02 / 0.13 / 0.02 / 0.03 / 0.70, 0.44
6. Security / 0.05 / 0.31 / 0.08 / 0.03 / 0.03 / 0.62, 0.42
Notes: Entries on the diagonal represent (1) Bagozzi’s (1980) construct reliability and (2) Fornell and Larcker’s (1981) average variance extracted . The non-diagonal elements are squared correlations among the constructs.
All intercorrelations are significant at 5% level.
a Single-indicator construct with fixed measurement error values
Table 2: Construct reliability and validity measures
It must be acknowledged that some of the construct reliabilities (e.g., comfort/convenience) are below the recommended thresholds and represent a weakness of the model. Unfortunately, this weakness is also present in other papers in the field that employ attitude constructs (e.g., Johansson et al. 2005, 2006) and indicates that future research to develop more reliable scales on attitudes toward transport modes is needed.
1