Quality of LV unstandardised computation

This forum is closed, and read-only.
Locked
silburn
PLS Junior User
Posts: 4
Joined: Mon Oct 22, 2012 6:46 am
Real name and title:
Location: Caribbean
Contact:

Quality of LV unstandardised computation

Post by silburn »

I expect that when my MV's are from a standardised scale like Likert, unstandardised LV scores for each case will be fairly close to the averaging on MV values.

When I run my rough checks, I am getting vast discrepancies. Eg the following LV scores unstandardised from 53 cases are listed.

ICOS(MV avg) ICOS (fr SmartPLS)
4.500 6.3185
4.667 5.7887
5.333 -1.2728
5.083 10.5586
5.583 3.9089
4.917 0.6819
5.583 -1.7692
5.833 11.4977
5.250 8.585
5.417 10.9784
4.500 -1.9397
3.917 2.5346
4.917 2.2356
5.917 5.7532
5.167 6.4968
5.083 4.0525
4.333 2.7217
4.750 -4.8526
4.500 3.0234
5.250 0.7384
5.000 13.5752
5.250 11.1464
4.333 7.658
4.750 -1.6361
5.917 15.7398
5.667 -5.9259
4.917 6.3316
6.500 -1.3146
4.333 4.6898
5.417 -5.3413
5.750 9.2126
4.000 5.0326
5.250 -8.2863
4.250 -0.2875
4.500 10.2452
5.167 -6.5026
4.250 -4.0465
5.167 -6.6662
5.333 20.2707
5.250 -3.593
4.833 2.5336
4.917 16.8636
4.167 2.2709
5.417 13.1566
5.417 11.6064
5.250 15.4356
5.333 5.7516
5.000 -4.9153
5.000 3.4813
5.500 7.9863
4.750 -3.1312
5.750 -3.2894
4.667 2.8922


Question: Why am I getting negatives in the SmartPLS LV scores and why are the scores greater than 7 if my likert scores are in the range 1 to 7 ?

This seem like a bug in the software to me.

Has anybody experienced this type of glitch before.

How do I get the software to consistently compute correctly ?
admin
PLS Senior User
Posts: 26
Joined: Sun Sep 11, 2005 8:09 pm
Real name and title:

Post by admin »

Do not uses SmartPLS with unstandardized data --> viewtopic.php?p=6129

The unstandardized LV scores for of the index values results are used for impact-performance matrix analysis (PLS-IPMA). You must ensure that all outer weights are positive. Otherwise, your LV scores fall out of the range of the scale of your input data. This is a consequence of the computation procedure for unstandardized LV scores for the PLS-IPMA. Some article which you can download here: viewtopic.php?t=1158

Höck, Claudia, Christian M. Ringle, and Marko Sarstedt. 2010. "Management of Multi-Purpose Stadiums: Importance and Performance Measurement of Service Interfaces." International Journal of Services Technology and Management 14 (2/3): 188-207.

Rigdon, Edward E., Christian M. Ringle, Marko Sarstedt, and Siegfried P. Gudergan. 2011. "Assessing Heterogeneity in Customer Satisfaction Studies: Across Industry Similarities and Within Industry Differences." Advances in International Marketing 22: 169-194.

Best
CR
Administration Team
silburn
PLS Junior User
Posts: 4
Joined: Mon Oct 22, 2012 6:46 am
Real name and title:
Location: Caribbean
Contact:

Post by silburn »

Appreciate the response.

I need to run some descriptive stats on my data outside of PLS

a. mean, mode, median, ranges
b. skewness and kurtosis
c. bivariate correlations (construct to construct)

Where do I get the estimates for the unstandardised LV's for normality tests (skewness/kurtosis) and for bivariate analysis ? Di I use the values from SmartPLS's tables for unstandardised LV ? Or do I do an averaging of the MV's outside of SmartPLS ?
silburn
PLS Junior User
Posts: 4
Joined: Mon Oct 22, 2012 6:46 am
Real name and title:
Location: Caribbean
Contact:

Post by silburn »

Visited the link to article by Edward E. Rigdon, Christian M. Ringle, Marko Sarstedt, Siegfried P. Gudergan (2011), Assessing Heterogeneity in Customer Satisfaction Studies: Across Industry Similarities and within Industry Differences, in Marko Sarstedt, Manfred Schwaiger, Charles R. Taylor (ed.) Measurement and Research Methods in International Marketing (Advances in International Marketing, Volume 22), Emerald Group Publishing Limited, pp.169-194.

It costs $35 to access a copy via Emerald Insight.

Is there an alternative to purchasing from Emerald.

Could you send a copy of the article directly to my email
admin
PLS Senior User
Posts: 26
Joined: Sun Sep 11, 2005 8:09 pm
Real name and title:

Post by admin »

Yep, see explanation above: www.smartpls.de -> announcements -> literature
Administration Team
Locked