Project Planning and Project Success

173

Appendix B: Details of

the Statistical Analysis

SUCCESS MEASURE ANALYSIS

For success, three items measured project eciency, and four measured

overall project success. All measures of success showed a high Cronbach

alpha score which shows they are all correlated. is ts with the results

of our review of the success literature and the view that all success mea-

sures are to some extent correlated (Dvir et al., 2003; Prabhakar, 2008;

Kloppenborg, Manolis, and Tesch, 2009; Zwikael and Globerson, 2006).

Table B.1 shows the results of this analysis.

Based on this result there is no strong reason to exclude any of the items.

e average is .905 and alpha will only improve marginally by deleting

project budget goals.

is also indicates that all of the factors above are interrelated to some

extent. However, we can note that the only factors close to the threshold

for removal are budget goals, scope, and time goals. ese are key compo-

nents of eciency. Scope is the lowest of this category which is in keeping

with Shenhar et al. (1997) who stated that scope was the most important

of the triple constraints for overall success.

e results of the Cronbach’s alpha analysis supported the initial

assumptions that the elements identied for measuring success (Müller

and Turner, 2007; Shenhar et al., 2001; Dvir et al., 2003; Zwikael and

Globerson, 2006) were valid measures of success for this survey and accu-

rately measured the judgments of respondents. Each variable achieved

a high alpha score greater than α = .85. In practical terms, this meant

there was a high degree of condence in the reliability and the data col-

lected through the survey, and they are accurate and meaningful for the

purposes of this research.

174 • Appendix B

Factor Analysis

In order to identify patterns of interrelationships among the variables

thought to be potential moderators, a factor analysis was completed. Factor

analysis is a technique for investigating whether a number of variables of

interest are related to a smaller number of unobserved uncorrelated vari-

ables (Cooper and Schindler, 2008). Factor analysis assumes the following:

• Multicollinearity

• Homogeneity of sample

Given the size and nature of the sample, these items were not a concern.

Anormalized varimax rotation was selected to achieve the highest load-

ings and best model t. A scree plot was created to conrm three factors

were appropriate, Figure B.10, at end of Appendix B.

Table B.2 is a summary of this analysis. ree factors were found to

be signicant. Factor 2 consists of three of four of the items identied as

planning factors in the hypotheses plus team experience identied within

the factor analysis. Next, Cronbach’s alpha analysis was conducted on the

components of the three identied factors.

e Cronbach alpha coecient is a number that ranges from 0 to 1.

Avalue of 1 indicates that the measure has perfect reliability. A value of 0

indicates that the measure is not reliable and variations are due to random

error. Ideally the alpha value should approach 1. In general, an alpha value

TABLE B.1

Cronbach Alpha Analysis of Success Measures

Summary for Scale: Valid N:1,378

Cronbach Alpha: .905 Standardized Alpha: .922

Alpha if Deleted

Project success rating 0.885

Project sponsors and stakeholders

success rating

0.884

Project budget goals 0.912

Project time goals 0.903

Scope and requirements goals 0.900

Project team’s satisfaction 0.888

Client’s satisfaction 0.884

End users’ satisfaction 0.889