What professional certifications do you have?

Of the 850 respondents, 793 answered the question. None, a listed answer, was the most frequent (236, 29.76%), followed by Six Sigma Green Belt (127), ITIL Foundation (126) and AWS Certified Cloud Practitioner (124). Each answer was given by respondents of a single role: individual contributors gave None or one of those three certifications, engagement managers gave PMP alone or with ITIL or Scrum Master, or Six Sigma Black Belt, and portfolio directors gave PMP with an MBA or ITIL Expert, or PgMP.

At engagement level, the certification assigned to each engagement was not associated with cancellation (chi-square 4.29, p = 0.235, n = 99 engagements), customer satisfaction among active engagements (Welch F 0.00, p = 0.992, n = 50 engagements), the engagement indicator mean (Welch F 1.60, p = 0.211, n = 99 engagements) or intent to leave (Welch F 1.15, p = 0.346, n = 99 engagements). Those tests compare the four answers of individual contributors, the only answers that were the plurality on 10 or more engagements. At respondent level, with standard errors clustered by engagement, the answers differed on all four measures, but those tests include answers given by one to twelve respondents and track role. Restricted to the seven answers given by 30 or more respondents, cancellation and customer satisfaction no longer differed.

Each engagement was assigned the answer given by the largest number of its respondents, with None counted as an answer. Fifty-one engagements were tied and carry no assigned answer, and 99 carry one. Every assigned answer is one of the four given by individual contributors. The MBA alone was given by one respondent.

Each share is computed on the respondents who answered in that row. Rows with fewer than 10 respondents report counts only. The respondent role rows are not tested, because the segment family records the same split the answers make.

Segment families recorded at engagement level were tested on engagements, and respondent role and vendor type on respondents. After Holm correction, the answers differed by respondent role (Cramer's V 1.00, Holm p < 0.001), a complete separation: each answer was given within one role only. No other family differed.

Cancellation rate is the share of engagements canceled by the study date, measured at engagement level (150 engagements, 56 canceled).

Across the 99 engagements assigned one of the four answers of individual contributors, cancellation did not differ by answer (chi-square 4.29, df = 3, p = 0.235). At respondent level, with every answer as a level and standard errors clustered by engagement, it differed (F = 25.68, df = 11 and 149, p < 0.001, n = 793 respondents). Restricted to the seven answers given by 30 or more respondents, the difference did not hold (F = 1.72, df = 6 and 149, p = 0.120, n = 735 respondents).

Customer satisfaction is the percentage recorded against each engagement. The engagement-level estimate uses the active engagements in the answer levels with 10 or more active engagements.

Customer satisfaction did not differ by answer at engagement level (Welch F 0.00, p = 0.992, n = 50 engagements). At respondent level, with standard errors clustered by engagement, it differed (F = 23.00, df = 11 and 149, p < 0.001, n = 793 respondents). Restricted to the seven answers given by 30 or more respondents, the difference did not hold (F = 1.55, df = 6 and 149, p = 0.165, n = 735 respondents).

The engagement indicator mean is the mean of an engagement's ratings across the rated indicators. At respondent level the measure is the respondent's own mean rating. The certification is not a rated indicator and does not enter either mean.

The engagement indicator mean did not differ by answer at engagement level (Welch F 1.60, p = 0.211, n = 99 engagements). At respondent level it differed (F = 26.31, df = 11 and 149, p < 0.001, n = 793 respondents). The respondents' own mean ran from 3.7 among individual contributors with the AWS certification to 5.2 among portfolio directors with PMP and an MBA, the order of the roles. Restricted to the seven answers given by 30 or more respondents, the difference held (F = 3.02, df = 6 and 149, p = 0.008, n = 735 respondents).

Intent to leave is the answer to C019, “How likely are you to leave your current job in the next 12 months?”, rated 1 to 10 by up to 798 respondents. For the indicator, the engagement-level comparison uses the mean intent to leave of each engagement.

Intent to leave did not differ by answer at engagement level (Welch F 1.15, p = 0.346, n = 99 engagements). At respondent level it differed (F = 2.12, df = 10 and 149, p = 0.026, n = 746 respondents). Individual contributors with the AWS certification reported the highest mean (7.3, n = 124 respondents). Restricted to the seven answers given by 30 or more respondents, the difference held (F = 2.71, df = 6 and 149, p = 0.016, n = 689 respondents).

Steiger's z compares two correlations, and a categorical indicator yields an effect size across its levels and not a correlation. The pairs of measures were therefore not compared. No engagement-level association reached significance.

Answers did not cluster within engagements (ICC −0.04): respondents on the same engagement gave different answers, as the mix of roles on each engagement requires. The design effect was 1.00, and the effective n equaled the 793 answers.

Customer satisfaction was equivalent to zero (equivalence p = < 0.001, n = 50 engagements). The engagement indicator mean and intent to leave were not, and their results are inconclusive rather than evidence of no association. The cancellation contrast carries no equivalence test.

A check holds when its test falls on the same side of 0.05 as the primary test. The engagement-level checks held. Each respondent-level check failed, because the respondent-level test found a difference where the engagement-level test found none (four of four). The respondent-level test compares answers that track role, and it includes answers given by one to twelve respondents.

Each of the four measures was computed from data the indicator does not contain. Each result is therefore a finding and not an arithmetic consequence.

A categorical indicator has no signed correlation with the study measures and therefore no profile against the study norm.

The certifications respondents reported divide along role without overlap, so the indicator restates role in another form. Among the engagements assigned an answer of individual contributors, the certification was not associated with any study measure. The respondent-level differences in the respondents' own ratings follow the order of the roles, and none of the respondent-level differences can be separated from role.

The evidence is the reported experience of the vendor's delivery teams, collected in a single cross-section. The associations reported here are not causal.

Only the four answers of individual contributors were the plurality on 10 or more engagements, so the engagement-level tests do not compare certifications across roles.

The respondent-level tests include answers given by one, eight, nine and twelve respondents. A level of a few respondents rests on a few engagements, so the size of the respondent-level F statistics should not be read as a measure of effect. Restricted to the seven answers given by 30 or more respondents, the differences in cancellation and customer satisfaction did not hold, and those in the engagement indicator mean and intent to leave did.

Tests by respondent role and by vendor type treated respondents as independent.