The three alteration questions by respondent role: C199, C205, C211 by C005

Three Part 5 questions asked respondents whether they had altered data provided by the financial institution (C199), the financial institution's processes (C205) and AI outputs (C211). Saying True was almost entirely confined to individual contributors. Among engagement managers the share saying True was 6.08% or lower on each question, and no portfolio director said True to any of the three. The difference between roles reached significance on all three questions.

For C199, on data provided by the financial institution, 63.43% of individual contributors said True (n = 659), against 2.00% of engagement managers (n = 150) and 0.00% of portfolio directors (n = 30). The role difference gave chi-square 215.77 on two degrees of freedom, p < 0.001, n = 839 respondents.

For C205, on the financial institution's processes, 72.66% of individual contributors said True (n = 662), against 6.08% of engagement managers (n = 148) and 0.00% of portfolio directors (n = 31). The role difference gave chi-square 265.42 on two degrees of freedom, p < 0.001, n = 841 respondents.

For C211, on AI outputs, 42.46% of individual contributors said True (n = 650), against 1.36% of engagement managers (n = 147) and 0.00% of portfolio directors (n = 31). The role difference gave chi-square 107.09 on two degrees of freedom, p < 0.001, n = 828 respondents.

The first table gives the base. The second gives the answers by role for each question, with the share saying True computed on the respondents who answered True or False. The third gives the test of the difference between roles.

Across all roles, 50.18% of the 839 respondents who answered True or False said True for C199, 58.26% of 841 for C205 and 33.57% of 828 for C211. The full distribution of each question, with its non-answers, is reported in the file of that indicator.

The analysis is itself a breakdown by respondent role, given in full in the second table. It was not broken down by any other segment.

Each question was tested for a difference between individual contributors, engagement managers and portfolio directors by a chi-square test on the respondents who answered True or False. The respondents holding the role of Customer Management Executive gave no True or False answer to any of the three questions, so they appear as counts and are not in the test. All three tests reached significance.

Cancellation was not part of this analysis. The association of each question with cancellation is reported in the file of that indicator.

Customer satisfaction was not part of this analysis. The association of each question with customer satisfaction is reported in the file of that indicator.

The engagement indicator mean was not part of this analysis.

Intent to leave was not part of this analysis.

The analysis concerns respondent role alone, so no comparison across the four measures applies.

The tests treat respondents as independent. Respondents on the same engagement share conditions, so the p-values are somewhat smaller than a test clustered by engagement would give. No clustered test was computed, so the p-values are approximate.

No equivalence test was computed, because every test found a difference.

No robustness checks were computed for this analysis.

Respondent role and the three questions are separate fields, and neither enters the other. No part-whole correction applies.

A derived analysis has no profile against the study norm, so no comparison applies.

The alteration of data, processes and AI outputs was reported by the respondents who do the hands-on work. Individual contributors said True far more often than engagement managers, and portfolio directors never did. The pattern fits work done at the level where the data, the process and the output are handled, and it also fits a reluctance among managers to report alterations made under their oversight. The analysis cannot separate the two readings.

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

The questions ask whether the respondent personally altered data, processes or outputs. A manager who knew of alterations made by the team would answer False, so the role difference measures who made the alterations, not who knew.

The tests treat respondents as independent and were not clustered by engagement.