The verb you choose signals what your design can support. Committees read verbs closely, and a mismatch is among the easiest comments to earn.
"Caused," "produced," and "resulted in" belong to randomized or quasi-experimental designs. "Associated with," "predicted," and "was related to" fit matched comparisons and correlational designs. "Improved from X to Y" or "gained" fit pre-post designs without a control group, as description of change rather than of cause. "Reported," "perceived," and "indicated" fit self-report and descriptive data.
"Contributed to" is the verb that most often slips through. It implies causation, because a contribution is a causal share of an outcome. It requires a comparison design behind it.
A practical check before the defense: search the document for the ten or so verbs that connect a program, practice, or variable to an outcome, and mark the design next to each. A mismatch has two fixes. Change the verb to fit the design, or state plainly that the design cannot support the stronger verb and name the study that would.
A short example. A survey finds that students who attended tutoring reported higher confidence. "Tutoring increased confidence" overstates it. "Students who attended tutoring reported higher confidence" fits a self-report design, and the limitations can note that a comparison design with baseline measures would be needed to support a causal claim.
Checking each verb against the design removes a line of questioning you do not need, and it makes the limitations section shorter, because the findings already say no more than the design allows.
This is practitioner guidance on reporting. It follows standard design-based conventions in education research, not a single source.
I coach Ed.D. and Ph.D. students in education through the dissertation. Details and pricing: cevallos.org/dissertation-thesis-coaching