Abstract (eng)
Introduction: Validity is considered the primary quality criterion of an assessment instru-ment. The aim of this thesis was to test and develop the content validity in addition to the se-mantic validity of the prevalence assessment questionnaire for bedridden persons and local confinement.
Method: For this purpose, a Delphi survey was conducted taking into account the qualitative and quantitative research approach. The Item Content Validity Index (I-CVI) and Scale Con-tent Validity Index (S-CVI) was used for the quantitative analysis. Significant changes in the unconformity between question round 1 and 2 were calculated by using the McNemar test. CVI values ≥ .90 were assumed sufficient. The open-ended responses were evaluated by a structured qualitative content analysis.
Results: The qualitative part of the research resulted in a redefinition of the phenomenon of local confinement (LC). The original three-level characteristic of LC was abandoned in favor of introducing a five-level radius of action (RA). The LC exists only as a dichotomous trait. A sufficient semantic validity (I-CVI = .91) was seen for the LC item in question round 1; never-theless, a non-significant increase of the I-CVI in question round 2 (I-CVI = .97, p = .625) could be achieved. For the bedridden persons (BP) item, the semantic validity was not given in question round 1 with an I-CVI of .77; however, in question round 2 it could be increased to I-CVI = .97 (p = .039). The linguistic discriminatory power between mild and moderate BP and between moderate and severe BP was also not given in question round 1 with an I-CVI of .77 and .87 respectively. Both values could be increased in question round 2 (I-CVI = .97, p = .039 and I-CVI = 1, p = .125). For the BP phenomenon, the S-CVI/Ave could be increased from .80 to .98 in question round 2. All five levels of the RA achieved I-CVI values of ≥ .90. The S-CVI/Ave for the RA was .99.
Conclusion: Although very high I-CVI values were achieved for RA, LC, and BP in question round 2, validity testing and development cannot be considered as complete. The results raised further questions, which have to be clarified in a subsequent research project.