Background
Methods
Design
Data collection and ethics
Participants
Attitudes to and knowledge of oral health questionnaire
Attitudes to oral hygiene |
Item 1- I think it feels nasty to take care of other people’s mouths |
Item 2- I think oral care is part of my job duties |
Item 3- I think it is practically difficult to perform oral care |
Item 4- The caregiver refuses to receive help with oral care |
Implementation possibilities |
What opportunities do you think you have when it comes to offering oral care to the healthcare provider you are responsible for? |
Item 5- I can take the time needed to provide oral care |
Item 6- I have enough knowledge to perform proper oral care |
Item 7- I have appropriate aids for the implementation of proper oral care |
Item 8- I know how to practically perform oral care |
Item 9- To caregivers who want to take care of their oral care themselves, I can give appropriate oral care advice |
Item 10- By actively informing “reluctant” caregivers, I can in the long run get them to accept help with oral care |
Knowledge of importance |
What skills do you think are important for being able to perform good oral care? |
Item 11- Assistive products and oral care |
Item 12- Diseases affecting the oral cavity |
Item 13- Various artificial (prosthetic) dental substitutes |
Item 14- What the healthy oral cavity looks like |
Item 15- Oral physiological function (e.g., chewing, swallowing, speech) |
Item 16- The psychosocial function of the oral cavity (e.g., appearance, well-being) |
Data analysis
-
1. In this theory, the indicators related to the test and the questions depend on the sample group. The sample group’s level of ability and its distribution strongly affects the item characteristics, such as diagnostic power, difficulty level, standard deviation, variance, and test mean. For this reason, the ability to generalize the results to other groups and communities is limited.
-
2. Revealing the level of ability of people depends on tests and questions. A person may score differently on two tests that measure the same trait but differ in difficulty levels.
-
3. The test, which is based on classical theories, is aimed more at people with moderate levels of ability. This means that the validity of the test is lower in both the upper-ability group and in the lower-ability group than in the middle group.
-
4.The standard measurement error is assumed to be the same for all individuals. This assumption leads to wrong decisions about people because standard error can vary based on ability level.
-
5. The degree of difficulty cannot be exactly the probability of answering a certain question correctly among people with different abilities.
Results
Item | Analyses from classical test theory | Analyses from Rasch | |||||
---|---|---|---|---|---|---|---|
Factor loadinga | Item-total correlation | Infit MnSq | Outfit MnSq | Discrimination | Difficulty | DIF contrast across genderbc | |
Item 1 | 0.543 | 0.226 | 1.24 | 1.14 | 0.83 | -1.68 | 0.16 |
Item 3 | 0.512 | 0.396 | 0.77 | 0.78 | 1.19 | 0.84 | -0.06 |
Item 4 | 0.317 | 0.214 | 0.96 | 1.00 | 0.97 | 0.84 | -0.09 |
Item 5 | 0.551 | 0.448 | 1.43 | 1.42 | 0.54 | -0.41 | 0.08 |
Item 6 | 0.793 | 0.668 | 0.70 | 0.70 | 1.35 | -0.10 | -0.16 |
Item 7 | 0.756 | 0.618 | 0.79 | 0.80 | 1.23 | 0.350 | 0.37 |
Item 9 | 0.617 | 0.502 | 1.08 | 1.08 | 0.90 | 0.01 | -0.27 |
Item 11 | 0.601 | 0.556 | 1.22 | 1.37 | 0.73 | -0.52 | 0.05 |
Item 12 | 0.703 | 0.665 | 0.96 | 0.95 | 1.07 | -0.27 | -0.07 |
Item 13 | 0.697 | 0.640 | 1.21 | 1.24 | 0.77 | 0.60 | 0.43 |
Item 14 | 0.794 | 0.726 | 0.84 | 0.83 | 1.18 | -0.29 | -0.05 |
Item 15 | 0.757 | 0.705 | 0.93 | 0.94 | 1.09 | 0.20 | -0.15 |
Item 16 | 0.791 | 0.732 | 0.80 | 0.80 | 1.20 | 0.27 | -0.35 |
ATa | IPb | KIc | Suggested cutoff | |
---|---|---|---|---|
Psychometric testing | ||||
Ceiling effects (%) | 1.3% | 2.0% | 39.9% | < 20 |
Floor effects (%) | 0.2% | 0.2% | 0 | < 20 |
Internal consistency (Cronbach’s α) | 0.460 | 0.760 | 0.869 | > 0.7 |
Confirmatory factor analysis | ||||
χ2 (df) | 66.372 (62) | - | - | Nonsignificant |
Comparative fit index | 0.998 | - | - | > 0.9 |
Tucker-Lewis index | 0.997 | - | - | > 0.9 |
Root-mean square error of Approximation | 0.011 (0.00–0.028) | - | - | < 0.08 |
Standardized Root Mean Squared Residual (SRMR) | 0.042 | - | - | < 0.08 |
Average Variance Extracted | 0.22 | 0.47 | 0.65 | > 0.5 |
Composite Reliability | 0.44 | 0.78 | 0.92 | > 0.6 |
Item separation reliability from Rasch | 1.00 | 0.96 | 0.95 | > 0.7 |
Item separation index from Rasch | 18.17 | 4.69 | 4.37 | > 2 |
Person separation reliability from Rasch | 0.70 | 0.71 | 0.54 | > 0.7 |
Person separation index from Rasch | 2.00 | 2.03 | 1.08 | > 2 |