*4.1. Identification of Critical Data Quality Dimensions of Highway Infrastructure Data*

The descriptive statistical analysis did not yield a whole number for the mean value of the responses. Therefore, for the purpose of interpretation, the impact of each dimension on data quality can be considered to lie between the midpoints of two adjacent scales [75]. The importance of the dimensions about the mean value (μ) greater than or equal to 4.5 was deemed to have a very high impact on the important data quality dimension. Similarly, the range of mean values 4.5 > μ ≥ 3.5 was treated as having high importance; 3.5 > μ ≥ 2.5 was treated as having moderate importance; 2.5 > μ ≥ 1.5 was treated as having low importance; and mean values less than 1.5 were treated as having very low importance on data quality. In the study, the key data quality dimensions for assessing highway infrastructure data for effective decision-making dimensions were deemed to be those that were both very high and of high importance.

#### *4.2. Importance of Data Quality Dimensions at Respective Decision-Making Levels*

Based on the questionnaire results, consideration was also given to the importance of dimensions. The data quality requirement may not be the same at all levels of decisionmaking. For instance, the project level focuses on the primary data collection and format. Hence, the dimensions critical at the project level are not critical at the remaining decisionmaking levels. Hence, the importance of dimensions at all decision-making levels was considered. The significance of data quality dimensions is determined at the strategic, network, program, project selection, and project levels of highway projects. Based on the ratings for the importance of dimensions at the decision-making level, decision-makers believe that all data quality attributes defined under the semiotic model are considered critical in data usage for information generation at all decision-making levels, with a rating of 4 out of 5. The context of data quality differs at each level of decision-making; consequently, data quality dimensions were determined, and the ranking of data quality dimensions was also calculated at each level of the semiotic framework, i.e., at the syntactic, pragmatic, empirical, and semantic levels.

#### *4.3. Ranking of Data Quality Dimensions within the Semiotic Framework*

Along with the level of importance, the decision-makers also prioritise data quality dimensions in each category of the semiotic framework. The priority of data quality requirements has changed from stakeholder to stakeholder at each decision-making level. The semiotic framework comprised the syntactic, empiric, semiotic, and pragmatic categories, which deal with the structure, meaning, information, and knowledge of data characteristics [32]. The prioritisation of dimensions was also taken in the questionnaire survey. The responses to dimensions given by the respondents were converted into a rank using Henry Garrett's ranking technique [76]. This technique provides the change of orders of problems into numerical scores. The prime advantage of this technique over simple frequency distribution is that the dimensions are arranged based on their priority from the point of view of decision-makers. Garrett's formula for converting the ranks into the per cent position is shown below as Equation (1):

$$\text{Percent position} = 100 \times (\text{R}\_{\text{ij}} - 0.5) / \text{N}\_{\text{j}} \tag{1}$$

where Rij = rank given for ith dimension by jth decision-maker

Nj = number of dimensions ranked by the jth individual.

The per cent position of each rank was converted into sources referring to the table given by Garrett and Woodworth [77]. For each factor, the scores of individual stakeholders were added together and divided by the total number of respondents for whom scores were added. These mean scores for all the dimensions were arranged in descending order; the dimensions were accordingly ranked.
