LibGuides: Research &amp; Publication in Medicine &amp; Health: RESEARCH DATA

Using Existing Data Sets

Using poor quality data can lead to erroneous results and even reputational damage. Before deciding to trust a dataset, assess data quality by considering both the source and the data itself.

Consider the Source

Where did you get the data files? From a source that can ensure the files haven't been changed or corrupted? Are you using the latest, most complete version of the data?

What do you know about the research methodology? Is there an associated article you can read to assess the research design? Do the data collection and processing methods meet the standards for high quality research in your field?

Source: UIUC Research Service Data Nudge 2024-09

Data Quality

6 Dimensions of Data Quality

Once you're assured the data comes from a quality source, assess the data itself. Although assessment tools for specific types of data might include additional measures, most data scientists agree that these six dimensions are essential for assessing data quality.

	Accuracy data correctly represents the source material All records in the Customer Table must have accurate Customer Name, Customer Birthdate, and Customer Address fields when compared to the Tax Form.
	Completeness expected values are fully present and known nulls are clearly marked Completeness measures the degree to which all expected records in a dataset are present. At a data element level, completeness is the degree to which all records have data populated when expected.
	Consistency data is recorded uniformly both within variables and across the dataset Consistency is a data quality dimension that measures the degree to which data is the same across all instances of the data. Consistency can be measured by setting a threshold for how much difference there can be between two datasets.
	Timeliness data represents a time period appropriate for research purposes Timeliness is the degree to which a dataset is available when expected and depends on service level agreements being set up between technical and business resources.
	Uniqueness data records are not unnecessarily duplicated Uniqueness measures the degree to which the records in a dataset are not duplicated.
	Validity data values fit within defined ranges or categories Validity measures the degree to which the values in a data element are valid.

Source: UIUC Research Service Data Nudge 2024-09

Examples of data quality dimensions can be found at datacamp's "Data Quality Dimensions Cheat Sheet".