In the practice of clinical science, statistics, as a methodology, runs through the whole process of clinical research, and needs the support of statistical knowledge from research design, data collection and arrangement, data analysis, results display and interpretation to the publication of final papers.

Before statistical analysis of data, it is necessary to make clear the types of research variables, and then choose the appropriate statistical analysis methods according to the research purpose and variable types.

In this issue, we will introduce the basic concepts of variables, the types of variables and the conversion of data types. In the next issue, we will briefly introduce the choice of statistical analysis methods for different types of variables, so please pay attention.

The collection of clinical data is recorded in the database through fields or variables. The type of clinical data mentioned here actually refers to the type of variables. Variables can be classified according to their characteristics, and there are many classification methods. For example, they are divided into independent variables and dependent variables from the perspective of causality, and they are named random variables according to their mathematical characteristics. It is important to understand the basic concepts and categories of variables, because different categories of variables need to be analyzed by different statistical methods.

1

basic concept

Examples of specific clinical observation items and variable names and labels in the database are shown in table 1.

Table 1

Note: Table 1 From Clinical Research Methods of Evidence-based Chinese Medicine edited by Liu Jianping.

2

Variable type

We sorted out the relevant knowledge points of data types with reference to relevant statistical knowledge, and divided the data into quantitative data and qualitative data according to the research needs. As semi-quantitative data, grade data also supplemented this. The specific data types are shown in the figure below.

2. 1 Quantitative data is also called measured data or numerical variables. Data obtained to observe the size of the index in each unit. Its variable value is quantitative, expressed by numerical size, and generally has a unit of measurement.

According to whether the observed values are continuous or not, they can be divided into continuous and discrete types.

2.2 Qualitative data, also known as enumeration data or disordered variable data, is data obtained by counting observation units according to a certain attribute or category and grouping and summarizing the number of observation units in each group. Its variable value is qualitative, indicating incompatible attributes or categories, such as negative and positive test results, family history and so on. , which can be divided into two situations:

2.3 Grade data: Grade data is also called semi-quantitative data or ordered classified variable data. In order to classify the observation units according to different degrees of an attribute and make statistics in groups, the data obtained by classifying and summarizing the number of observation units in each group has semi-quantitative nature, which is expressed by grade size or attribute degree. If we observe the curative effect of a drug on a patient, the results can be divided into four grades: cured, effective, improved and ineffective.

three

Conversion of data types

In data analysis, the data of different types of variables can be transformed according to the analysis needs. Generally, variable data conversion is carried out in a way that includes information from high to low. Quantitative variables can be transformed into qualitative variables according to research needs. Qualitative variables can't be converted into original quantitative variables, so quantitative data should be collected as much as possible in the data collection stage, and quantitative data contains more information than qualitative data.