Вивчаючи статистику

4. Read and translate the texts:
Categorical data are concerned with characteristics of cases, such as an individual’s name, gender, colour (say, of eyes, hair, car or socks) or opinion about some issue (which may be given as an answer in an interview or questionnaire in terms of categories such as: agree, have no opinion, disagree, or better, stay the same, worse). In the main, categorical data tend to be described using percentages. To calculate a percentage is easy:
1. Take the number of cases in a certain category.
2. Divide this by the total number of cases.
3. Multiply the result by 100.
With numerical data, the variables that you are concerned with include measurements and counts such as weight, height, age, or income, and so on, and these are obviously expressed as numbers. Because these data have numerical meaning, you can describe them in more ways than you can describe categorical data. What is comforting is that these descriptive statistics can often be done without needing complicated statistical software – your fingers and some mental arithmetic, or a calculator, will do just as well. Descriptive statistics allow you to describe where the centre is; that is, where the middle point of the data is, or what a typical value might be, and how much the data differ from (are dispersed) around this middle value. And, clearly, the method that you choose to find the centre of a data set influences the conclusions that you can draw about your data.
Inferring differences and relationships
One way you can begin to look for relationships between variables in your categorical data is by creating a cross-tabulation (often referred to as a ‘crosstab’ by statistical software). This enables you to summarise information about two categorical variables at the same time – for instance, gender and age. Using statistical software, crosstabulations can also give you the percentage of individuals in each combination of categories that you have chosen – for instance, where gender and age are concerned, the percentage of women who are aged 20–29. Cross-tabulations present data in tables.
The usefulness of cross-tabulations is that you can move on from them to conduct statistical tests to find out whether there is a link between the two variables of concern – in formal language, infer whether there is a significant relationship between the two identified variables or if the difference, for example in the number of males and females within each age group, is significant. As we have already noted, the statistical tests you might use to do this are called inferential statistics because they allow you to draw conclusions or inferences about a population on the basis of the data in your sample.