Choosing the Proper Statistical Test Type Of Data

Loosely speaking, there are two types of data: categorical data and continuous data. The categorical data can be further divided into nominal data and ordinal data.

The possible values for continuous data typically will form a whole interval or range. For example, data collected from measuring the body weight of adult mice in line A could range from very low, say 150 grams, to very high, say 450 grams. The body weight of a randomly selected member of this population could assume any value in this range. For instance, a random sample of n = 5 subjects from this population could have the following set of weight measurements: {177, 283, 222, 155, 399}.

Categorical data are a collection of observations that can be categorized by classification or on an ordered scale. Nominal data are based on classifications. Suppose that, as a researcher, you are interested in collecting information on the distribution of body weight throughout a given mouse line according to the breeders from which they were bred. Each subject would be measured for body weight and then classified by the breeder. The body weights of mice could also be categorized according to relative size. For instance, mice with body weight less than the 25th percentile may be categorized as ''lean''; mice with body weight between the 25th and 75th percentile may be categorized as ''normal weight''; and those over the 75th percentile may be categorized as ''overweight.'' These categories have a natural order or scale to them, and this type of data is called ordinal. The information on the scale of ordinal data can be used to the advantage of the researcher by way of testing for trends that would not be possible were the data treated as nominal.

The distinctions between these data types are not always clear. One must look beyond the measurements themselves and examine the quantities or qualities being measured before giving meaning to the numbers in a dataset. Continuous data may be divided into categories that can be treated as being either ordinal or nominal. Although there may also be many justifiable reasons for doing this, such as nonlinear relationships between factors under study, continuous data may be categorized simply to ease the interpretation of the results of an experiment. In our previous example, we showed how the mice could be categorized according to preset cut points into either lean, normal weight, or overweight classifications by their body weights. In this case, researchers may prefer the ordinal categorization in order to make statements about the relationship between some factor(s) and being normal weight or overweight rather than trying to relate the continuous measure of body weight with the factor(s). Another example is that we can categorize animals into those with hypertension and those with normal blood pressures based on their systolic blood pressure. In this case, we transform a continuous variable into a nominal one. The ease in interpretation from such analyses does come at a cost. Categorizing continuous data will strip it of some portion, perhaps a very significant portion, of the information on the relationship the researcher is investigating.

Blood Pressure Health

Blood Pressure Health

Your heart pumps blood throughout your body using a network of tubing called arteries and capillaries which return the blood back to your heart via your veins. Blood pressure is the force of the blood pushing against the walls of your arteries as your heart beats.Learn more...

Get My Free Ebook


Post a comment