Shreemathi S. Mayya, Ashma D Monteiro, and Sachit Ganapathy
You are watching: Is bmi an interval or ratio variable
Identification and description that variables supplied in any study is a necessary component in biomedical research. Statistics analyses rely on the kind of variables that are involved in the study. In this short article, we present the different species of biological variables. A researcher needs to be familiar with the type of change he/she is managing in his/her study to decide about appropriate graphs/diagrams, an overview measures and statistical analysis.
Research question is the initial and integral step in any type of research work. Relying on the research concerns to be answered and the data available, researchers decide around the statistical approaches to be provided for analysis. Researchers have to be acquainted with the variety of variables connected in their study to choose suitable diagrams/graphs and summary measures because that presentation, and also valid statistics tests for the evaluation of data.
Information collected about a sample of topics (often patients) comprises attributes which vary amongst the subjects. Any characteristic, which different from separation, personal, instance to individual is dubbed a change (1). The characteristics such as age, sex, height, weight, human body mass table of contents (BMI), blood group, body temperature, blood glucose level, blood pressure, love rate, variety of teeth, severity of an illness (mild, moderate, severe) etc. Are several of the examples for biological variables in research. A basic distinction in the nature in between these variables is your quantitative or qualitative (categorical) measurements (1,2).
Quantitative variables are those features which deserve to be a count or measured numerically. They deserve to be continuous or discrete. Constant variable can theoretically take infinitely numerous values in a given range. This way that, we can constantly find an intermediate value between any two values, however close lock are. For example, in a given variety of 5–10 cm size one can write infinitely plenty of values favor 5, 5.1, 5.12, 5.01, 5.003 cm… etc, depending upon the degree of accuracy decided by the researcher. Height of a person, weight, age, arm length, blood pressure, temperature, glucose level are few of the examples for constant variable. Right here the acquired measurements have the right to take any value in a provided range.
Discrete variable (discontinuous variable) deserve to take just specified number of values in a provided range. For example, variety of children per household in a given selection of 0–5 deserve to be 0, 1, 2, 3, 4 and 5. No an ext values in this selection can be written. Variety of visits to hospital in a year, number of children in a family, number of admitted patients in a hospital ward, number of missing teeth etc. Are several of the examples for discrete variables. Discrete variables are usually counts.
Qualitative (categorical) variables space those qualities which space not numerically measurable. This variables room either in the name (no herbal ordering) or ordinal (ordered categories). Usually, for the function of data entrance and analysis using software, categories room coded assigning number values.
Nominal variables permit for only category or categorization based on some distinctively different characteristic, however we cannot rank order those categories. Typical examples of in the name variables room sex, religion, blood group, symptom of disease, cause of death etc. Numerical worths assigned to different categories are valuable for the function of identification just (e.g., 1= male, 2= female). Once a qualitative variable has actually only 2 categories (alive/dead, male/female, diabetic/non-diabetic), it is dubbed a binary or dichotomous variable. In the name variables are summarized with counting (frequency) and also expressing relationship of each classification (percentage).
Ordinal variables allow us to location order the categories in terms of which group has less and also which group has more of the high quality represented by the variable, but the distances in between categories are not known. A usual example of an ordinal variable in medication is the step of a diseases (stage ns to stage IV). For example, we know that “stage I” is less severe 보다 “stage II” that a disease but us cannot tell the exact difference between the two stages. Socioeconomic condition of families (low, middle and high socio-economic status), BMI category (underweight, normal, overweight, obese), disease condition (deteriorated, same, improved), ache score etc. Are a couple of examples for ordinal variables. Numerical values assigned for various categories are helpful for identification and rank bespeak (e.g., 1= low, 2= middle and 3= high earnings group). Ordinal variables space summarized through counting (frequency) and also expressing ratio of each classification (percentage).
Categorizing a continuous variable
Quantitative variables are frequently converted to categorical ones using “Cut-points”. Instead of presenting the typical fasting glucose level of male and female subjects, one might prefer to existing the ratio of diabetics in male and female population using a fasting glucose level that 110 mg/dL as the cut-point to categorize the subjects as diabetic/ non-diabetic. However, categorizing a continuous variable bring about loss of details (3). For example, while categorizing, topics with fasting glucose level the 85 and also 109 mg/dL room treated as equal and also classified together non-diabetic. Similarly, topics with glucose level 111 and 150 mg/dL room classified as diabetic. The difference in the values will no be noticed while presenting just the variety of diabetic and also non-diabetic cases.
Dealing through Likert type data
Likert range is developed with a rule of measuring mindsets by asking people to respond come a collection of statements around a topic, in terms of the extent to which castle agree with them (4). A statement (Likert item) such as: “It’s vital for all biologists to find out statistics” can be request to it is in rated together 1= strong disagree, 2= disagree, 3= no agree nor disagree, 4= agree, or 5= strongly agree or periodically on 7 values rather of five, including “very strongly disagree” and also “very strongly agree”. Variables measure on Likert item are a form of ordinal variables. Likert range is the result of including together the scores on number of Likert items. Likert scale may be treated as a constant variable. Descriptive and also inferential statistics depend on the distribution of scores, symmetric or skewed.
Qualitative data (nominal or ordinal variable) might be gift in the kind of frequency tables. We count the variety of subjects/units in each group of the variable along with percentage and also present the numbers and also percentages in a table. E.g., we summarize Blood group circulation of 100 subjects in the kind of a table mirroring blood group and also corresponding frequency in addition to percentages. If we have the data for 2 categorical variables, data might be gift in the type of a contingency table showing frequency and percentages.
As ordinal variables are likewise categorical variables with a pre-determined order, the descriptive measures such as frequency and percentage has to be reported once the number of categories space few. In addition, median, inter-quartile selection along through maximum and also the minimum worth is considered suitable for summarizing ordinal variables.
Nominal data and also ordinal data with minimal number of categories can additionally be gift in a diagrammatic form, such as a bar chart and also pie chart. In a bar chart, size of the bars to represent the frequency or relative frequency that each group of the variable. Commonly the bars room of equal width and there is a room between them. A pie graph is essentially a circle separated into segments with the area of every proportional come the it was observed frequency in each category of the variable. Full area to represent the full frequency.
Mean and standard deviation room appropriate an overview measures for constant variables with symmetrical distributions. Median and inter-quartile range are to it is in computed to summarize quantitative variables v skewed distributions. Selection is much information if supplied as a supplement to typical deviation or inter quartile range. Discrete variables might be summarized and also analyzed either together a continuous variable or as an ordinal variable relying on the variety of distinct values.
Quantitative data can be represented graphically by way of a histogram. Histogram is advantageous to decide around the form of the distribution, symmetry or skewed. But, with little samples, histogram may not be helpful to identify the shape. As a dominion of thumb, if the median is smaller than twice the standard deviation the data are most likely to be skewed for change with positive values (5). Quantitative data can likewise be shown as stem & leaf plots, period plots, box & whisker plots and also scatterplots, depending on the situation (6).
Analysis of data
Type that the variables decides the kind of statistical analyses to it is in performed, parametric or non-parametric. Parametric methods, such as t-tests, ANOVA, Pearson’s correlation, and also regression, require the presumption that the data monitor a normal distribution and also that variances that the distributions room equal. Commonly used nonparametric approaches are Mann-Whitney or Wilcoxon rank sum test, Wilcoxon signed location test and also rank correlation. Non-parametric methods, do no assumptions about the distribution of the data; they use the rank order of monitorings rather than actual dimensions (7). Chi-square check (or Fisher’s exact test if the number are really small) is the most often used method to compare categorical data. Fail to pay attention to assumptions and their implications can lead to rise in kind I or kind II errors.
We analyze data from comparable studies, totally differently relying on the type of variable involved. For example, let us say the our target population is 50+ age group in a certain population and we have actually measured the change systolic blood push in a sample that 40 male and 40 mrs subjects, and also our null theory is “Male and female population have the very same systolic blood pressure”. We would certainly compare the average blood pressure in males and females v a two-sample t-test (parametric test). If the variable is convert to hypertension standing (hypertensive/normal), the is a nominal variable, and we would compare the hypertension frequencies in males and also females through a Chi-square test (non-parametric test). We uncover smaller P worth for t-test contrasted to chi-square test. Crucial message that we shot to convey below is that, statistics tests will have more power for a consistent variable than the matching nominal or ordinal variables (2). In various other words, to achieve the same power together that of a parametric test, non-parametric test require larger sample dimension than a parametric test. Therefore, one may categorize the data because that the purpose of presentation (e.g., hypertensive/normal), yet not because that statistical analysis (3).
Detailed conversation of various tests is the end of the border of this article. Campbell & Swinscow (2) have summarized the tests an ideal for various types of variables in a single table. Because that computation procedure and an ext details about various parametric tests, researchers might refer some conventional text books (1,3,8). For a great discussion of a number of nonparametric test readers might refer Siegel and also Castellan (9) and also Conover (10).
See more: That'S My Dad And Me Or My Dad And I " Is Correct Or "My Father And Me" ?
The kind of descriptive and analytical procedures to be offered in data summarization and analysis, all rely on the type of variables. Therefore, to achieve the relevant measures for dataset in ~ hand, us recommend the researchers to examine the attributes of data (categorical, quantitative) and shape that the frequency circulation (symmetrical bell shaped, skewed) prior to deciding around the descriptive measures, graphs and diagrams, and statistical tests an ideal for the presentation and evaluation of data.