Stat 1 variables & sampling

Bio-Statistics

Prepared by:

Assistant Prof.

Namir Al-Tawil

Definition of statistics

It is the science that is concerned with collection, organization, summarization, and analysis of data; then drawing of inferences about a body of data when only a part of data is observed.

Data

Are the raw material of statistics. Simply defined as numbers. Two main kinds of data:

– Result from measurement (such as body weight).

– Result from counting (such as No. of patients discharged).

Each No. is called datum.

Sources of data

Routinely kept records. E.g.: hospital medical records.

Surveys. Experiments. External sources. E.g.: published

reports, data banks, research literature.

Definitions:

Biostatistics:A term used when the data analyzed are

derived from biological sciences and medicine.

Variable:The characteristic takes different values in

different persons, places or things, so we label a characteristic as variable. E.g. : blood pressure, weight, height, hemoglobin…….

Definitions:

Quantitative variableA variable that can be measured in the usual

sense. E.g.: Weight of pre-school children, age of patients ……

Qualitative variable Can not be measured as the quantitative

variable, e.g. ethnic group, possessing a characteristic or not such as smokers and non-smokers. Here we use frequencies falling in each category of the variable.

Classification of variables:Random variable :Results only by chance factors i.e. can not be

predicted. I. Classification based on GAPPINESS Continuous random variable Does not possess gaps. E.g. height and weight. Discrete random variable Characterized by gaps or interruptions in the

values that it can assume. E.g. No. of admissions per day, or No. of missing teeth.

Categorical (e.g. sex and blood groups). Numerical discrete (No. of episodes of angina).

Classification, cont.Note: To summarize discrete variables we measure the proportion of

individuals falling within each category. For continuous variables we need measures of central tendency and measures of dispersion.

II.Classification by DESCRIPTIVE ORIENTATION Independent variable: Is a factor that we are interested to study. E.g. meat

intake in grams per day. Dependent variable (outcome variable): Is the factor observed or measured for different

categories of the independent variable. E.g. hypercholesterolemia.

III. Classification by levels of measurement

The nominal scale: Consists of classifying the observations into various mutually exclusive categories. E.g. males & females.

The ordinal scale: Observations are ranked according to some criterion, e.g. patients status on discharge from hospital (unimproved, improved, much improved).

Levels of measurements, cont.

The numerical scaleSometimes called quantitative observations.There are two types of numerical scales:1.Interval or continuous scales e.g. age.2.Discrete scales (e.g. No. of pregnancies). Means and standard deviations are generally

used to summarize the values of numerical measures.

Definitions

Population: The largest collection of entities

for which we have an interest at a particular time.

Sample:Part of a population.

Random (probability) Sampling methods

1. Simple random sampling:

Use random number table.

(see next slide).

Random (probability) Sampling methods

2. Systematic sampling: Include individuals at regular intervals. E.g. individuals No. 4, 7, 10, 13, …. Will be included.

The interval in this example is (3), measured by dividing the No. of the population by the required sample. E.g. 60/20.

The starting point must be chosen randomly.

Random (probability) Sampling methods

3. Stratified sampling: Divide into subgroups according to age and sex for example, then take random sample.

4. Cluster sampling: It results from 2 stage process. The

population is divided into clusters, and a subset of the clusters is randomly selected.

Clusters are commonly based on geographic areas or districts.

Convenience sampling

Note: It is not always possible to take a random sample, e.g. a busy physician who wants to make a study on 50 patients attending the out-patient clinic. This is called a convenience sampling (non random).