İSTATİSTİK I

Several characteristics define a data set’s structure and properties. Element, Variable, Case and Observation are the key components of data sets.

Ratio scales of measurement, in addition to having all properties of the interval scale, have a natural or zero-valued base value that cannot be changed. For example, an individual’s age, weight, height, systolic blood pressure are ratio scale variables because they have natural base value. For example, John and Mary are 20 and 40 years old, respectively. We can say that Mary is two times older than John.

Ordinal scales of measurement have the property of both classifying and magnitude. Subjects are categorized into different rank ordered groups. Each value on the ordinal scale has a unique meaning, and it has an ordered relationship to every other value on the scale.

In an observational study, researchers simply collect data based on what is seen and heard and infer based on the data collected. Researchers observe subjects and measure variables of interest without any intervention to the subjects. Experimental study is a study in which the researcher manipulates some of the variables and try to determine how the manipulation influences other variables.

Data collection by observation procedure has some advantages and limitations. The advantages are:

1. The direct recording of the data avoids problems such as incomplete or distorted recall.
2. Data can be obtained continuously over an extended period of time.

The limitations are:
1. The observer or the instrument to be used for data gathering must be able to record the events correctly. For example, human observers must get through training about the study and the data to be collected and so that different observers will record the same events in the same manner.
2. The individuals included in the study might be aware of this fact and then altered their behavior, decision or answers. This leads to bias in the study.

In many surveys, closed-ended questions are preferred because close-ended questions lead to obtain higher response rates when responders don’t have to type so much.

Likert type of scale is generally used to measure the attitudes of an individual towards a specific subject.

When the data set is large or the measurements are obtained using ratio scale, grouped frequency is more appropriate for summarizing the data.

Interpretation of the frequency distribution table can be easier or clearer when we use the percentage of the frequency.

A frequency distribution table provides information of how many observation or elements occurred for each value or group of values of a variable. Cumulative frequency is used to determine the number of elements that falls above or below a particular value in a given class interval.

Analysis is not a component of data sets. Analysis can be done over datasets.

Nominal scale classifies things into qualitative different classes. Gender is qualitative.

That type is ratio scale. An example of this type scales is age. The age of a person can be zero at minimum. So if one person is 60 years old, he/she is 3 times older than a person who is 20 years old.

The researcher is making an observation.

It's a open-ended question type because the respondent can freely answer whatever he thinks of. He doesn't choose among or rank the given options.

A pretest over a small sample can help us in determining whether questions are clearly understood and whether the order of questions cause a difference in results.

Since almost half of the population is female, he is making a mistake in sampling. The parfume preferences of males and females could be different and his sample is male biased.

Class Width=Range/Number of classes

where Number of Classes=√n, where n=number of observations

Thus Number of Classes=√500=22.36

But when the number of classes is larger than 20 we take it as 20. So in this case:

Class Width=(40-20)/20=20/20=1kg

There are 200 workers in factory and in total 190 them are below 50 years old. Thus the cumulative frequency is 190/200=0.95 (95 %)

The definition corresponds to random error. Random errors are caused by unknown and unpredictable factors that randomly affect measurement of the variable across the sample.