Free Descriptive Statistical And Inferential Methods Research Paper Sample

Descriptive statistical procedure is a form of presenting numerical facts and data in either tabulated or graphical form, it does not involve making conclusions beyond the analysis. Inferential statistical method involves taking representative samples from a population, analyzing the sample and making generalized inferences for the whole population. Inferential methods of data analysis include hypothesis testing where statistician test a proposed theory based on collected data (Al-Benna, 2010). Other inferential analyzes are correlation testing and regression analysis. The following information has its basis in a case study of products marketing business, where the researchers used descriptive and inferential methods to research the problem of decreased sales.
Data collection for statistical analysis is a prerequisite for making predictions on the population. The descriptive statistical method has defined data collection procedures, which make use of different instruments. A questionnaire is a primary tool for collecting data from the field used by various research organizations. It is a list of questions devised for the statistical study of answer choices provided. It has advantages of imposing a cheaper operational cost to the researcher. Because the researcher does not interact directly with the respondent, there is the elimination of interview bias and data collected is trustworthy. The questionnaire also gives the respondent enough time to consult on the study topic and thus give a comprehensive answer. The disadvantages of questionnaire use are that to avoid confusing the respondent; interviewers provide simple questions and surveys requiring on spot solutions are impossible. The researcher has no control over the reply speed and response may delay beyond the time allocated for the research. There is use of telephone interviews for spontaneous data collection, and this method is relatively cheaper and quick. The disadvantage is that researchers can`t use the method in countries where few own telephones because the findings are biased to upper class. (Al-Benna, 2010)
In cases where the respondent is available, a researcher uses personal interviews to collect information on study topic. This method allows them to gauge the accuracy of interviewee`s information, and the researcher can guide the respondent to fully understanding the presented questions. It has the advantage of allowing persuasion of interviewee towards disclosing needed information. The disadvantage arising from this method is that theinterviewee will not willingly disclose confidential information because of the face-face embarrassment. Use of Interviews is most appropriate for studying a local population. Data collectors can alternatively use observation to collect information from the field (Besselaar 2003). The advantage is that there is actual observation of actions of the people as opposed to relying on secondary information, and the system remains undisturbed. This collection method is dependent on skills of the observer, and it is hard to make future predictions based on observations made in the present. In our marketing business, use of questionnaires is the most effective means since it can collect data from a wider range of customers. Face-face interviews are running over a short period to ensure data collection process is not nuisance to the customers.
Descriptive data analysis uses different statistical methods in presenting the collected data, such as graphical method and frequency distribution. The principal method used by researchers is a graphical presentation of statistical information in visual form. This method gives a chart picture with a scale caption to show the distribution of the data. The graphs are clear and concise, and the researcher should ensure they are well labeled and legible. (Besselaar 2003) Pictorial graphs come in different forms such as pie charts, simple bar charts, and two-directional bar charts. In making quick comparison of related data sets, researchers use a multiple bar chart. The researcher may choose to use frequency distribution to present data based on the findings they want to make known to their audience. They summarize data from surveys, experiments or census, and the pertinent descriptive information extracted. Frequency distributions provide information on required central measurements such as mean, median, variance, and kurtosis. The distributions presented in form histogram and frequency curves are easy to read and make conclusions. Frequency polygons use to display more distributions on the same axis for comparison purpose.
Sampling is the process of selecting units of a population to act as a representative study sample so that the inferences made generalize characteristics of the entire population. Researchers pick the different sampling methods based on the extent of accuracy required from the data and funds available to service the operational cost. The methods broadly categorized into probabilistic and non-probability sampling methods. Data analysts use probability-sampling methods in cases where every element of a population has a known chance of selection. It has stratified, simple, and systemic sampling sub-categories. Researchers use the simple random sampling where all members of the population have an equal chance of selection. The estimates obtained from this method are unbiased. In systematic sampling, elements of the population are arranged based on certain criteria, usually the study criterion, and members selected at regular intervals. (Besselaar 2003) The stratified sampling method involves grouping the units of a population into sub-population samples, which are independent and analyzed separately. It has the advantage of enabling researchers to make new inferences on particular subgroups. The statistical estimates made are more efficient.It, however, has a limitation. The researchers have to select different stratification variables thatare demanding. The method is expensive and is not useful when the case study is a homogenous group.
Non-probability sampling methods are efficient in cases where determination of the probability of selection of some members of a population, is impossible. The elements to act as a sample are chosen based on researcher’s assumptions and interests towards the population. Accidental sampling is the commonly used non-probability method. Selections of components of a population that is readily available as the representative sample.(Al-Benna, 2010) In Quota sampling method, the data analyst divides the population into mutually exclusive sub-groups and analyzes them independently. The main disadvantage of the non-probability methods is the non-random attribute. Exclusion bias arises because of researcher`s favored interest on certain elements of the study population.
A researcher can choose to take a combined approach todescriptive, inferential statistics to improve on data presentation. Both methods use the same data collection procedures and equipment. The difference arises in the mode of presenting collected data to the audience. Descriptive analysis tools show a quick method of comparing different data sets and spots the progress of study variable against time. It also shows data dispersion and makes a presentation that is easy to understand. Inferential tools explain the distribution of the data and give reasons for the patterns observed in the data.(Besselaar 2003) Working with a single method would not bring the entire results that are trustworthy for decision-making. A statistician will perform better in data analysis if they integrate the two techniquesfor presenting findings from their research.

References

Al-Benna,S.,Al-Ajam,Y.,& Steinstraesser,L. (2010). Descriptive and inferential statistical methods used in burns research.Burns. doi:10.1016/j.burns.2009.04.030
Besselaar,P. V. (2003). Descriptive statistics, inferential statistics, rhetorical statistics. Journal of the American Society for Information Science and Technology. doi:10.1002/asi.10304