Call Us: US - +1 845 478 5244 | UK - +44 20 7193 7850 | AUS - +61 2 8005 4826

Inferential Statistics

Inferential Statistics

The study of statistics could be generally classified in two major branches – the descriptive statistics and the inferential statistics. While the descriptive statistics helps in the description of the data, for instance charts, tables, and graphs, the inferential statistics helps in the making predictions or inferences from the data. With inferential statistics, the researcher is able to take data from the sample, carry out the analysis and make generalization about the overall data. This statistical analysis is applied when it is not possible to carry out an examination of every variable in the population. For instance, a researcher might stand in a mall and ask a sample of 100 people I they like shopping at the ABC mall. From the data collected, the researcher may develop a pie-chart of Yes or No answers (descriptive statistics), or could use the findings and reason that around 70-80% of the population likes shopping at ABC  malls (inferential stataisctics).

Categories of inferential statistics

There are two categories of inferential statistics:

Parameter estimation: this implies taking a statistic from the sample data such as the sample mean, sample standard deviation etc. the sample statistic is then applied in then applied to say something regarding the population parameter (the corresponding population mean or standard deviation).

Hypothesis Testing: this is application of the sample data to help in answering the research questions. For instance, the researcher could be intersected in finding out if the new cancer drug is effective. Or if children could perform better in school if they are given the breakfast. To answer these question, a hypothesis is developed and analysis carried out to determine if it is true or not. 

Statistical Tests in Inferential Statistics

Using inferential statistics, a sample data is obtained from a small number of people and determine whether the data predicts the behaviour of the whole population. For instance, determine whether a certain drug would work for everyone (population) by testing it on a number of people. This could be done using various ways. One way is the calculation of the z-score. Z-score would show where the data would lie within the normal distribution. Another way is the use of post-hoc or advanced testing. In addition, inferential statistics helps in the comparison of data using statistical models with other sample or the previous research. some of the models applied in this case include the generalized linear model such as the student t-statistics analysis of variance (ANOVA) and regression analysis.