Robert Chipman
BUS308: Statistics for Managers
Instructor Edward Kaplan
September 25, 2015
Descriptive statistics
I can describe the data with the help of descriptive statistics. The data can be described with the help of various types of graphs and central tendency.
Measures of central tendency are mean, median and mode. Mean is reliable for normal data. Median is reliable for a data which has outliers. And mode is preferred when data is categorical. There are various types of graphs like pie chart, bar graph which gives us a pictorial representation of data.
Measure of variability which is also known as spread and scatter can be calculated by range, Interquartile range, variance and standard deviation. The easiest …show more content…
It includes estimation of population parameter and hypothesis testing.
In inferential statistics null hypothesis is a generic statement or certain default position that there is no evidence to establish relationship among two measured phenomena, or when there not any difference among groups. And alternate hypothesis is the claim of researcher which he wants to establish. We use a hypothesis test to make inferences about population when sample data is available.
P value in inferential statistics is the probability of null hypothesis being true, if the p value is less than level of significance than we reject null hypothesis and if p value is more than level of significance we do not reject null hypothesis.
Hypothesis development and testing
A statement about population which is to be tested is referred as Statistical hypothesis. A test of hypothesis is a statistical procedure used to make a decision about the assumed value of a parameter. We will make our decision based on observed values of a statistic. The purpose of hypothesis testing is to determine whether there is enough statistical evidence in favor of a certain belief about a parameter. We use a hypothesis test to make inferences about population when sample data is …show more content…
Choosing an appropriate test depends on following factors. Measurement of data scale, whether it is nominal, ordinal, interval or ratio. Whether the data follows Normal distribution or not which can be done using histogram, box plot, Shapiro-Wilk test etc.
I decide whether parametric test or non-parametric test have to be applied based on certain underlying assumptions. Non parametric test incudes, Wilcoxon rank-sum test, Wald-Wolfowitz runs test, Mann-Whitney U test, and Kolmogorov-Smirnov, Kruskal-Wallis analysis of ranks and the Median test etc. while non parametric test includes, one sample t test, t test for testing means of 2 independent samples, t test for paired data, one way ANOVA and repeated measures ANOVA.
Almost all exact sample tests are based on the fundamental assumption that parent population is normal. And if test deals with the parameters of the population from which samples are drawn then it is called as parametric test. If no assumption is made regarding the parameters of population, then test is non parametric. Assumptions are as follows.
• Sample observations are independent.
• Variable under study is