Objective

After the research question is formed, and after results are

collected; the next step would be to extract meaning out of them. This is done

through result analysis, where the research team examines the results from

different perspectives and in different ways, in order to confirm or deny any

underlining assumptions relating to the research topic. In other words, data

analysis is used to interrupt the data obtained throughout the research, and

use it to gain insight about the research topic.

The Process

Data analysis is to done to prove through mathematical and

empirical means; the existence of a relationship between the dependent variable

and independent variable(s), and if such relation can’t be supported after the

statistical models are applied; then this affects the conclusion of the

research and it ultimately means that the research question and any assumptions

previously made, must be reexamined. This analysis would utilize several

methods and models of empirical and statistical nature; this is discussed in more

detail below.

Relationships

The research topic is concerned with the attitude of the workforce

towards automation, and whether they fear losing their jobs to it or not. Thus

the analysis would be done to establish a relation between the work force (as

function of the independent variables) and the fear of losing job to automation

(the dependent variable). If the

underlining assumptions are true; then one would expect to see correlation

between (for example) skill level and fear of job loss to automation.

Methods of Analysis

The research team means to employ several statistical methods to

inspect the existence of a relation between the variables. These methods

include the following:

Descriptive statistics

Descriptive statistics are basic statistical tools and concept that

give a general and over the top view of the main characteristics of the sample

and what are the most distinct traits and possible causes of error. These tools

include but not limited to:

·

Frequency

Distribution & Visual Representation (ie: Histograms)

Frequency

distribution is simply a way to arrange the data in and orderly fashion and

classify it according to the number of observations. Frequency distribution is

obtained through grouping data in an classes of adequate size and listing the

frequency of observation included in each class, and this method can summarize

data in an effective manner and allow for visual graphs to be used such as

Histogram. Lastly, it is worth noting that frequency distributions can be

applied to qualitative data sets and quantitative data as well

Mean

The

expected value of a population is known as the Mean, it also refers to the

central tendency of a given data set. This value is obtained by dividing the

value sum over number the of observations, and it can be sometimes used as an

expression of the average of a data set. The mean is useful when dealing with

data set; for it can reveal what is the overall tendency in one value (relative

to the set), which can lead to finding any skewness present in the data.

·

Standard

Deviation

The

term Standard Deviation is an expression used to indicate (in quantifiable

amounts) the amount of dispersion present in any given data set. This value

gives a measure of how close the points in the data set are to the expected

value or the mean.

·

Contingency

Table

Contingency table

is sets of values presented in a tabular matrix that show the frequency

distribution for multiple variables. They are one common way that is used for

survey analysis. They give a big picture of all of the variables which allows

for comparison between the variables and reveals initial relation between them

Factor Analysis

In the field of statistics, one method which can be used to examine

the interdependence of variables is Factor Analysis. This method is used to

express the changeability of observed variables in terms of possibly fewer

hidden or latent variables. An example of that may be lowering the variables

from eight to four due to the fact the only four are essential and the rest

change as these essential variables change. This method is useful to examine

the degree of dependency of each variable on one another; which in turn allows

to researchers to focus on the most relevant variables to the topic.

Correlation Coefficient

Correlation Coefficient is a value used in statistics that indicate

how strong is the relation between two variables. This method can be a very

useful way to find out if there is any relation between the two variables and

if so how significant it is. This means if a change in variable happens; the

other variable would exhibits some sort of change as well tough that doesn’t

necessarily means that one causes the other. If the change is proportional;

than this means that it is positive correlation, and if it is disproportional;

then it is a negative correlation.

Bivariate analysis (Trend Analysis) & Multiple

Regressions

Trend Analysis or Bivariate analysis is a way that is used in

research, engineering and since; to analyze the relation between two variables

(one of which is dependent on the other). it is a useful way to examine how

strong of a relation the dependent variable has to the independent one. This is

expressed by the degree of accuracy of a prediction of the dependent variable

based of the independent variable within a specific range of the data set. The

relation is usually expressed in terms of a linear function Y=mX+b.

Bivariate analysis applies only for two variables. For more than

two variables; multiple regressions is used. Multiple regression is similar to

linear or Bivariate regression in the sense that it is used to examine the

relation between variables, but it can be used for multiple variables. Same

concept applies in term of how good is the relation, obtained from multiple regression,

in anticipating the value of dependent variable. It also can be used to see

which one of the independent variables can best predict the value of independent

variables, and this may lead to finding out which variable is the most

dominant.