What is Data Analysis in Marketing Research and how to deal with it? MBA Marketing study material.

Before going into a conclusion part, Data Analysis is a mainstream process of getting an end result of marketing research. It is an obvious process in which the researcher can reach the final stage of the marketing research. This is time to examine the amount of data and information collected and draw a final conclusion from it.

Once the data has prepared for analysis as shown in the previous post, there are some basic steps which include some statistical procedures. A wide range of collected data should be accurately measured and recorded for analysis to get an accurate conclusion.  Data analysis can be done in two way that are descriptive statistics and inferential statistics. Let’s sort it out one by one:

A.    Descriptive Statistics:

Descriptive statistics is a process of describing and summarising the data in knowledgeable conclusion. This method does not allow a researcher to go beyond the information collected or assumed hypothesis. It simple describe the data to the point as it is. Following are the types of descriptive statistics:

·         Measures of Central Tendency:

Measures of central tendency tools are used to calculate an average value of data or sample. There are three types of widely using averages for research as explained below:

1.   Mean: This is the most popular average type of central tendency. The mean is to be calculated by the sum of all values of data divided by a number of values in data. It is a basic calculation of central tendency.

2.  Median: Median can be calculated by arranging the data in ascending or descending order and selecting a middle value from the arrangement as an average. In an even set of data, the average of two middle numbers can be a Median.

3.   Mode: Most frequently appeared number in data can be a mode average. In a data set, it can be possible that researcher finds more than one mode due to scattered data.

·         Measures of Variability:

Central tendency represents the single value of an average but it cannot describe the data observation fully. Measures of variability can calculate the reliability of the data observation. It also can be used to measure the differences between variable. It is more consistent than a central tendency. Types of measures of variability are:

1.    Range: Range is a basic measure of calculating an average for data set. The range can be calculated by defining a difference between smallest value and largest values from data.

2.    Mean Deviation: Mean Deviation is an average calculated from mean and median of data. It also called for an average deviation. The average can be a mean or median in this type of measure of variability.

3.    Standard Deviation: Standard deviation is very popular in and mostly used a method of getting an average of data. It is a value or an average value calculated from overall data which differ from a mean value.

B.     Inferential Statistics:

In Descriptive statistics, we have seen how to make data analysis with the data or set of data available to us. But in Inferential statistics, we can calculate, measure and analyze the information to check out the reliability and consistency to look into beyond the descriptive statistics. It allows us to compare the group of data and tests hypothesis.

In marketing research, we need the information or data for analysis, but we cannot study the whole population, so we choose a sample size as per a sampling theory. After this process, we need to make some hypothesis. In this case, a sample should be perfect which can represent the whole population. The final stage of inferential statistics is to compare the analysis of data and hypothesis to study the difference between them and make some decision about the research.  

As we can see, Descriptive and Inferential Statistics are two main steps of Data analysis which allows making a decision on research.

Data Processing and Preliminary Data Analysis in Marketing Research. MBA Marketing study material

Data Processing and Preliminary Data Analysis is a point of doing marketing research at first place in order to collect raw data and transform it into a knowledgeable form for analysis and interpret the final result. Initially, after designing a suitable Questionnaire, researcher need to take a field survey to collect the data. This blog comprise the practice of transformation of data into a logical and expressive content.

Data Processing:

Data processing simply the process of produce a data that can be convertible into knowledge. Data processing is a first step of final marketing research. After taking hundreds and thousands of survey, researcher need to check out the data that can be convert into some meaningful form. There are several steps involved in this process as below:

·         Questionnaire examination:

Researcher need to check the question formation, incorrect questions, rejection of unacceptable questions, removal of incomplete questions and also need to check past cut-offs date or missing pages. 

·         Editing:

Editing is a process of generate data which is correct and accurate. It corrects the errors detected from the data where ever possible. The purpose of editing is to ensure that each questionnaire is accurately completed. It confirms that the information collected from questionnaire is error free and ready to transfer into final analysis.

·         Coding:

Coding is nothing but assignment of numbers, symbols, or alpha to various data categories through some measurement and scaling techniques. Some information need to take into a featured category and so it need to record it by its number or symbol. Coding is a process of ensuring every category is in numerical form accurately to push into final analysis. It is usually applied on open-ended questions.

·         Transcribing:

Transcribing is entering a data into various software that can be accessible to people. Data entry software in this process should be user friendly to easily reachable to people. There are various methods and statistical formats for entering data easily.

·         Cleaning:

While doing a transcribing, cleaning of raw data is very important step of data processing. Raw data may have defective logins, number of errors which need to fix it with the step of entering data. Purpose of cleaning process is to make sure the consistency of the data which is transcribed.

·         Statistical Adjustment:

Some data need to represent visually by graph or statistical chart to show the result more effectively. Statistical adjustment is a step to ensure the graphs, charts or tables involves in data are correctly entered. Researcher also need to check the graph or chart are exactly placed on assigned location. 

Researcher can take these above steps according to the need of research.

Preliminary Data Analysis:

Preliminary data analysis is the first stage of research which mostly concerned with descriptive statistics. Descriptive statistics is the process of describing the characteristics of primary data which provides initial analysis of assumptions. This is important process for detecting any violation of assumptions in data. It gives a clear vision to researcher as to where the violation occurs. It embraces some statistical and arithmetical terms like Mean, standard deviation, range of score, skewness etc.

Nowadays, SPSS (Statistical Product and Service Solutions, SPSS Statistics) is a commonly used statistical analysis software package as well as SAS, Stata, Minitab are also widespread for descriptive statistical analysis.

In usual preliminary research process, there are three mainstream stages:

1.      Exploratory Analysis: This analysis shows a limitations to the primary data analysis and convert it into an evidence. It helps to resolve the difficulties by taking questions as a proof for final result.

2.      Developing the findings: It is a process of clean the data set from violations and generate the summery, relationship, narratives, and interpretations also recommendations for research users.

3.      Archiving: In archiving stage, data processor can keep the record of non-transient data for further analysis. Due to vast and complicated process of research analysis, researcher can only take fraction of content. So this step helps to save other inconsequential sets of data.

Hope this blog helps you to understand Primary stage of marketing research that is data processing and preliminary data analysis.

What is Hypothesis Testing in Marketing Research? MBA Marketing study material.

Marketing research process involves collecting of information through a questionnaire from a selected sample size and measuring this data by various techniques to record it systematically for further study. Now before heading forward let’s look into the most important topic that is Hypothesis. Hypothesis testing in marketing research helps to give a direction towards the study by making and testing an uncertain statements. It is a process of getting direction through the predictions.

What is Hypothesis Testing?

Hypothesis testing is a process of making an assumption or predictions of an uncertain events or statements to study the relationship between two or more variables and test it whether the assumptions are specific and correct to achieve the purpose of research. Benefits of hypothesis testing are:

·         Clear the purpose of research.

·         Gives the direction to re-analyse and rethink towards the research efforts.

·         It stats the relationship between two or more variables which are considered in research.

·         Gives opportunity to different industries to have debates.

How to write a Hypothesis?

Writing a hypothesis is very important and complex procedure. It is very vital to consider the relationship between two variables while writing a hypothesis. Let’s consider the below example:

“Girls and boys have different grades in school.”

In above statement, there is no existence of generic relationship between variables.

“Girls have more grades than boys in school”

This statements shows some measurable values that girls have “more” grades than boys. It gives you the clear direction towards the research as:

Why the girls have more grades than boys in school?

After getting a direction toward a research, let’s write a hypothesis by assuming some uncertain facts about previous statements:

“Girls have more grades than boys because girls are more attentive than boys in school.”

Now researcher have to test the above hypothesis to make sure whether it is correct or not.

How to test a Hypothesis?

There are two types of testing a hypothesis:

1.      Null Hypothesis (H₀):

Null hypothesis is the statement to believe as correct throughout the research. Null hypothesis is assumed to be true unless it has some proof to reject it. It is subjective by random cause. It is statement of equality. Null hypothesis represent as “H₀”. Let’s look an example:

“Average fare charges of a car in city is $50”

In this example, Null hypothesis can be indicate as:   H₀ = $50

If the data represent the fare charges is less than or more than $50 then null hypothesis reject the statement. So it comes to second type of hypothesis that is alternative hypothesis.

2.      Alternative Hypothesis(H₁):

The name itself states the alternatives or options for the Null hypothesis. Alternative hypothesis statements shows every possible alternatives other than null. This statement influenced by non-random cause. It accept all expect null statement. Alternative hypothesis represent as “H₁” or “H_a”. If we consider the above example:

“Average fare charges of a car in city is $50”

In this case, Alternative Hypothesis can be indicate as:  H₁ < $50; H₁ > $50 but H₁≠ $50

Both the Null Hypothesis and alternative hypothesis should be considered and stated in research before collection a data. It gives you desire and expected conclusion of marketing research.

Hope this blog helps to clear your idea about Hypothesis Testing.

What is Sampling in Marketing Research? MBA Marketing study material.

In Marketing Research, recognize and select the sample is very crucial part to collect information. This article helps to understand what is sampling? And How to identify and select a correct sample size?

What is sampling?

It’s a process of identifying and selecting a number of units from a whole population for our marketing research. Sample is a subgroup of larger population of individuals or objects.

In short, it is a process of selecting small amount of people from whole large population for study. It is more feasible, less time consuming and cost-effective to analyse or study of smaller amount of individuals rather than to study large population. Sample size is a group of an individuals for whom the questionnaire is to be designed aims marketing research.

Sampling Methods:

Generally, sampling is classified into two groups that are probability sampling and non-probability sampling. Let’s have a look how it work:

1.      Probability sampling:

In this method, every object or an individuals has an equal chance of getting selected for research. This method of sampling have many types as follows:

• Simple Random sampling: This is a basic type of sampling in which every individual or object has an equal probability to get selected. Lottery method is the best example of Simple random sampling. This type is simplest, cheapest and fastest but unreliable in some points because of its uneven population.

• Systematic Random sampling: It is modification of simple random sampling. In systematic random sampling, Object or an individual is to get selected by calculating desired sampling fraction. E.g. Every 5^th person or object, every 15 mins, every 15 houses, every 2 kilometers, every 3^rd shop etc.

• Stratified Random sampling: In this type of sampling, Whole population is divide into strata according to its characteristics and then sample is going to select randomly. In this type, stratification can take place even of selecting an Individuals. Stratification of population can be done by age group of people, income of people, and number of dependents accordingly.

• Cluster Random sampling: ‘Cluster’ is the name itself shows the meaning of this type of sampling. It is a process of forming a group or bunch of an individuals or object according to their features and within this group we can choose a set of sample randomly. E.g. there are 2000 schools in one town for research, if we apply this method, we cluster the geographic area of town where each area has 50 schools and within 50 schools we have to select every 10^th school as our sample set.

2.      Non-probability sampling:

As the name stated, non-probability sampling is a process of identifying and selecting sample set not by randomly but as a subjective.  It is much less formal than probability sampling method. It involves:

• Convenience sampling: This non-probability sampling method is used to select set of sample from those who willing to involve in research or who are volunteers to get selected.

• Judgement sampling: Judgement sampling is exactly opposite of simple random sampling. This method involve the process of selecting sample by own personal judgement and opinion.

• Quota sampling: It is a process of selecting a representative sample by dividing whole population into quotas according to their variables like age, income, location etc. Sample is getting selected by drawing its quota from each variables.

• Snowball sampling: In this method, selected sample size gives the references from their personal network with same features and characteristics. This method is supportive for organisation to get additional samples for its research.

Measurement and Scaling techniques in Marketing Research. MBA Marketing Study Material

After getting a knowledge of Marketing research and how to collect data/ information in the previous post, here is the blog on Measurement and Scaling techniques in Marketing Research.

Initially, we have to look into the meaning of Measurement and Scaling in marketing research

Measurement: It is a process of observing, recording and assigning numbers or other symbols to a characteristic of the object according to certain rules.

Scaling: It is a process of assignment of objects to a number. Scaling is developing a continuum/range/series upon which measured objects are located.

In simple words,

There are two forms of data, first comprise of quantitative variables which can be measured in terms of numbers like price, income, expense etc. and second comprise of qualitative variables which cannot be measured in numbers like emotions, feeling, sense, intelligence etc. For further analysis, the organisation needs to convert this qualitative data into the numerical form. Here, measurement and scaling techniques help to convert the data into the measurable form. Here is how:

Measurement and Scaling techniques:

1.      Nominal Scale:

This scaling technique is basic and simple to understand. It is a technique of assigning labels or numbers to the variables. In this type, there is no existence of numerical significance. There is no link between the variables and numbers or labels allocate to them. It can be called as a “Label Scaling technique”.

e.g.

Gender-    1. Male   2. Female

Marital Status-    A. Married   B. Unmarried

2.      Ordinal Scale:

Ordinal scaling is a process of ranking the objectives according to their characteristics or features. Ordinal scale helps to convert the data from unmeasurable to measurable one, so that researcher can analyze the particular value of responses and rank it as per its feature. This scale is basically used to measure non-numerical concepts like emotions, satisfaction, intelligence etc.

e.g.

How would you rank our product? -   1.Satisfied   2.Unsatisfied   3.Delighted

3.      Interval Scale:

Interval scale can be measured an absolute value or difference of values between scale points. This scale not only can rank the data but also can convert the difference in numerical form so that researcher can find correct figure to record it properly. This type of scale gives a strong figure for non-numerical responses in numbers.

e.g.

Rate our performance on a 0-10 scale- The answers can be 8.2, 9.2. 1.3, 5.6.3, 4, 6.6 etc.

4.      Ratio Scale

The ratio scale is the ultimate level of scale by its use as it allows the researcher to find or classify the object, rank the object as well as compares the difference between the intervals and gives the result in ratio form. In short, Ratio scale includes nominal, ordinal as well as interval scale pattern to measure and record the data in sequence.  It compares both distinctions in rating.

e.g.

What is the average temperature measured in the USA in last three months?-

In the first month- 20 degree- 2^nd rank

In the second month- 22.5 degree- 3^rd rank

In the third month- 10 degree- 1^st rank

Above explained techniques are the basic in measurement and scaling the research which can be used on collected data through Questionnaire as described in the previous post.

How to design a Questionnaire for Marketing Research? MBA marketing study material.

Every organization needs a valuable and truthful information to take a decision about future of the organization. Correct information leads to influence the marketer to take a right decision.

The previous article explained types of market research and how information can be collected for research. Primary/qualitative type of marketing research consist a survey method to collect data. In the survey, Organization needs to make a Questionnaire. Here is the answer of How to design a Questionnaire for Marketing Research?

What is Questionnaire???

A questionnaire is a series of questions with a choice of answers devised for the purpose of research or statistical study.

Therefore, the decisive part of good research is allied with making sure that the questionnaire design as per the need of the research.

How to design a Questionnaire for Marketing Research??

Designing a questionnaire is not as simple as it at first sight. A research team has to plan for collecting information is required to be extremely careful in deciding vital queries like:

What kind of data/information should be collected?

What kind of questions to be formed?

What type of wording should be used in questions?

Arrangement of sequence of questions

How to present a form of questionnaire?

Finalization of Questionnaire etc.

Let’s look into this step by step:

1.      Determine what information is required:

The basic step is to determine exactly what kind of information is required to accomplish the goal of a survey. An organization needs to make a list of objectives of the survey.

2.      State sample size:

This is a step of selecting a set of people or respondent of questionnaire or sample size. It means Organization should select a target customer’s sample size to whom the questions to be asked.

3.      Choose a question type:

The organization should be specific about what question type to choose, whether it will be multiple choice questions or open-ended questions or projective type questions like fill in the blacks and word association etc.

4.      Decide content of questions:

This is a very important step in which organization should develop each question with a specific purpose. Each question should be logical, easy to understand and answerable. The content of questions should be addressed to need of market research.

5.      Arrange sequence of questions:

An arrangement of questions plays a vital role in designing a questionnaire. Neutral questions should be arranged at the beginning to set an empathy to set a person at ease. The questions should be in order to bring logic and flow to the interview.  

6.      Finalise the questionnaire:

After the arrangement of questions, the form needs to format with instructions to respondents with the good introduction. It will give a clear idea to the respondent about organization’s objectives.

7.      Test and review:

After making it finalize, Questionnaire should be tested on small sample size (probably on associate or friends). Pre-test of questionnaire aims to detect faults and correct where ever it needs before going through the main survey.

And finally, your questionnaire has been designed to work. It all set to collect information/ data from a respondent.