Analysis of Quantitative vs Qualitative Research Tutorial 2019
This tutorial explains Quantitative vs Qualitative Research with best examples. Quantitative data are analyzed at the conclusion of the research while qualitative data are analyzed while research is conducted.
The data from quantitative research are in the form of numbers that can be totaled, averaged, compared and contrasted. The resulting statistics are used to describe consumer behavior or preference within a targeted population segment.
Management may not understand how the statistics were obtained, but they can easily understand the averages and percentages and, therefore, trust the resulting information.
In contrast, qualitative research results in verbal data and images including recordings, written words and sometimes photos or videos. These data cannot be statistically manipulated, compared and contrasted.
Rather than proving facts using statistics, the analysis of qualitative data has as its focus the search for meanings. This is because qualitative research is used to answer the question of ‘Why?’
The answer will always be more complicated to explain to management than just a percentage or average.
The art of qualitative research
Qualitative research, as with quantitative research, is conducted in order to answer a research question. However, a skilled qualitative researcher may find more in the data than just the answer to a research question. Quantitative research is usually conducted using a survey methodology.
This limits the responses of participants to the issues on the survey form. Because the subjects involved in qualitative research are allowed to provide any additional information they feel is important, there will be a wealth of data.
These data may provide new insights to help answer the research question. In fact, it may turn out that the subjects have an entirely different view of the solution to a problem or a new opportunity.
Because the research process may reveal unexpected ideas and opinions it is important that regular meetings are held between researchers and the management of a company during the research process.
These meetings will be to share the unexpected insights that have resulted from the research. This will give the company’s management the opportunity to discuss whether they wish to have these insights further explored by adjusting the research methodology.
The analysis of qualitative data is also unique because it starts without predetermined categories in which to place ideas or opinions.
Rather, new insights will reveal themselves as the data are repeatedly analyzed by researchers. The goal of such analysis is a full description of the attitudes, values, and opinions of consumers.
The Analysis Process of Quantitative vs Qualitative Research
The steps involved in analyzing qualitative data include the organization of the data, review of the data, coding, and analysis. The most significant difference between quantitative and qualitative analysis is that with qualitative analysis, rather than merely quantifying responses researchers are looking for patterns or themes in the data.
While the analysis of qualitative data is more art than science, there are still distinct steps to the process. First, researchers must organize the data. The next step is to transcribe any verbal information onto tape.
Once the data are organized and transcribed researchers review the material by reading through it with no preconceived idea of what will be found.
Researchers will again examine the data after this initial review, all the while coding common concepts. Once this level of coding is complete, the researchers will again review the material to find any concepts that need to be broken down into more than one category.
They will then analyze these concepts and categories by questioning what relationship they have to the research question. The final step is to interpret this information into recommendations for action.
Data organization involves both the collection and the transformation of the collected information. Organization of data according to the research question can be particularly important for qualitative research, as the findings may need to be compared with the findings from quantitative data on the same subject.
Some of the data resulting from the research might be lists that have been collected on large pieces of paper.
Focus groups commonly use this method so that everyone present can view the responses being provided to questions asked by the moderator. The lists are commonly written in marker pen in large print and then hung on walls around the room.
These lists must be taken down, labeled and dated. Each label should include the research question that resulted in the list of information. The date of the focus group and the name of the researcher should also be noted on each piece of paper.
Once these pieces of paper are back in the office, the data should be typed up. If the lists include additional notations, such as item numbering when participants have been asked to prioritize, this information must also be added to the typed list. Sometimes arrows or other visual notations will be on the paper.
If they cannot be typed they should be described in the typed document. This typed list cannot totally duplicate all the information contained on the original sheets and, therefore, these should be saved in case they are needed for analysis.
Written materials from any other projective techniques should also be collected. These may be ideas written on index cards, cartoons with speech bubbles or sentence completion forms. Drawings made by participants should also be organized and saved by having them scanned into a computer program.
These scanned drawings will also now be available to be added to the final report. Digital photographs that have been taken by participants can also be downloaded into the computer, both for analysis and for future use in the final report or presentation.
When all the material has been written down, the data from focus groups can be organized by research question. If more than one group was conducted, the information should also be organized by group. Projective data should be organized by technique. For example, all sentence completion forms and all drawings should be kept together.
Interview transcripts should be organized by research question. Intercept interviews must be organized by location and expert interviews by topic. Observational research forms are usually organized by the location where the observations took place.
Ethnographic research materials should be organized by a site visit, while the material produced by grounded theory research needs to be organized by topic
Organization of data
Focus groups organized by research question and group
Projective techniques organized by technique
Interviews organized by topic
Intercept interviews organized by location
Expert interviews organized by topic
Observational research organized by location
Ethnographic research organized by online community
The art of transcribing recordings (Quantitative vs Qualitative Research)
Qualitative interviews and focus groups will both result in tape recordings that will need to be transcribed. If possible, a word-for-word transcription should be produced.
However, if there is a great deal of taped material this may not be economically feasible. A good transcript will allow a company to experience the research process as if they were there.
It is best if the researcher who moderated the focus groups or conducted the interviews listens to the tape before transcription. While they are busy conducting the research, it is always difficult for a researcher to retain all that is happening. Listening to the tape will refresh a researcher’s memory of what was said.
In addition, the tone of voice used and even the silences between speech can provide insights.
After listening to the tape, transcribing the material is the next step in the analysis process. While a researcher is listening to a tape, they can be typing up notes on the main points being made.
Once the transcription process starts, the researcher can pause the tape while they make notes on new insights or memories. For example, there may have been a focus group member who spoke rarely but followed the proceedings closely with evident interest.
This gives more weight to his or her opinions than if the researcher remembered that the focus group participant seemed bored and distracted. It is difficult for researchers to take notes on behavior during a focus group or interview, so these notes should be added to the transcript.
The same or different participants will frequently repeat information, and a researcher can quickly pick up patterns and develop a shorthand notation process.
Transcribing tapes is much more than just a technical task, which is why it is best if a researcher prepares a transcript personally. The goal of transcription isn’t a word-for-word perfect transcript.
Instead, while the transcription is taking place, a researcher should concentrate on recording information that addresses the research question. While the conversation may stray and other topics be discussed, these topic areas are important only in how they relate to the main research question.
For example, if the research concerns the cost of a product, such as cell phones, participants may start discussing where they like to shop.
A researcher will then analyze this conversation for its relevance to the research question on cost, by noting whether the stores mentioned are discount outlets or high priced specialty stores. Information on a great sale on shoes that was added by one participant can be ignored.
When transcribing a tape, it is not important to note the names of speakers as they do not need to be identified personally.
However, it is still important to attribute comments to individuals to determine if there is consistency in the comments they provide throughout the focus group. Instead of names, a researcher can add a code or a number to represent each speaker.
In addition, comments on the tone of voice or emotion can be added using the same system. One means of adding such detail is to type the transcript in a column format. The first column will have the words spoken during the session, whether the interview or focus group.
The second column will contain the researcher’s notes on who is speaking, their tone of voice and any other observations noticed by the researcher when the research was conducted.
The third column will be used for coding purposes. Because this column transcript will be referred to frequently during the analysis process, it is important to leave additional space where handwritten comments can be added.
Coding Qualitative Data
Once the transcriptions are complete, researchers should review all of the data. This review should be conducted with an open and relaxed frame of mind. At this stage of the analysis, researchers must let the data reveal insights rather than impose ideas that were formed while conducting the research.
While the impressions formed during the research are important and should be retained, it is also important for researchers to look at the data with fresh eyes. It might be that comments and ideas which were initially overlooked can now be seen as being important.
Once data have been transcribed and reviewed, researchers will begin to code. Coding is used to note the repetition of ideas, opinions or facts. The first coding will be conducted to examine the data for answers to the research question.
For example, the research question might have asked how a visit to the dentist could be made more pleasant. A focus group of clients would be asked for their ideas for improvements that could be made to a dental clinic.
A transcript would be coded for the times when any mention of the ideas for improvements was mentioned. These instances are coded so that researchers can then return to the information to analyze if many of the responses gave similar ideas or if any unique suggestions were provided.
The transcribed notes will then be analyzed again to code other topics that arose during the research. For example, besides discussing ideas for improvements, researchers might find that on multiple occasions trouble reaching the dental office due to a lack of convenient parking or public transportation options was mentioned.
Another issue discussed might have been the services now offered at a new, competing, dental office. By coding these data, researchers may find that it is this new competition that received the most mentions. Below is an example of how data are being collected from online sources.
The first step in the analysis of all of the written material is coding for the main concepts that appear in the transcripts. Researchers read through the documents to see if there are any concepts that are raised repeatedly.
If a study involved the reasons why students leave the university before completing their studies, the main concepts that might appear could be ‘money’, ‘studies too difficult’ and ‘unfriendly staff’.
However, a more surprising concept that might be discovered would be ‘got a good job offer’. This analysis takes skill, as the wording used in each individual comment may not directly describe the concept and it will certainly take more than one reading of the material by a researcher before all of the concepts become clear.
It is a researcher’s responsibility to notice the similarities in comments and that they may all belong to a single concept.
Categories (Quantitative vs Qualitative Research)
Once researchers have finished coding for concepts, they may find that some need to be further broken down into categories. These concepts and categories are important as they are the building blocks from which researchers will make their recommendations for action.
For example, many comments in a transcript may involve the concept of the price of a product. Several participants may state that they don’t buy a product unless it is on sale.
Other participants may state that they buy a competing product because it is cheaper, while some may state directly that the price of a product is too high. While all of these involve the price of a product, a researcher may decide they are too dissimilar and break them down into three categories: ‘don’t buy because can’t afford’, ‘competing product purchasers’ and ‘non-purchasers’.
The researcher may then make different recommendations for attracting each of the first two groups and recommend no action on the third.
Researchers can code transcripts using highlighting markers with a different color for each concept and category, or alternatively, the concept and category name can be noted in the margins of a document.
As researchers code, the material for the answer to a research question, each time a researcher comes across a statement that deals with the issue of the price being more than a consumer is willing to pay this will be coded.
However, with qualitative research, there may be other concepts that arise that are worth noting (which is one of the benefits of qualitative research). For example, in the discussion on price, a researcher might note that several consumers expressed how much they enjoyed the design of the product.
While the research question had not addressed design issues, this would still be useful information for management.
Using coding to develop recommendations
Once coding is completed, all material will be reviewed again to develop recommendations based on the coded concepts and categories. For example, a research question might have asked ‘Why do consumers not purchase automobiles produced by our company?’
The coded material may have revealed infrequent comments made on color, appearance and amount of chrome. All of these comments the researcher will code under one concept – ‘style’.
Other comments made about the cost of the automobile a researcher will code under the concept of ‘price’. Further analysis might now reveal that the concept of ‘price’ is actually two categories.
One involves comments on the cost of the automobile, while the researcher might find that a separate category is now needed for those comments that involve the cost of maintaining such an automobile, including comments about gas mileage, insurance and repair costs.
Based on this coding researchers might recommend that promotional material should address the reasonable cost of maintaining the vehicle and not just the low purchase price.
Software tools for coding (Quantitative vs Qualitative Research)
Software tools that assist in the analysis of qualitative data are now available. However, marketing researchers must decide if it is worth the money to purchase such software.
If a research process has only involved one or two focus groups or interviews, the time saved in using software may not justify its cost and the time it will take for researchers to learn to use it. In this case, researchers may decide to rely on hand coding and analysis.
If researchers conduct qualitative research on an ongoing basis or have a large qualitative research study planned, then it may be worth their while to purchase and use coding software.
While these software packages can save researchers the tedium of coding, they do not replace the analytical process of determining the concepts and analysis of the concepts and categories that result in recommendations.
A software package designed for qualitative research will help with the development of a system of coding and then applying the system to the transcribed text.
It will also allow researchers to add brief comments to the data. These comments might explain nonverbal behavior that a researcher had noticed while the comment was being made.
Researchers will also be able to link different codes between transcripts, memos, and notes on different focus groups, interviews, and projective techniques.
The software can also help in the preparation of final reports by displaying coded categories in a graph format. If there is a sufficient amount of data, once it is coded the number of responses that include a particular type of content can be quantified for management.
This process is started by a researcher selecting text in a transcript that belongs to a specific concept, which the researcher then names. Once this task is done the software will then search all the data files for identical text and code these occasions as the same concept.
For example, the comment that a product is ‘too expensive’ will be coded as belonging to the concept of ‘price’.
The software will search all the transcripts and other written material for similar comments. In addition, the software can search for additional phrases that have the same meaning. Terms such ‘costs too much’ and ‘not worth the price’ can be added to the same concept.
Computer-assisted qualitative data analysis software (CAQDAS) while not new, is becoming increasingly sophisticated. The more advanced programs can now analyze and code not only text but also audio files. The purpose of the software is to enable researchers to analyze across numerous files of different types to find common themes.
Analysis of Qualitative Data Content
After data have been organized, transcribed and coded, the next step in the process of analysis is to determine if there are any relationships between the concepts and categories.
The purpose of developing relationships is to generate new ideas to answer the research question. These new ideas will be the basis for making recommendations for action.
After all, management will want actionable recommendations from a study, not just analysis. A report that simply describes researchers’ impressions will not be considered useful enough to justify the cost of the research.
Possible recommendations might involve how to target new types of consumer segments, descriptions of the process of consumer behaviors, a comparison and contrast of consumer motivation, or a new research question of a relation between variables that will need to be verified by future quantitative research.
Coded qualitative research data can be analyzed for information on possible new market segments to target.
A company may be aware of how to market their product to their current demographic and geographic segments, but qualitative data might reveal entirely new psychographic segments of which that company was unaware.
These new segments will have been identified based on common values and attitudes that have been verbalized or displayed during the qualitative research process.
For example, a research focus group on a product for older consumers might have found that people aged 65–75 years old do not consider themselves as being older. Because they are still leading active lives, this segmentation category based on age may have no meaning for them.
Instead, they may identify themselves as ‘active adults’ who just happen to be retired or on their second career. They may also think that they have no attitudes in common with people in the traditional category called ‘senior citizens’.
Likewise, qualitative research data may uncover segments of individuals who identify with various types of hobbies. What they will all have in common is a specific interest, say in crafts, and will, therefore, identify themselves by this, for example as ‘crafters’.
As a result, researchers might recommend that a company commissioning research on this topic considers producing products aimed at this new segment.
Qualitative research might also uncover new usage categories. In discussions of food consumption, it might be found that food ordinarily consumed at breakfast is also enjoyed at other times of the day.
Based on a finding that cereal is also eaten at the office, researchers might recommend a new promotional campaign based on this usage.
Consumer behavior processes
Besides new market segments, qualitative research can provide insights into consumer behavior processes. A company that makes readymade dinner entrées may be interested in the meal preparation processes of today’s busy dual career families.
Analysis of ethnographic data might reveal that parents would like to have everyone sit down for meals together, but that children have their own dietary preferences.
Using this knowledge, researchers could recommend that a company produces prepackaged dinners with a choice of side dishes so that everyone can eat together and yet still have the food they each want.
If it is found that parents still want to have their families maintain a little formality when dining, researchers might then recommend that the packaging includes decorative paper napkins.
An observational study on how people drive their cars might find that drivers need cup holders that can keep their beverages hot or cold when they spend long periods of time in the car. In addition, observing children traveling in their car seats may have revealed a need for a small storage area for their food.
These are ideas that might not otherwise have been discovered in quantitative survey research. However, analyzing the data from qualitative research can reveal useful ideas such as these that can be recommended to companies.
Comparing and contrasting consumer traits
While researchers are analyzing data, they may note some differences in the consumer behavior process based on demographic or psychographic traits. For example, a qualitative research study might have been specifically designed to examine and compare the differences in cell phone usage for different age groups.
These types of differences will appear in qualitative data from focus groups, interviews or ethnographic studies.
In this case, researchers might perhaps find that women were using the photo feature to take pictures while shopping of possible purchases for their home that they can then view later.
Meanwhile, it may also be found that males were using their cell phone cameras to take candid photos of their friends. These are ideas that can be developed into recommendations.
Development of new research questions
Another recommendation that may result from an analysis of qualitative data is a new research question about the relationship between two variables. This research question cannot be said to be proven, based on qualitative research.
However, it might be so intriguing that the researchers recommend quantitative research be conducted to determine the answer to this question.
For example, qualitative research might find that the consumers who are nonusers of a product believe that the product is too expensive to operate. This fact could then be tested further with survey research.
The information provided by observational research will not be in a verbal format. Instead, the data will be in the form of notes on behavior, photos or video. Observation forms and notes must also be analyzed, but not by coding for words. Instead, researchers will be looking for unique or repeated behavior that has been noted on the forms or in the photos or videos.
Researchers can look for these data concerning the process of using a product, new ways of using a product, where consumers use a product and the mistakes they make when using a product – all of which may have been noted on the forms.
For example, observational research of consumers shopping at a clothing store can show how they travel through the store, which products they tend to buy first, and how long they spend in the store. If researchers notice that people seem to have a problem finding the fitting rooms, better signage may be recommended.
Often netnography research may reveal that people use a product in a way that was not originally intended by the company that designed that product. These insights can be used to make recommendations on the redesign of a product or the development of a totally new product.
For example, a netnography study may have been conducted on students living together in the university-owned housing. Analysis of discussions on sites used by students may have found that students like to study while lying on their beds. From this study, it might be recommended that better lighting is provided above beds.
1. The differences between analyzing qualitative and quantitative data include the fact that the analysis of quantitative data results in statistics that describe behavior. However, qualitative data are analyzed for insights into the motivation for behavior.
Quantitative data are analyzed at the conclusion of the research while qualitative data are analyzed while research is conducted. The analysis of qualitative data is an art that relies on the knowledge and skill of researchers.
The analysis must only be conducted by researchers as they alone will have experienced the incidents that occurred during the research. In order that these incidents are not lost, researchers should hold debriefing meetings as soon as the research study has been concluded and even during the research process.
2. While qualitative analysis is an art, there is still a process to be followed. First, the data must be organized and any verbal information transcribed. The data are then reviewed and coded for concepts and categories. Finally, the relationship between concepts and categories is questioned and the findings interpreted into recommendations for action.
Data are organized based on the methodology and notes are then transcribed. This transcription can be verbatim or in note form. The transcription should be in a format that allows researchers to easily add insights and coding. The transcription is then reviewed for insights.
3. The most important step in the qualitative analysis process is the coding of the data. Both repeated and isolated incidents and comments are coded by theme and named as concepts. This can be done by physically marking the words and then distinguishing the type of comment by words or colors.
This coding will be used to build categories with common elements. There is now software that helps to make this task more manageable, but the ideas for the coding of concepts and categories must first come from researchers.
4. Analysis of coded data will include questioning the relationship between categories and looking for insights that can be interpreted to answer the research question. The interpretation might reveal information on new potential consumer segments.
It also might reveal information on consumers’ behavior processes. Consumers could thus be analyzed for an interpretation of traits. In addition, new research questions may be determined. Finally, analysis of nonverbal ethnographic and observational data can be used.
Analyzing Numerical Data
Does Everyone Shop on Amazon?
In the United States, the answer is almost yes! Only 17 percent of the primary shoppers in American households have never bought a product on Amazon. How often do Americans shop on Amazon? Twenty-eight percent shop weekly, 25 percent one to three times a month and 30 percent less than once a month.
Who doesn’t shop? They tend to be older: with an average age of 57 for non-shoppers versus 49 for shoppers. In addition, they earn less, $45,700 to $62,800. The question is why not?
After the survey results were analyzed, additional interviews with nonusers of Amazon have conducted that uncovered three reasons why people do not use Amazon.
First, some people enjoy browsing in stores, Second, some people find a large number of offerings for a single product more irritating than helpful. Finally, there are people who do not use Amazon because it is too easy to use and therefore they spend too much money!
Question: How could these research findings be helpful as Amazon is designing its first physical stores called Amazon Go?
Measuring Differences (Quantitative vs Qualitative Research)
Quantitative marketing research methods analyze consumers’ current or future behavior which can be expressed using numbers or percentages. When analyzing the data from quantitative research, consumers’ physical characteristics such as gender, age, religion, ethnicity, income, education level, or even their height, hair color or weight, can be quantified.
In addition, consumers’ behavior can be quantified by the frequency of purchase, consistency of purchase, place of purchase or size of purchase.
Using statistical analysis researchers will explain behavior using numbers rather than words. Furthermore, if the sample for a population is sufficiently large and properly selected, researchers will be able to say with some certainty that the research findings are probably true for the total population.
Scales of measurement
While a computer will be used to tabulate the results of a survey, the findings will still need to be analyzed by the researcher. Before the methods of analysis can be discussed it is important to understand measurement scales, as the type of scale used in designing the answer will affect the type of statistical analysis used.
People use measurement scales on a daily basis. When consumers decide to purchase a rug for their bathroom floor, they may decide to use their own feet and pace out the floor space to be covered.
When they go to the store they can again measure off the space using their feet and find a rug that fits. Of course, it is simpler to use a measuring tape already marked off with standard units of measurement.
In research, it is also easier to use standard measurement scales. The four standard measurement scales researchers have available are nominal, ordinal, interval and ratio. The choice of statistical procedure to analyze the data will depend on the scale being used.
The nominal scale
The nominal scale is used for characteristics that can be defined as different states of being. There are only the two states and a research participant must be one or the other.
University graduate and non-university graduate are an example. A research participant can be one or the other, but not both. Nominal data are usually analyzed by simply counting the responses.
The report might also state that of the 100 participants, 78 did not graduate from the university while 22 did graduate. When analyzing central tendency, the mode will be used.
Ordinal measurement scale
The ordinal scale is used when there is not just an absolute difference, but rather a degree of difference, such as preference. A question using the nominal scale would be ‘What type of pizza do you prefer?’
This question, using an ordinal scale, would ask consumers to rank order their favorite types of pizza. Analyzing these data would provide a ranking, such as participants’ favorite type of pizza topping is pepperoni, followed by olives and then sausage.
While this type of measure provides information on ranking, what it does not show is how much more popular pepperoni pizza is than olive or sausage. When analyzing central tendency, the median will be used.
The interval scale
Using an interval scale will provide more information than just ranking. An interval scale adds a unit of measure with a start and finish, and the difference between each unit of measurement being the same.
Researchers construct the scale by creating the starting and end point of the scale and the units of measurement. In this type of ranking, participants are given a choice of degrees. The question might ask if the pizza was very delicious, delicious, good, alright, or inedible.
While the construction of the scale is arbitrary, it is assumed that the amount of difference between ‘delicious’ and ‘very delicious’ is the same as the amount of difference between ‘good’ and ‘alright’. With this type of data, researchers will be able to provide an average opinion of the pizza by all survey participants.
When constructing questions that will provide interval data, researchers can use a category rating scale, Likert scale or differential scale. A category rating scale gives a rating such as excellent, very good, good, poor, very poor, awful.
On a Likert scale, five to seven choices are given that measure a participant’s opinion or agreement with a statement. This could strongly agree, agree, undecided, disagree, strongly disagree. A number is attached to each category that allows researchers to provide an average.
A semantic difference scale is also commonly used in interval measurement question construction. In this type of measure, two opposing statements are made with seven points in between. A participant chooses where they stand on the subject by circling a number.
The ratio scale
A ratio scale has given start and end points that already exist and are not created by researchers. A consumer’s weight is an example of a ratio scale, as no person’s weight can be zero pounds.
There is also an upper limit on what a human can weigh and still survive. On a ratio scale, there is also the ability to measure exactly the difference between units. A person who weighs 200 pounds is exactly twice as heavy as a person who weighs 100 pounds.
A person who spends $900 on food per month spends exactly twice as much as a person who spends $450 on food. The use of the ratio scale in marketing is not as common as the use of the interval scale because marketing often asks questions about people’s preferences and opinions using an interval scale. When analyzing central tendency, the geometric mean will be used.
The Process of Quantitative Data Analysis
After conducting a survey, researchers will be faced with either a pile of survey forms or, if the survey was entered directly into a computer, an electronic file. Researchers must now begin the task of analyzing the data.
This process of data analysis will begin with a pre-analysis stage where researchers will review the data, including its validity, completeness, and accuracy. They will then code any open-ended questions and enter all the data into a computer software program.
Using an online form can save time at this stage because the data will already have been entered into a computer. Another advantage of online surveys is that researchers can monitor the data as they arrive.
However, the electronic file still needs to be reviewed for completeness and any responses to open-ended questions must be coded. This is also true of online forms that ask for complaints or suggestions.
After the pre-analysis, the next step in the statistical analysis is determining the frequency of responses. The analysis may be of a single variable, or the frequency of more than one variable could be cross-tabulated.
Once this has been completed, the data are analyzed for central tendency by calculating the mode, median or mean. Researchers will also want to know how widely the participants’ responses varied from each other, so dispersion will be examined by calculating standard deviation, range, and variance.
Finally, statistical tests will be performed to determine if differences in the data are due to chance or random error, or if such a difference has a statistical significance.
Statistical testing can include chi-square, t-test, ANOVA, correlation and regression. This statistical testing will be used to determine if the hypothesis is supported.
For example, the researchers may have hypothesized that more men than women attend the cinema. If the data show that men buy fewer theater tickets than women, statistical testing will confirm if this difference is significant enough for the alternative hypothesis to be supported. A study may contain more than one hypothesis as long as questions are written to address each.
Pre-analysis of survey data
Once survey research has been conducted, the first task confronting a marketing researcher is to review the questionnaire forms. After this review is completed, the researcher is ready to code any open-ended questions. The final step in the pre-analysis process is to enter the data into a computer software program.
Researchers may have taken great care in deciding upon the research question. In addition, the questionnaire form may have been painstakingly written and tested. However, unless the survey sample was very small with only a few participants, the actual conducting of the survey will have been out of the control of researchers.
For an administered survey form, assistants will have been hired to conduct the survey. Self-administered forms will have been completed without any help from researchers or assistants. People may start an online form, answer only two or three questions and then quit. Because of these facts, the survey forms must be checked for validity, completeness, and accuracy before the data are entered.
The issues that arise when validity is considered are whether the survey was actually conducted by an assistant and whether a participant was eligible for the survey. It is unfortunate but true that sometimes those people hired to conduct surveys may actually have completed the forms themselves.
This may be due to frustration, because of an inability to obtain the cooperation of potential participants, or it may be due to dishonesty. Whichever might be the case, all forms should be checked to see if it looks as if an assistant has completed them.
Clues would include answers that are extremely random or answers that are constantly duplicated. In addition, the demographic portion of a form should be checked to ensure that any eligibility requirements, such as age and education level, have been respected.
Not all survey forms will have every question answered and so should be checked for completeness. A form may be incomplete because a participant chose not to answer some of the questions or it may not have been completed because of time constraints.
In addition, the form may have consisted of more than one page and a participant may not have noticed additional questions on a second page. Researchers must then decide what percentage of completion is required for a form to be included in the study or for it to be discarded.
Lastly, forms must be checked for accuracy. Researchers must review the forms to determine if the answers can be read and understood. They should be able to easily distinguish which answers have been marked. In addition, the answers to open-ended questions should be able to be deciphered.
[Note: You can free download the complete Office 365 and Office 2019 com setup Guide.]
Coding survey data
After checking for validity, completeness, and accuracy, the forms are ready for data entry. Since surveys are a form of quantitative research, the data need to be entered as numbers that can then be statistically analyzed. Close-ended questions should have been pre-coded on the survey form with the numbers that will be entered into the computer.
However, open-ended questions will require manual processing before they are ready to be entered into the computer. To accomplish this task, researchers should begin by listing all the answers that were given to an open-ended question and then consolidate the responses.
For example, if a question asked for reasons for shopping at Crossroads Mall, the responses might be consolidated into the groupings of ‘convenient location’, ‘variety of stores’, ‘parking’ and ‘special events’. Each grouping is then assigned a numerical code and it is this code that is entered into the computer.
If a survey is web-based, the data entry step is not required as the answers are entered automatically. Data entry for traditional paper survey forms can be performed by anyone with the skill to use a computer. A software package can be used for statistical analysis.
Data Analysis Using Descriptive Statistics
Once the data have been entered into the computer software program, marketing researchers are now ready to start the process of analyzing the data. They should never forget that the purpose of the analysis is to provide information that can be used for making strategic decisions.
There are two types of statistical analysis that can be used and these are descriptive and inferential. Descriptive analysis collects, summarizes and presents a set of data. This type of analysis is simple for researchers to conduct and for management to understand.
Descriptive statistics help researchers to see patterns in research data. A basic concept used to analyze consumer characteristics and behavior is frequency, including one-way frequency and cross-tabulation.
Using frequency researchers can identify how many participants’ responses were similar. A second concept is a central tendency, which includes the mode, median and means. In addition, dispersion of central tendency must be examined including range, variance, and standard deviation.
Frequency: one-way, cross-tabulation
Central tendency: mode, median, mean
Dispersion: range, variance, standard deviation
Frequency can be understood using the example of a survey question that asked why consumers shopped at Steve’s Sporting Goods store. The first issue faced by the researchers helping Steve is that not all of the 1,000 survey forms distributed may be usable as some of the participants may not have answered the question, while others may have attempted an answer but the mark they made is difficult to interpret.
In addition, some respondents may not have completed the questionnaire. As a result, of the 1,000 survey forms distributed perhaps only 950 will be useable. When the data are analyzed for one-way distribution (how many people responded to each potential answer to the question) it may be found that helpful service is the most frequent response.
This information will quickly inform Steve of the ranking of the responses, with helpful service being top, followed by best selection and then good prices. However, it is difficult when just reading the numbers to understand how much more important to the participants was service over the selection.
Adding a percentage to the table makes the relationship between the responses easier to grasp as most people can see the relationships between percentages more quickly than those between raw numbers.
The percentage is calculated from the total number of responses used (950) rather than the total number of respondents (1,000). The percentages show that helpful service is almost twice as important as the best selection.
The survey question asking consumers for their motivation for shopping at Steve’s could have been designed to allow for multiple responses. The question would have read ‘Which of these reasons are why you shop here?’ In this case, the numbers would look different as some people might tick more than one response.
Because the question allowed respondents to tick more than one answer, the total response is now 1,095. However, the percentage is still calculated on 950 respondents which is why when added together, the responses total more than 100 percent.
People often use the term ‘average’ when they are referring to a middle ranking. However, there are a number of ways to measure average or central tendency that include mode, median and mean.
Mode refers to the response that is the most common for all participants. The mode is used when describing nominal data, which can have one of either two states of being, but not both. In the research for Steve’s store discussed above, the respondents may have also been asked their gender.
The question might have found that of the 950 survey participants who responded to the question on gender, 505 were male and 445 were female. Obviously, gender cannot be averaged, as the answer would always be exactly half or 500. Instead, the concept of mode (most frequent response) is used, with the mode being male instead of female.
Median is the response that measures the halfway point of the responses. Median is used in ordinal data, where there is a degree of difference.
In the question of the motivation for shopping at Steve’s, best selection is the median response as one answer received more responses and one answer received fewer responses. Median cannot be used when analyzing nominal data as there are only two possible responses so there cannot be a midpoint.
Mean is the average of all of the responses. The mean is calculated by adding all the responses and then dividing by the number of participants. If the survey participants were asked their age, it would be simple to determine the mean age. The ages of all the participants would be added together and then divided by the number of responses, or 950.
Central tendency measures (Quantitative vs Qualitative Research)
Mode: the most frequent response
Median: the response that divides a series of responses in half
Mean: the average of the responses
One of the issues that researchers must analyze is how varied the responses are from the calculated mean. To do so researchers use the concepts of range, variance, and standard deviation. These statistical concepts allow researchers to compare the dispersions of two sets of data.
While two sets of data may at first seem similar because they have the same mean, researchers know that the individual responses that comprised the mean may be dispersed very differently.
The ways to examine this issue of dispersion of responses include range, variance, and standard deviation. The amount of dispersion may depend on the way the rating or ranking question was structured.
The range is the easiest dispersion measure to understand and tells researchers how widely answers are dispersed. To calculate range, the smallest value expressed in the survey is subtracted from the highest value. This gives the range of responses.
The variance of the set of numbers around the mean helps researchers to understand how dispersed each individual response is from the mean. One way to calculate this number would be to subtract each individual data number from the mean and then add the differences. However, this will not work because the negative numbers and the positive numbers will always cancel each other out and the answer will be zero.
To solve this dilemma, the difference between each individual data number and the mean is first squared and the answers are then summed. The final step is to divide the sum by the number that is one less than the number of responses.
This step allows researchers to compare the variance between datasets that have different numbers of responses. The final number calculated is the variance.
The higher the variance, the more dispersed are the responses in the set of data. As can be seen by looking at the data, the males’ spending pattern is more dispersed. The problem with the variance number is that being squared, the number no longer has any meaning.
If the square root of the variance is calculated the answer will be the standard deviation, which is in the same units, currency, as the original numbers.
If the standard deviation is added and then subtracted from the variance, this tells researchers that this range is where most responses will fall.
Data Analysis Using Inferential Statistics
The other type of statistical analysis that researchers can conduct uses inferential statistics. These statistical methods go beyond just describing the data discovered during the research. Of course, no marketing research study that uses a sample can ‘prove’ anything with absolute certainty.
What the analysis of quantitative research data can do, however, is indicate whether a hypothesis is most likely to be false.
Using inferential statistics, researchers can perform statistical tests to determine if responses from a sample can be used to draw conclusions about an entire population. In fact, more than one statistical test can be conducted on the same set of data.
Statistical testing process
The first step in using statistical analysis to indicate the truth of a hypothesis is to state the hypothesis or guess, about some characteristics of consumers or their behavior. The research methodology will then be designed to ensure that these characteristics, whether about people or their behavior, are measured.
Once the research study has been completed and the data entered into a computer, the measured variable for the sample of participants will be compared with the expected outcome stated in the hypothesis.
The type of test that will be used to determine if the difference is significant depends on both the type of measurement that was used and the type of resulting data. These tests can be used on their own or in combination.
The z-test is used to determine if the differences in proportions or mean of characteristics are statistically significant or not, while the t-test also looks for statistical significance but between the means of two unrelated groups.
The term hypothesis is used when describing the scientific method. This is a process that attempts to determine if something is true. A hypothesis is simply a statement of fact that needs to be proved using statistics. After all, if there were no scientific method all that would be left is a person’s opinion or idea.
A business cannot spend money on developing a new product or starting a new marketing campaign based only on someone’s opinion that they will be successful! The cost of failure would be too great.
Of course, it is impossible to absolutely prove that the outcome will be successful. However, conducting research and then analyzing the results will help.
A hypothesis is a guess that is made by the company or individuals commissioning the research. Perhaps an academic publisher has come up with an idea for a new textbook that can be easily read on a smartphone. The question is whether they should spend the money to develop and introduce the product.
Qualitative research has indicated that many students would be interested in this product. However, the finance department of the company has stated that at least 20 percent of current student textbook purchases will need to purchase the new product to cover development costs and make it financially viable.
This first hypothesis is the null hypothesis and will be stated as what the company does not wish to be true. (The symbol H0 is used to designate the null hypothesis.)
The null hypothesis is considered true until proven false. For the publisher in this example, the null hypothesis is that fewer than 20 percent of students will be interested in purchasing the product.
The alternative hypothesis would be that 20 percent or more of students will be interested in purchasing the product. (The alternative hypothesis is designated H1.) One hypothesis is the opposite of the other and so both cannot be true. A quantitative survey will then need to be conducted to test the hypotheses.
Formulas for stating the hypotheses
H0: π = 0.20
H1: π = 0.20
This is an example of a simple hypothesis. It only contains three elements: the population (students), the variable (new product) and the outcome (sales). A complex hypothesis adds more variables to the question.
The company may want to know how pricing will affect the sales of the new product. So this may be added to the research question, which becomes more complex because now two variables are involved. First, do they want the product and how does price affect this desire?
The hypothesis can also be directional. It can be stated that students will not purchase if the cost is 20 percent higher than the current textbook. The statistical tests cannot be used to prove the hypothesis true.
This is impossible as the only way to know with 100 percent accuracy if a hypothesis is true is to survey the entire population.
If the null hypothesis is proved false, then the alternative hypothesis (that 20 percent or more of students will be interested) can be accepted as being true.
The null hypothesis needs to be expressed in such a way that its rejection leads to the acceptance of the preferred conclusion – developing the new product. These stated hypotheses are an example of a one-tailed test, the kind most commonly conducted in marketing research.
The publishing company surveyed a sample of 1,100 students (more than the sample size of 1,024 that would have been needed to make the study valid at 95 percent confidence) and found that 22 percent stated they were interested.
While this is over the required 20 percent, researchers know that taking a sample will never be as accurate as asking everyone. However, the question remains – if 22 percent is so close then is it simply an error that made it over 20 percent?
Therefore, the next step is to calculate whether the difference between the hypothesized outcome and the survey outcome is statistically significant. While the word ‘significant’ usually means important, in statistics it means ‘true’. The test to find out if it is significant would be automatically calculated by a statistical computer software program.
However, the formula is actually easy to understand. To calculate the significance all that is needed is three numbers: the hypothesized percentage, the sample percentage and the standard error of the percentage.
Researchers already have two of these, the hypothesized and sample percentages. To calculate the standard error of the percentage, the researcher could use the following formula, but of course, a computer program can provide the answer more easily.
The formula for calculating the standard error
= √π (1 −(π)/np
= √0.20 (1 − 0.20)/1100 = 0.010
Using this number as the standard error, the Z-score can be calculated.
Formula for calculating z-score
z = (p − πh)/p
2.0 = (0.20 − 0.22)/0.010
This z-score (sometimes referred to as the p-value) can be compared with the numbers found on a table of z-scores to determine if it indicates that the null hypothesis is not true. It is standard procedure to have the computer software do the comparison.
However, a rough calculation can be done by remembering the standard numbers for confidence levels. For a 95 percent confidence level the number was 1.96 and for 97 percent confidence, 2.58.
The z-score of 2.0 tells the researchers that they cannot say with 95 percent confidence that the null hypothesis is not proved false. Therefore, the company will not go ahead with production.
Interestingly, if the company wanted to be 97 percent confident – the company would not start production. The same type of calculations can be done for comparing a hypothesized mean and the mean that was found by surveying the sample.
Steps in the analysis process
1.State the hypothesis
2.Conduct the research
3. Compare the measured value with the hypothesized value
4. Decide the necessary level of confidence
5. Choose a statistical test for significance
6.Calculate the test value
7.State a conclusion and any recommendations
Level of confidence
Everyone is familiar with political polls that try to predict the outcome of an election. It might state that based on a survey, 63 percent of the population is in favor of a change in the government.
It will also include an error percentage such as plus or minus 5 percent. No one can know the outcome of an election before it happens because not everyone can be surveyed.
Nonetheless, the survey taker may be 95 percent certain, or perhaps 97 or even 99 percent certain that the survey results and the actual election will be the same. The level of confidence will depend on the number of people surveyed.
The possibility that the null hypothesis will be rejected as false when it is indeed true is called a Type I error, which is signified by using the lower case Greek alpha ( ). The amount of possibility that a Type I error has been committed is called the level of significance of the statistical test.
Researchers must decide on the amount of risk they are willing to tolerate of committing a Type I error. There are standard levels of risk that are considered acceptable when conducting statistical analysis. These standard levels, or value of , are 0.01, 0.05 or 0.010.
Another way to express these values is that there is a 1 percent, 5 percent or 10 percent chance of the hypothesis is rejected when it is indeed true. The traditional value used by researchers is 0.05, or there is a 5 percent risk that the null hypothesis is false, but it isn’t rejected.
Another type of error, Type II, happens when the null hypothesis is not rejected when it should be. The Greek letter beta ( ) is used for this type of error.
A statistical test to check for Type I errors is called a one-tailed test, while a statistical test to check for Type II errors is called a two-tailed test. Most researchers will only use a one-tailed test.
The chi-square test is used for what is called ‘goodness-of-fit’ when analyzing frequencies of responses in a frequency table using cross-tabulation. Marketers often want to know if there is a relationship between a specific group of consumers and some preference for a product benefit.
Marketers also may want to know if men or women prefer the product in a smaller size bottle, or they may want to know if young consumers would prefer the product in a new color.
While these statements could be presented as hypotheses this is not necessary. In fact, when using chi-square all researchers need to do is think – the computer will handle the rest.
Using the example of the publishing company and the textbook, the company believes that men will be much more interested in the product than women. The researchers will use software to calculate a cross-tabulation of preference (yes or no) with gender (male or female).
The computer, based on the proportion of men to women in the sample, will calculate what the expected percentages would be if there was no difference in preference versus if gender makes a difference.
It would be simple to compare these numbers using percentages if the groups were all the same size. However, this is unlikely to be true. For this reason, the chi-square test can be used to determine if there is a statistically valid difference in the relationship between age and reason.
Another statistical test is the t-test. With this test, the idea is to compare the outcome of a variable with two different groups. It would be used if a researcher wanted to compare the average number of sporting events attended by male versus female students.
A survey would be conducted to ask how many events were attended and then the mean for females and males would be calculated. It would be unlikely that the average would be exactly the same.
The t-test tells the researcher if the difference is significant to warrant action. ANOVA is a test that is similar but can handle more than two groups. For example, if the study on sporting event attendance wanted to determine the difference between first, second and third-year students.
Another way of examining data statistically is a correlation, which looks at how one variable affects another. Using the example from above, another question that might have been asked is the students’ address. A correlation could be run between attendance and distance from the stadium to see if it affected attendance.
If the question is whether the distance from the stadium and also a student’s academic standing affect attendance, multiple regression can be run. The analysis that will result will show if either, distance or academic standing, have a significant effect on attending sporting events.
1. Quantitative research produces statistical findings that, if the sample is sufficiently large and has been carefully chosen, can be used to support a null hypothesis.
The measurement scales that can be used are nominal, with two states of being, and ordinal, which also shows preference. In addition, interval scales provide a standard unit of measurement, and ratio, which has a given start and end point.
2. The process of qualitative data analysis starts with the pre-analysis of data including review, coding and data entry. The descriptive statistical analysis includes frequency, central tendency, and dispersion. The inferential analysis includes analysis of statistical significance used to test a hypothesis.
3. Frequency analysis provides a count of the frequency of responses. Cross-tabulation shows the number of responses along with at least one other variable. Central tendency analysis of the data includes the mode (or most frequent response), the median response and the mean (or average). Dispersion measurements include range, variance, and standard deviation.
4. Data analysis can also be used to test the hypothesis or guess, made before the research begins. The purpose of the analysis is to try to prove the hypothesis false (and the alternative hypothesis true) with a certain level of confidence.
A z-test is used to determine if there is significance in the difference in the expected hypothesized result and the result and from the surveyed sample. A chi-square test finds the goodness-of-fit in a frequency table between two variables.
Report Writing and Presentation
Marketing researchers do not conduct research just for the sake of ‘knowing’. Research is conducted to find a solution to a problem. Even though the data have been collected and analyzed, marketing researchers’ work is still not done. After all, the analysis that the marketing researchers have completed does not solve management’s problem.
Instead, it is the raw material that provides the insights that researchers will use to make the recommendations that will solve the problem. Once researchers have completed the analysis and developed the recommendations, the next step is to communicate this information in a written and oral format that is both understandable and actionable.
Researcher responsibilities after completion of a research study
Analyze data for relevance to research question Make recommendations for action based on the data
Prepare a written and oral report to communicate the data, analysis, and recommendations
Reasons for preparing a report
Unfortunately, too often researchers do not allocate enough time or importance to this last step of marketing research. Perhaps this is because researchers enjoy the research process more than report preparation and writing.
As a result, researchers might simply prefer to move on to the next research project. However, there are important reasons why a written report is necessary. First, the report gives legitimacy to any recommendations by describing the research methodology.
There is also a need to preserve information for the future. A report also communicates recommendations while providing documentation that can be used to clarify any misunderstandings.
After all, the company that commissioned the research will be paying for recommendations that can be implemented to improve their performance, rather than just facts based on findings.
Today there are other methods used to present research findings than a traditional written report. For example, reports can be prepared in the form of videos. This type of report is more common when conducting research with trendy products and young research participants.
A video can capture ‘attitudes’ that are difficult to communicate in writing. Even reports that are written can benefit from video clips that show either parts of the research project or the actions of consumers. Such clips can help emotionally engage the audience.
Researchers must prepare a thorough report as management may not be able to understand the research methodology or the analysis process without a clear explanation. Terms such as ‘stratified sample’, ‘projective techniques’ or ‘confidence level’ may have no meaning to those whose responsibility it is to make decisions based on the research.
Without an explanation of these terms, the data will either be meaningless or misunderstood. Marketing research is also used in fields other than just consumer marketing. For example, a study done on political issues may well have a reader who is familiar with political strategy but unfamiliar with research methodology.
The second reason for a written report is so that the knowledge that is obtained from the research continues to be available in the future. The research data and recommendations need to be maintained for both the marketing department and management.
All companies have personnel changes and it is particularly common for marketing professionals to change positions frequently.
Even the manager who commissioned the research may be promoted or leave the company. If there is no written report, the new manager will have no access to the knowledge that resulted from the research effort. In this case, duplicate research may be conducted.
The purpose of the report is not only to report data or information but where the recommendations that result from the data and analysis are explained. These recommendations are the result of marketing researchers’ analysis and interpretation of the data. These recommendations should be reported as actionable ideas that management should consider implementing.
A final reason for writing the report is to ensure that the marketing researchers who have conducted the research have a document that details findings and conclusions in case of future misunderstanding.
If management makes a decision that causes a future loss of revenue, it might be easy to state that the decision was based on research conducted by the marketing department. If there is a written document this misrepresentation can be clarified.
Types of the research report
One size does not fit all when preparing research reports, so more than one type of report may be needed. First, preliminary reports may be prepared to reassure those who commissioned the research that the process is on track.
In addition, different groups, such as marketing staff, management, company employees and participants, will have a desire and need for varying levels of detail in the report.
When preparing the various reports required, researchers should consider carefully both the length of the report and how it is to be disseminated.
If the research process spans a considerable length of time, a marketing researcher may prepare a brief preliminary report. A preliminary report can assure management that the research is proceeding as planned. Having read the report, those who originally commissioned the research may decide to make changes in the methodology based on the preliminary results that have been presented.
It is better to find that management wants changes before the process is complete. Of course, a more detailed report will be prepared after the research and analysis have been finished.
Marketing staff report
A report being prepared for the marketing staff of an organization should contain all details of planning and methodology. Marketing researchers will be interested in all of the results and analysis, whereas management will only be interested in the main findings. This document will also be used for reference when planning future research.
Because the management of an organization will generally not have the time or inclination to involve themselves in all the details of the process, their report will provide a quick overview of the research process.
Management will usually want more information on any findings and recommendations. A shorter report that leaves out most of the technical details has the advantage of being less expensive both to write and reproduce.
In addition, researchers should remember that a shorter report is more likely to be read and understood. Because managers are busy with many responsibilities, a long report may be put aside to be read ‘later’ when there is more time.
Unfortunately, this ‘later’ (with ample free time for reading the report) may never arrive. Researchers working together with management from the beginning of the research process will prove advantageous for both.
Research + Management Goals = Success
Marketing research practice now emphasizes partnerships between researchers and management. Today researchers understand that their research must deliver more than numbers.
Research findings must translate into concrete benefits for companies by helping them achieve business goals. If not, management will not commission research again. Ten tips to ensure that the research findings are seen as useful by management are:
1.Align market research with the marketing plan
2.Link market research with business results
3.Link business activities with future business results
4.Integrate the voices of customers and employees into the strategy process
5.Link customer attitudes with customer behavior
6.Engage stakeholders in the research process
7.Define success measures
8.Track those measures
9.Create a process for driving insight
10. Create a process for driving improvements that produce action Question: How else can market researchers convince management that research is useful?
If the research report recommendations involve a change in the product line or targeting a new consumer segment, a company may wish to share the research that resulted in these recommendations with their employees.
If the company wants the support of its employees in making changes, the employees need to understand the research that resulted in the recommendation.
However, company staff may not be interested in all the details of how the research was conducted, so the report needs to be kept short. This type of report can also be posted on internal websites so it is shared with everyone in an organization.
If the research findings are to be made public they can be placed on the organization’s web and social media sites.
When conducting the research, participants may have expressed a wish to learn about the outcome of the research. This is especially true if the research was undertaken for a nonprofit organization. If they are interested, participants should be provided with a short summary of the findings.
After all, it is the participants who made the research possible. However, participants should not be provided with a copy of the recommendations as this is information that belongs to the company or organization commissioning the research.
Components of a Written Report
Reports may vary in writing style, but all reports should contain an introduction, a section on methodology, any findings and recommendations and appendices. The introduction will identify those involved in the research and provide a brief overview of the contents. The next section of the report will explain the research question, objectives and methodology.
The third section would discuss the research findings and recommendations. The appendices will provide background information that will more fully explain the report’s contents.
The introductory material contained in a report will include a title page, a letter of transmittal, a table of contents and an executive summary.
The report will start with a title page that provides the name of the research study, the date of the report submittal, the names of the researchers, the names of the people or organization that commissioned the research and contact information for both groups.
A letter or memo of transmittal will be included for marketing research studies that were conducted by an outside firm.
The letter of transmittal formally concludes the research study and transmits ownership of the information contained in the report to the management of the company.
The letter will also describe the legal ownership of any supporting material, such as tapes or completed survey forms. Because of the sensitive nature of some research, the commissioning organization may wish to have this information destroyed and in this case, the letter will also state that this has been done.
The report should next contain a table of contents and should be indexed in some way so that each section can be easily found. This table of contents and the indexing are useful when the report is discussed in meetings, as each section can be easily and quickly located.
A table of contents should not only list the main sections, such as the introduction and findings, it should also list the pages for any subtopics under the main topics.
A table of contents should also provide a separate listing for any graphs and tables. The more detail that is provided in a table of contents, the more easily material can be found.
An executive summary quickly states the research question, the methodology, findings, and conclusions. As most people in positions of management are under time pressure, the executive summary is essential to communicate quickly what a report contains.
In addition, if the executive summary does not communicate effectively that the research methodology, findings, and recommendations are important, the report may never be read. While included early on in the report, the executive summary is actually the last part that is written.
The body of a report will include information on the research question, the research objectives, and the research methodology. First, the problem that resulted in the research being conducted should be explained and the main and any secondary research questions should be stated.
Since the findings and recommendations should directly address the research question, it should be highlighted from the body of the report by using bolding or italics so it can be easily referred to later.
The research objectives should also be explained. If they are not included, anyone reading the report may wonder why certain aspects of the research findings are not discussed further in the recommendations.
This section of the report should also describe the sample selection process. It should clearly explain how it was determined who should participate in the study and how these specific individuals were chosen from the total population.
The legitimacy of the findings of a quantitative study depends upon whether the appropriate people were asked to participate. Even for qualitative research, if the wrong people are asked to participate in the research, the information will not be relevant.
Therefore, to give legitimacy to the research findings and recommendations, it is important to describe the profile of participants and how they were selected.
Finally, the body of a report will explain the research methodology. For a survey, this will include how the questionnaire was developed and tested. The report will also inform the reader of how many surveys were conducted, the method of contact and the dates of the research.
For a focus group, similar information will be included but the name of the moderator will also be given. Readers may be less familiar with research techniques such as observation, projective techniques, and ethnography. If less well-known research techniques have been used, the report will need to explain the methodology in more detail.
Findings and recommendations
The next section should discuss the research findings and recommendations. This section should include a summarization of the research findings.
It should not provide all the data that were compiled during the research process, instead, it should summarize the data that were used to develop the recommendations. Research always provides a wealth of information. However, providing all the detail at this point in the report will only confuse the reader.
For example, a survey might have been conducted to discover what activities consumers want to experience on a cruise ship. The research will have also asked survey respondents demographic information.
During the analysis stage, researchers discovered that the geographic location of consumers’ homes made no difference to preference.
This fact will be stated. However, there is now no need to present a detailed breakdown of the home location of each participant in the body of the report. These data can be added in the appendix.
On the other hand, if age was a very important variable that affected what activities people wanted to experience, detailed information on age and activity preference should be included in the body of the report.
Recommendations are the most important component of a research report. After all, researchers are not paid just to collect data. They are also paid to analyze and interpret the data.
The recommendations should directly address the research question and the research objectives. The goal of all research is that it proves useful and it doesn’t just sit on someone’s desk.
It’s All about the Implementation and it: Starts with the Report The marketing researchers should start thinking about the implementation even while they are preparing the report and presentation.
While the researchers have conducted the research, and have a deep understanding of what the findings mean, this is not true for the report’s audience. A way must be found for other people to understand the findings and internalize what they mean for the organization. Here are some ways to do so:
Always Start with the Why: Explain why the research was conducted and why the findings matter. If you don’t there is a good chance that the report will end up unread.
Provide a Story: Don’t only talk about data in the abstract. Use real examples. If the data were based on only statistical analysis, give hypothetical examples of typical customers to make the story real to the listener.
Use Diverse Media: Not everyone wants the same amount of information so provide long and short forms of the report. In addition, the report can be provided verbally, with an online video or be using infographics.
Crowdsource Implementation Ideas: Rather than give your opinion on what actions should be taken, give opportunities for everyone in the organization to provide their ideas. This can easily be done using social media.
Question: What social media could be used in an organization to get ideas for actions to be taken based on the research findings? Source: Jolls, 2015
The final section of the research report will provide the full data that were obtained during the research process. It will also include information that provides further details on the research sample and methodology.
For example, details on the research sampling method can be explained. If referrals were used, the organization and individuals contacted can be listed. Examples of research methodology such as survey forms, projective techniques or focus group scripts can also be included.
Writing a Professional Report
If the research report is going to be read, instead of just sitting on someone’s desk, it needs to be readable, interesting and concise. A report that contains too much jargon is poorly organized and is visually unattractive will not be read.
This is especially the case because most people have multiple tasks to accomplish each day and will complete the easiest task first.
Preparing a readable report
Issues that researchers should consider before writing a report must include the writing style. Before beginning to write a report, researchers should determine for whom the report is being prepared.
If a report is being written for someone in the marketing department, a more professional style will be used. Here it can be assumed that the reader will be familiar with research terminology.
However, if the report is being written for someone who runs small business researchers must write in a more colloquial style while being careful to explain every term.
Making the report interesting
Any report should also be interesting to read. Researchers should not just state facts but also give examples of interesting incidents that occurred during the research process. This type of detail will help to bring the information to life in the mind of the reader.
Another way to add interest to a report is to use photos of the participants as they were involved in the research process. If this is not possible, actual quotes can be used to give the readers a feeling that they were at the research sessions.
A report should be kept as short as possible while still including all the necessary information. It should also be arranged so that readers can process the amount of information they wish without necessarily reading the entire report.
This can be accomplished by providing a well-written executive summary and clearly labeling each section. In addition, headings and subheadings should be used throughout the report so that a specific issue can be found quickly.
Using visual material
The research findings section will contain a wealth of detailed numerical data. A report writer should use visuals to assist readers in understanding the relationship between different sets of data.
For example, readers can quickly visualize the relationship if 46 percent of the respondents were female while 54 percent were male.
However, if part of the research question addressed the age of the survey respondents these demographic data can be quite detailed. For example, if 12 percent of the respondents who used the product were aged 18– 22, 16 percent were aged 23–29, 26 percent were aged 30–39, 23 percent were aged 40–49, 14 percent were aged 50–59 and 9 percent were 60 or older.
It is extremely difficult for readers to visualize quickly and see the relationship between these numbers. To assist readers, such detailed data can be presented in the form of tables, charts or graphs.
A table simply lists numbers in rows by categories. Individual numbers will still be discussed in the body of the text. In the example given above, the writer might include in the body of the report that the largest age group of product users was 30–39, at 26 percent, with 23 percent of the users aged 40–49.
The writer might then add that the smallest group was aged 60 or older, at 9 percent. This comparison thus makes the point that most users are middle-aged. The remaining numbers, while still relevant in showing the distribution, should not be discussed in the body of the report as it makes it too difficult to read.
A pie chart is a visual method to present raw numbers and their relationship as percentages of a whole. Pie charts are useful when representing numbers that make up a whole, such as the percentage of customers in different age groups.
The different sections of a pie chart should be shown in various colors or in shades of gray for each group. Using color to differentiate the sections makes a pie chart easier to read and use.
Bar charts represent the values of different items so that these can be easily compared. The difference from the pie charts is that bar charts do not necessarily show items that are part of a known whole.
However, a bar chart will be used to show the relative size of the raw numbers of people in each age group who are product users.
A bar chart can be shown with the bars vertical or horizontal. Bar charts can also be used to show changes over time by combining two sets of numbers collected at different dates. Each set of data can be shown next to each other on the chart.
A line chart is designed to show changes in data over periods of time. Time is shown on the horizontal axis, while the numerical measurement is shown on the vertical axis. A line chart can show the percentage change in a product’s sales by any measurement of time, such as by the day, month, season or year.
The advantage of a line chart is that it can easily display movement in value over time for more than one variable. For example, a company may collect sales data for five different products all of which could be shown on the same line chart. This could then reveal if there is any relation between changes in sales figures.
People find it easier and faster to process visual information than text. Infographics are a way to take almost any data and present it in a pleasing visual format. They present both the numbers and an explanation of what the numbers mean.
They are not a substitute for a written report, but rather a way to get the interest of the reader who will then read on for more detail. There are many online sites that can be used to generate infographics. They can be used in both the written report and oral presentation.
An Oral Presentation
An oral presentation is an opportunity for researchers both to explain and to ‘sell’ to management any research findings and recommendations.
An oral presentation also allows researchers to provide a more effective description of the research methodology.
This is because during an oral presentation a presenter, by observing their audience, can become aware of when they are encountering difficulties and can explain any confusing details more fully.
In addition, a presenter can more clearly explain how report recommendations are related to any research findings. Finally, an oral presentation provides a means to clarify any misunderstandings about the research process.
Reasons for oral presentation
A clearer explanation of the methodology
An opportunity to explain the tie between research findings and recommendations
An opportunity to clarify misunderstandings
Presentations, just like written reports, must have a structure as it helps the audience anticipate, and concentrate on, information that is of particular interest. A structure also helps a presenter stay on topic.
No one would throw together a written report at the last minute as its poor organization would leave readers confused. It would be quite obvious to them that researchers had not taken time to present the material in a logical manner.
However, some people do throw presentations together at the last moment and believe that their audience will not notice. Unfortunately, a poorly prepared oral presentation will leave listeners just as confused as a poorly prepared written report.
A well-prepared presentation will have four major components: an introduction, methodology, findings and recommendations, and a conclusion.
It is important when planning a presentation that most time is devoted to the recommendations section. Please note how the time has been divided to ensure that all topics will be covered before the conclusion of the presentation. Not all sections are given equal time as not all have equal importance.
The presentation’s introduction should identify the researchers and the commissioning company and also explain how long the presentation will take. During the introduction, the presenter should inform the audience whether questions may be asked during the presentation or if the audience should keep their questions until the end.
The introduction should then very quickly state the research question and describe what information the presentation will contain. For example, in a 30-minute presentation, the introduction should only last about three minutes.
The section of the presentation on methodology is where the presenter will first state the research question and objectives. The presenter will then briefly inform the audience of how the research participants were selected. The presentation should not be used to describe the technical details of the sampling procedure.
If the audience is interested, the report will contain all the necessary information. The purpose of describing the sampling procedure during the presentation is simply to give credibility to the findings and recommendations.
The same holds true for the methodology, although more time should be spent on this topic so the audience will better understand how the findings were obtained.
At this point in the presentation, a sample survey form can be distributed, projective techniques can be displayed, or a short video clip from a focus group can be shown.
Findings and recommendations
The presenter should spend more time presenting the findings and recommendations. They can use visuals such as graphs to quickly show to the audience what has been learned from the research. The presentation should never try to explain all the findings, as there simply isn’t the time and the audience will get lost in the details.
In addition, it is the presenter’s responsibility to sift through all the findings to determine what is relevant for answering the research question. However, any findings that have an impact on the recommendations should be presented.
When Can Research be Considered Effective?
The answer to this question is clear to those market researchers who attended a seminar sponsored by Research Solutions, based in New Zealand. In their opinion, marketing researchers must ‘come out from behind their pie charts’ and become partners with the organization commissioning the research.
What does such a partnership mean? A true partnership was defined as exhibiting four factors:
1. The partnership must be based on trust. Of course, a good market researcher will produce accurate findings but management must also trust the recommendations made by researchers.
2. Management must be willing to risk making changes based on research findings. This risk-taking approach must start at the research proposal stage. If the research proposal is written to allow only a conservative research approach, no breakthrough findings can possibly result.
3. Researchers must be in contact not just with management and marketing but also with other departments throughout a company. This must be the case because the recommendations that result from the research will impact more than just the marketing department.
If these departments do not trust the marketing researchers, they are more likely to argue against taking action based on their recommendations.
4. Management must accept that research findings are based on consumers’ viewpoints, which can be surprising and sometimes uncomfortable for management to hear. It is not the responsibility of researchers to only bring good news.
Question: Which of these issues do you believe is the most critical in developing a partnership?
The conclusion to a presentation should be brief. The presenter should restate the research questions and the main recommendations. They should also thank the audience for their attention.
The presenter should allow adequate time to answer any questions. The audience should also be informed of who it is they can contact if they have any questions in the future.
A presentation has a different purpose than that of a written report. Giving a good presentation is a skill that can be learned. Everyone understands that being able to produce a clear, concise and interesting written report takes time and effort;
however, too often presentations are afterthoughts that people expect will happen automatically once they are in front of an audience. After all, while not everyone is skilled in writing, everyone can speak. Yet nothing could be further from the truth.
The general rules for an effective presentation are to be interesting, organized and brief. A presentation should never be thought of as simply an oral presentation of all the information in a written report. The purpose of a presentation is not just to communicate information.
After all, the audience at a presentation can read the written report for themselves. The purpose of a presentation is to ‘sell’ the ideas contained in the report by persuading an audience to act upon its recommendations. If an oral presentation is successful, the audience should be eager to read the written report for more details.
In fact, if possible a presentation should be practiced in front of a similar audience in the same room that will be used later. A practice presentation will mean that when the actual presentation is given, the presenter will be relaxed and able to concentrate on the audience instead of the presentation.
Unforgivable sins made during presentations
A successful presentation depends on preparation. In addition, a good presentation must be interesting. If it is not interesting, the research recommendations may be ignored because the audience simply lost attention and they weren’t heard. Everyone has probably had the experience of having to sit through a poorly prepared presentation.
There are actions that are guaranteed to result in a poor quality presentation. A presenter should never read anything longer than a short quotation. After all, the audience came to hear an oral presentation and they could have stayed in their offices to read the report on their own.
In addition, audiences should never be frustrated by being shown any PowerPoint slides or other visuals that can’t be easily read. They should also never be bored – life is difficult enough without struggling to stay awake during a presentation.
A presenter should never be so rude as to ignore the audience, nor should they overwhelm an audience with too much detail. It should be remembered that humans can only assimilate so much information at a time.
Finally, a presenter should care about the information that is being presented. After all, if the presenter does not care, why should the audience?
Unforgivable sins of oral presentations
Reading out the report
Presenting PowerPoint slides that are unreadable Boring the audience
Not interacting with the audience
Using too much-supporting material
Not being emotionally involved with the presentation’s contents
Using visuals during a presentation
Because of today’s multimedia environment, people have become accustomed to receiving information in more than one form at the same time. People seem to have no problem understanding and using these multiple sources of information even while performing other tasks.
So it is not surprising that people have a difficult time simply listening to a verbal presentation with no other visual interaction. Even during the most interesting oral presentation, the minds of the audience might wonder as to what they need to pick up at the store on their way home from work or order online as soon as they get the chance.
Therefore, using visual material not only helps to communicate the research information it also helps an audience stay focused on a presentation. Visuals used during a presentation may be computer generated. However, low-tech methods, such as using a whiteboard, flipcharts, handouts, and photographs, can be just as effective.
PowerPoint is probably the best-known method of presenting visual information during oral presentations. However low-tech methods can also be used, including writing on a whiteboard, using flip charts and distributing handouts with relevant information and photographs of products or research participants.
They can be effective because the low-tech approach can seem more personal. Whiteboards and flipcharts can be used to draw attention to important facts.
If a surprising 78 percent of consumers surveyed were unaware of a company’s new promotional campaign, this number can be written out in large print in red on the board or paper. The number becomes a clearly drawn exclamation point for this fact.
Even if a presenter chooses to use PowerPoint during a presentation, it is still useful to use a low-tech method such as handouts with a summary or pertinent information.
This summary, which should only list the main findings and recommendations, serves two purposes. It can be used to reinforce what a presenter is explaining and it can also provide a place for an audience to write notes as they are listening.
Handouts can also be used that contain other material, such as maps of the study area or copies of the survey form or projective techniques.
Photographs of the product under discussion can also assist an audience in better understanding the relevance of data. These photographs can be used as displays around the room where a presentation is taking place.
High-tech presentation aids include the projection of an online source, videos, and PowerPoint. Using a PC and a projector, presenters can bring up on screen material that is online. For example, if the research involved the perception of a company’s image the home page of that company’s website can be shown.
Video clips of participants, perhaps in focus groups or as part of the observational research, can also be used during a presentation. The use of videos is becoming increasingly common when making research presentations as they will bring the material to life in ways that other material cannot.
Sometimes research studies have relied on videos to record events or behavior that are then analyzed. Of course, these studies should include some of the video clips in the presentation.
1. Marketing research is conducted to find answers to questions which will lead to recommendations for future action. This information is usually communicated through a written report and oral presentation.
Reasons for the written report include the need to explain the research methodology, to preserve the knowledge for future employees, to document the actionable ideas that resulted from the research and to provide clarification in case of future misunderstandings.
A researcher may wish to prepare reports of different lengths for other researchers, management, staff, and participants.
2. A research report will consist of an introduction that will identify those involved in the research and provide a brief overview of the contents. The methodology section of a report will explain the research question, the objectives, and the methodology.
The third section, which is the most important part of the report, will discuss the research findings and recommendations. The appendices will provide background information that will more fully explain the report’s contents.
3. A professional report is readable, interesting and concise. If this is not the case, the report will not be read and, therefore, the recommendations will not be implemented. Numerical data should be communicated using tables and charts. These will help readers visualize the relationship between groups of data.
4. An oral presentation is needed to explain the methodology, the relationship between the findings and the recommendation, and to provide an opportunity to clarify misunderstandings.
The presentation will include an introduction, methodology and findings, recommendations and a conclusion. A presentation should be interesting, organized and brief. Both low-tech and high-tech visuals should be used to better communicate information.