RESEARCH BRAINTEASER
You are conducting a research study on consumer attitudes towards a local home improvement store. You plan on creating a survey and you know, from your research classes, that measurement is a vital part of an effective survey in order to achieve reliable and valid results. You realize that measuring "attitudes" is hard as attitudes can't be directly observed. So, how do you measure consumer feelings?
QUESTION
What is "measurement"?
What is "measurement"?
A good way to observe consumer attitudes with the use of measurement would be to devise a scaled response where the varied opinions would be assigned a number. this would give the respondent several options so the researcher could pin point their true feelings most accurately.
ReplyDeleteYou're 1/2 there in your answer.
ReplyDelete"Measurement" is the process of assigning #s that reflect the amount of attribute possessed by persons, objects or events. Only the attribute is measured, such as attitude, age, income, etc.
What is a “scale”? How does it differ from measurement?
ReplyDeleteScale is providing a variety of varied answers such as poor, very poor, great ect.... measurement is the process of assigning numbers and labels to persons
ReplyDeleteNot quite, although you're 1/2 there again!
ReplyDelete"Scale" is similar to measurement in that you are assigning #s to attributes of a person, object or event in order to give it some numerical characteristic.
There are four types of "scales", in which the researcher (as you stated) can get a variety of answers from respondents. What are these four types of scales?
I do not recall the four types of scales
ReplyDeleteYou may want to reread Chapter 8. The four types of "scales" are: nominal, ordinal, interval and ratio.
ReplyDeleteWhich of the four types of measurement scales are best suited to measure “attitude”?
i'm here!
ReplyDeleteFrankie - you're late. Please answer the most recent question: Which of the four types of measurement scales are best suited to measure “attitude”?
ReplyDeletemaybe interval scales?
ReplyDeleteI would have to say multidimensional scale would be the best to find attitudes. this type of scaling is designed to measure several dimensions of concept, respondent or object.
ReplyDeletei chose this option because it measures a consumers trait over another consumer
ReplyDeleteWhy an interval scale?
ReplyDeleteActually, the answer is the two best types of scales in which to measure attitude are: nominal and ordinal.
ReplyDeleteBe sure to review Chapter 8 for Thursday's class!
wait... does interval scales only measure temperature like the example in the book?
ReplyDeleteOkay, on to new material. This week, we will talk about data processing.
ReplyDeleteWhen I say “data processing” or “data analysis”, what am I referring to?
sorry, i was on the wrong page
ReplyDeletethe process of analyzing gathered data
ReplyDeleteFrankie - to answer your question re: interval scale. No, it is not used to solely measure temperature. An interval scale assumes that the points on a scale are equal, i.e. a thermometer in which all points are equal and you have an absolute zero.
ReplyDeleteResearchers use interval scale to measure how much of a trait one consumer has (or does not have) over another.
obtaining data and then doing an in depth study of said data
ReplyDeleteExactly, Correne. Once we’ve designed our survey and administered the survey to respondents, now we have to analyze the results.
ReplyDeleteThere are five steps involved in this analysis process. They are:
- Step 1: Validation and editing
- Step 2: Coding
- Step 3: Data entry
- Step 4: Machine cleaning of data
- Step 5: Tabulation and analysis
Let’s go through these five steps now. Let’s begin with Step #1 – validation and editing.
ReplyDeleteWhat is “validation” and “editing”?
validating that the information youve obtained is correct
ReplyDeletevalidation is finding if the researcher was trying to measure what was actually measured.
ReplyDeleteSort of... "Validation" of data results means asking yourself if the interviews conducted as specified, i.e. did the interviewer follow instructions. Can you detect interviewer failures to follow instructions (i.e. cheating)?
ReplyDeleteediting is going through and correcting/finding the data that might have been misrepresented.
ReplyDeleteFrankie - that's the definition of a "valid" measurement. Good memory, though.
ReplyDeleteYes, "editing" is asking yourself if the surveys filled out properly and completely?
ReplyDeleteEditing also involves checking for interviewer and respondent mistakes (i.e. interviewer failed to ask question or record answer, were skip patterns followed, did the interviewer paraphrase the respondent’s open-ended answer or record each word?)
On page 395, Exh. 12.2, there is a great example for recording of open-ended questions.
Do we remember what a "skip pattern" is?
ReplyDeleteAny ideas?
ReplyDeleteskip pattern is when you have a selected number you want to skip like every other person to answer your survey
ReplyDeleteYes, that is correct Frankie. What you are describing is a skip pattern as it pertains to your selecting your sample, i.e. you'll survey every 10th person.
ReplyDeleteWith regards to skip patterns within an actual survey... If the respondent answers "yes" to Question #4, then they move to Question #6, for example.
So, should you have designed your survey with such skip patterns, during the editing process you want to ensure these skip patterns were followed by the interviewer. If they weren't, then the survey is not as reliable and valid.
ReplyDeleteLet's move to Step #2 – coding.
ReplyDeleteWhat is “coding”?
Coding- the process of grouping and assigning numeric codes to the various responses to a particular question.
ReplyDeleteExactly, Correne. So, why is "coding" important when analyzing our survey results?
ReplyDeletecoding is more or less labeling groups with numbers and codes to particular questions to identify them individually
ReplyDeleteRight, Frankie. So, why is "coding" important when analyzing our survey results?
ReplyDeletethis will allow us to determine what percentage of our respondents felt a certain way about a particular topic.
ReplyDeleteyou need to know how the consumer answered the questions because most of the questions are closed-ended
ReplyDeleteCorrene, you hit the nail on the head. "Coding" allows us to assign a # to each answer, regardless of open- or close-ended, in order to get a numerical value for how ALL respondents who took our survey felt.
ReplyDeleteThere are two different ways to code answers, depending on whether it's a close- or open-ended question.
Closed-Ended: We pre-code the answers respondents can select from, i.e. answer “0-2” = code 1, “3-5” = code 2, etc.
Open-Ended:
1. Researcher prepares list of possible answers, SEE PAGE 397, Exh. 12.3
2. Group similar answers into categories and assign code, SEE PAGE 397, Exh. 12.4
3. Enter codes into data processing, i.e. Excel spreadsheet
Any questions on coding open- versus close-ended questions? I would study page 397, esp. if using open-ended questions in your own personal research project.
ReplyDeleteOkay then, moving on to Step #3 – data entry.
ReplyDeleteWhat is “data entry”?
data entry- the process of converting information into a into a form that can be read by the computer.
ReplyDeleteExactly. "Data processing" is converting information from surveys to computer program, i.e. Microsoft Excel.
ReplyDeletechanging information that is gathered and putting it into computer lingo
ReplyDeleteStep #4 – machine cleaning of data.
ReplyDeleteWhile this definition sounds more high tech than it really is, what is “machine cleaning of data”?
Machine cleaning- a final computerized error check of the data. this allows the researchers to confirm again that the data they've entered in the computer was processed correctly.
ReplyDeletechecking for errors in the computers data. its a routine check to make sure everything is a-ok
ReplyDeleteYou got it. It's a fancy way of saying double-check your # before moving on to Step #5 - tabulation of results.
ReplyDeleteThere are two manners in which we can tabulate results.
The first is with a One-Way Frequency Table: the # of respondents choosing each answer to a survey question; also indicates the total % of answers to a given question, SEE PAGE 403, Exh. 12.8.
The second is with a Cross Tabulations: responses to one question relative to the responses to 1+ questions, SEE PAGE 406. Exh. 12.10 & 12.11.
We'll review these two types of tabulation in class on Thursday when we're all together. In the meantime, look at pages 403 and 406 to familize yourselves.
ReplyDeleteBe sure to bring your textbooks to class!
Any questions on the materials covered today?
ReplyDeleteNot so far
ReplyDeletei'll bring a couple in class
ReplyDeleteLet's move to Ethics then...
ReplyDeleteHOMEWORK REMINDER
- Complete Chapter 7 chapter questions 5, 10 & 11 due Thursday, April 23rd
- Review Chapter 8 for Thursday, April 23rd
- Answer chapter 8 questions #2, 13, 14 & 17 for Thursday, April 23rd
- Read Chapter 12 for Thursday, April 23rd
- Mystery Shopping assignment with oral presentation due Thursday, April 23rd