Overview of Survey Research Methods o Largescale research

  • Slides: 24
Download presentation
Overview of Survey Research Methods o Large-scale research design procedures for collecting large amounts

Overview of Survey Research Methods o Large-scale research design procedures for collecting large amounts of raw data using questionnaires.

Conceptualization and Constructs o A concept is an abstract idea to explain how things

Conceptualization and Constructs o A concept is an abstract idea to explain how things work. Closely related, a construct is more abstract and developed for scientific inquiry

Operationalizing A Construct When a research team operationalizes a construc, they go through a

Operationalizing A Construct When a research team operationalizes a construc, they go through a process of “breaking-it-down. ” In other words, they create questions to reflect a construct. How would you operationalize … Ø Customer Loyalty 12 -5

Various Variables o Independent variables o Dependent variables o Extraneous (uncontrolled) variables n Lead

Various Variables o Independent variables o Dependent variables o Extraneous (uncontrolled) variables n Lead to spurious conclusions

Guidelines for Question Quality (to control for non-sampling error) o Survey questions must tap

Guidelines for Question Quality (to control for non-sampling error) o Survey questions must tap into the research problem • Use simple, interesting, non-threatening opening questions • Do your respondents understand the question? . • Use the funnel approach, asking broad questions first • • Carefully design branching questions Place difficult or sensitive questions near the end Ask for classification information last Pre-test the survey

A Funnel Sequence How would you rate your last experience when eating at ABC

A Funnel Sequence How would you rate your last experience when eating at ABC restaurant? o Extremely o Poor So-So Good Extremely Good How would you rate the quality of the service? o Extremely o Poor So-So Good Extremely Good

Leading Questions A question framed to give the respondent an indication about how he

Leading Questions A question framed to give the respondent an indication about how he or she should answer. Example Do you feel that limiting taxes by law is an effective way to stop the government from picking your pocket every payday? Yes No Undecided

Double-Barreled Question A question that calls for two responses Example Are the food and

Double-Barreled Question A question that calls for two responses Example Are the food and service at the local Green Mill good? Yes No

Branching Question Example 1. Did you give money to Canopy of Care last year?

Branching Question Example 1. Did you give money to Canopy of Care last year? ___ Yes ___ No {If Yes, go to #3. } 2. 8. 9. We are interested in why people do not contribute. The following is a list of answers others have given. Please tell me which, if any, apply to you. a. ___ Someone else in my household had already contributed. b. ___ I did not have the money at the time. c. ___ I gave to other charities. d. ___ I volunteered my services to Canopy of Care instead of contributing money. e. ___ I volunteered my services to other charities instead of contributing to Canopy f. ___ I did not give because Canopy of Care spends its money inefficiently. g. ___ None of the above. 3. 4. 5. 6. To how many different charities do you think Canopy of Care contributes to? a. ___ 0 -20 d. 81 -100 b. ___ 21 -40 e. More than 100 c. ___ 41 -80 f. Don’t know. 3. 4. 5. 6. 7.

Creating the instrument

Creating the instrument

Open-Ended Questions o the respondent can reply in his or her own words. “What

Open-Ended Questions o the respondent can reply in his or her own words. “What advantages, if any, do you think ordering from a mail order catalog company offers compared with local retail outlets? ______________________________________

Closed-Ended Questions o Closed-Ended Questions n n response is “forced” does avoid problems (such

Closed-Ended Questions o Closed-Ended Questions n n response is “forced” does avoid problems (such as subjectivity) of open-ended questions. But you can lose compelling data.

Closed Questions o RESPONSES MUST BE. . . o Exhaustive o Mutually exclusive o

Closed Questions o RESPONSES MUST BE. . . o Exhaustive o Mutually exclusive o Should have a “Don’t know”

Types of Scale o Nominal Scale n labels and categories (e. g. gender, major)

Types of Scale o Nominal Scale n labels and categories (e. g. gender, major) o Ordinal Scale n order (ie. , rank restaurants in order of preference) o Interval Scale (it is inherently ordinal) n equal differences (e. g. temperature) o Ratio Scale n Natural zero (e. g. weight, height)

Scale Characteristics o Order – numbers on scale are ordered o Distance – differences

Scale Characteristics o Order – numbers on scale are ordered o Distance – differences exist between the ordered numbers o Origin – There is a unique origin indicated by the number zero

o Dichotomous Questions n n The simplest form of a closed-ended question. Nominal response

o Dichotomous Questions n n The simplest form of a closed-ended question. Nominal response 1. Where do you live? A. Large urban area B. Small rural area 1 2

Nominal Scale Which of the following soft drinks do you like? Check all that

Nominal Scale Which of the following soft drinks do you like? Check all that apply. � Coca-Cola � Mountain Dew � Seven Up � Dr. Pepper � Pepsi � Sprite

Ordinal Scale 1. Now that you have read a sample of the Wall Street

Ordinal Scale 1. Now that you have read a sample of the Wall Street Journal, would you say that you would…(Check one) Definitely buy it Probably buy it Might or might not buy it Probably will not buy it Definitely will not buy it

Interval Scale What is your overall opinion about each of these brands? Coca-Cola Dr.

Interval Scale What is your overall opinion about each of these brands? Coca-Cola Dr. Pepper Pepsi Sprite unfavorable 1 2 3 4 4 5 5 favorable 6 7 6 7

Likert Scale o How much do you agree with the following statements: SD SA

Likert Scale o How much do you agree with the following statements: SD SA o People who live in BH are rich 1 2 3 4 5 o People who live in BH are sad 1 2 3 4 5 o

Semantic Differential o People who live in Beverly Hills are: o rich o friendly

Semantic Differential o People who live in Beverly Hills are: o rich o friendly o young o happy 1 1 2 2 3 3 4 4 5 5 poor snobby old sad

Ratio Scale (numeric response) Divide 100 points among these soft drinks according to your

Ratio Scale (numeric response) Divide 100 points among these soft drinks according to your likelihood of purchasing each within the next week: ___Coca-Cola ___Mountain Dew ___Seven Up ___Dr. Pepper ___Pepsi ___Sprite Age ___

Hierarchy of Measurement Scales Ratio – the “elite” scale. Possesses order, distance, origin (absolute

Hierarchy of Measurement Scales Ratio – the “elite” scale. Possesses order, distance, origin (absolute zero) Data Analysis? ______ Interval – possess order and distance, but not origin (zero point is arbitrary. Data Analysis? ______ Ordinal – ranking scales. Possess only the characteristics of order. Data Analysis? ______ Nominal – names, labels, tags. Does not possess order, distance, origin Data Analysis? ______

Scale Validity and Reliability o Validity: Does the measurement “mean” what you wanted it

Scale Validity and Reliability o Validity: Does the measurement “mean” what you wanted it to mean? n Construct validity p p n Convergent Discriminant Predictive validity o Reliability: Has to do with how accurate the measure is