Exam 2 MKTG 352 - Gill
Construct
- An abstract idea or concept formed in a person's mind - Hard to measure - Indirectly observed variable - Need more than one question to measure constructs
Non probability sampling techniques
- Convenience - Judgemental - Snow ball - Quota
Variable
- Directly observed - Demographics
Ordinal level of measurement
- Express relative magnitude or rank-order - Implies more or less, but not how much more or less
Ways to reduce sampling errors
- Increasing sample size - Using the appropriate sampling method
Nominal level of measurement
- Labels only, no arithmetic properties - Represents mutually exclusive categories
Interval level of measurement
- Like ordinal, BUT numbers in this measure are meaningful and represent equal increments - Zero has no meaning
Types of itemized rating scales
- Likert scale - Semantic differential scale - Stapel scale
Four levels of measurement
- Nominal - Ordinal - Interval - Ratio
Types of questions
- Non structured (open-ended) - Structured (close-ended)
Types of non-sampling errors
- Nonresponse error - Response error -sampling frame -measurement -coding errors
Types of comparative scales
- Rank order - Constant sum
Ratio level of measurement
- Ratio measure is an interval measure with a fixed zero - Zero is meaningful
Determining sample size when using probability samples
- Sample size important - Researches want a big enough sample to be confident that the results are an accurate reflection of the population
Determining sample size when using non-probability samples
- Sample size less important - Researches know that they cannot generalize to population
Types of errors
- Sampling - Non sampling
Probability sampling techniques (4)
- Simple random - Systematic - Stratified - Cluster
Response error
- When respondents have impaired memory and do not respond accurately - Social desirability bias
Sampling errors
-Bias attributable to mistakes in either drawing a sample or determining the sample size - The difference between the findings based on the sample and the true values for a population - Caused by the method of sampling used and the size of the sample
Nonresponse error
-Final sample not representitive of planned sample) - A systematic bias that occurs when the final sample differs from the planned sample - Occurs when a sufficient number of respondents in a pre-selected sample do not participate
Steps for stratified random sampling
1. Divide the target population into homogenous subgroups or strata 2. Draw random samples from each stratum 3. Combine the samples from each stratum into a single sample of the target population
Steps for systematic random sampling
1. Obtain an acceptable frame of the target population 2. Determine the number of units in the list and the desired sample size 3. Compute the skip interval 4. Determine a random start point 5. Beginning at start point, select units by choosing each unit corresponding to the skip interval
Factors that influence questionnaire quality
1. Type of questions 2. Relevance and wording of questions 3. Sequencing of questions 4. Overall appearance of questionnaire
Summary of steps for hypotheses testing
1. state null and alt 2. compute t value 3. figure out critical value of t , given the degrees of freedom and level of significance 4. Make statistical decision 5. Draw managerial conclusion
Semantic differential scale
A 7 point rating scale with end points associated with bipolar labels that have semantic meaning
Stapel scale
A unipolar rating scale with 10 categories numbered from -5 to +5, without a neutral point (zero). Presented vertically
Identify the scale: Allocate 100 points to the following three insurance providers. After you are finished allocation points, please make sure that your total adds up to 100. National _____ Geico _____ State Farm _____
Constant sum scale
The true value of a construct is 6. Now we measure this contract 10 times, and the measured scores are 8, 8.5, 9, 8.8, 8, 8.2, 9, 8.6, 9, 8.4. Which of the following is the most appropriate description of the measurements? A. High validity but low reliability B. High validity and high reliability C. Low validity and low reliability D. Low validity and high reliability
D. Low validity and high reliability
Crit value > Calc Value
Do not reject the null
P Value > Level of sig.
Do not reject the null
Univariate
Frequency distribution of one variable (ex. Male of female)
Bivariate
Frequency distribution of two variables (ex. Male/Female and Private Label/Branded)
Identify the measurement level: "I am highly satisfied with the service that my personal banker provides me." Strong disagree 1 2 3 4 5 6 7 Strongly agree
Interval
Identify the scale: Sears has poor in-store service. 1- Strongly agree 2- Agree 3- Neither agree nor disagree 4- Disagree 5- Strongly disagree
Likert scale
Identify the scale: Sears sells high quality merchandise. 1- Strongly Agree 2- Agree 3- Neither agree nor disagree 4- Disagree 5- Strongly disagree
Likert scale
Measure of central tendency
Mean, median, mode
Identify the measurement level: Please indicate gender below. 1. __________ Male 2.__________ Female
Nominal
Identify the type of question: How old are you? _________
Non structured/ratio
Identify the measurement level: What is your age? ____ younger than 20 ____ 20-39 ____ 40-59 ____ 60-79 ____ 80-99
Ordinal
Calc Value > Crit Value
REJECT THE NULL
P Value < Level of sig.
REJECT THE NULL
Measure of dispersion
Range, Standard deviation (for quantitative data)
Identify the scale: Please rank order the following insurance per your preference. National ______ Geigo ______ State Farm _____
Rank order scale
Identify the measurement level: Rank these soft drinks according to which ones you like the best (1= most preferred, 3 = least preferred) Pepsi _____ Coke _____ Sprite _____
Rank order scale/ Ordinal
Identify the measurement level: "How many times during a typical month do you travel by air? ____
Ratio
Identify the measurement level: What is your age? ______
Ratio
Likert scale
Requires the respondents to indicate a degree of agreement or disagreement with each of a series of statements about an object
Continuous rating scale
Respondents rate the objects by placing a mark at the appropriate position on a line that runs from one extreme of the criterion variable to the other -(Interval)
Sampling Frame Error
Sampling frame not representative of ideal population - can be caused by sampling frame that contains irrelevant units
Cluster sampling
Sampling units are divided into mutually exclusive and collectively exhaustive subpopulations called clusters
Non probability sampling
Sampling units do not have equal opportunity to be selected for the study - sample size is less important
Probability sampling
Sampling units have equal opportunity to be selected for the study
Identify the scale: SEARS IS: Powerful --:--:--:--:--: Weak Unreliable --:--:--:--:--: Reliable Modern --:--:--:--:--: Old-fashioned
Semantic differential scale
Questionnaire
Set of questions to measure constructs and record variables of interest
Identify the type of question: In which of the following categories does your age (in years) fall? _____ Less than 18 ______ 18-30 ______ 31-45 ______ 46-60 ______ Over 60
Structured/ordinal
Itemized rating scales
The respondents are provided with a scale that has a number or brief description associated with each category
frequency distribution
Univariate and Bivariate
Way to increase validity
Use multiple dimensions
Way to increase reliability
Use multiple questions for each of the multiple dimensions
Validity
Whether or not we are measuring the right thing
Reliability
Whether or not we can get the same results with our measurement again and again
alternative hypothesis
a statement in which some difference effect is expected
Null Hypothesis
a statement of the status quo, one of no difference or no effect
Structured Questions
closed ended -dichotomous (2 category levels) -or multi-category (having more than 2 answer options)
Data Error
distortions in collected data and mistakes in data coding, analysis, or interpretation
Non-comparative rating scale
evaluates a single product by itself (ex. Likert Scale)
Response category sequence
follows a natural order
Relevance in questions
make the questions so that the respondent CAN and WILL respond. They wont usually respond to questions about personal finances or sex
Conceptual Model
pictorial display of variables and hypothesized or proposed relationships among them
Response category content
says responses should be -collectively exhaustive (taken together) AND -mutually exclusive (not overlap)
Balanced scales
will have a "neutral" or "zero" option
