Marketing 3633 Final Exam Review

Sample

small subset of the population we collect information from with the goal of making projections about what will be true for the entire target audience

Target Population

A collection of individuals/objects/firms that possess the information about which a researcher wishes to make inferences. All members from target population can be surveyed or interviewed but we don’t have infinite resources

Sampling Frame

A population members from which a sample will be actually drawn – Geographic areas, institutions etc.

Probability Samples

Simple Random, Stratified Random, Cluster

Simple Random Sample

Each possible unit has an equal chance of being picked in the sample, useful when entire sampling frame is easily available and accessible, ensures general view of the population

Stratified Random Sample

Divides population into subgroups based on some attribute, useful when subgroups are homogeneous

Cluster Sample

Divide the population into subgroups based on some attribute, useful when subgroups are heterogeneous

Non Probability Sample

Convenience, Judgment, Snowball

Convenience Sample

The sample is selected based on the unit being present at the right place and at the right time.

Judgment Sample

Sample elements are hired by the researcher because they believe that these elements can handle the tasks in the research.

Snowball Sample

A type of judgment sample where a person is chosen by the researcher, and that subject helps identify others with similar characteristics.

Criteria for selecting sample size

degree of homogeneity, precision, and confidence

Degree of homogeneity

If all the people in the population are similar, smaller sample can be sufficient, If large heterogeneity in the population, then larger sample is required.

Precision

How close does the estimate need to be to true population value?, You don’t know the real average of income of the population, You are estimating the average income for the population using a sample.

Confidence

What is the degree of certainty with which we can claim that the population value falls in aprecision range? You select a degree of precision required (Within +/- $1000, +/- $100 or +/- $10 of the average population income) Given your sample size, you can be confident to a certain degree your selected degree of precision.

True

T/F for the selected precision range, if more confidence is required, the sample size must get larger.

Types of Errors

Sampling, Non-coverage, Non-Response, Response

Non-Coverage Error

Errors that arise when one DOES NOT include an individual even though that person is eligible to be part of a sampling frame

Non-Response Error

Occurs when we FAIL to get a response from an individual who is part of the population and was selected to be in the sample

Response Error

Occurs when an individual provides a response to an item, but the response is inaccurate

Solution to Poorly written/designed questions 

Pre-test, follow rules of questionnaire design to ensure smooth flow of the survey

Response rates

serves as an indicator of the overall quality of a data collection effort

Poor questions, lack of interest, failure to gain attention

Very low response rates may indicate

Response Rate

total valid responses/total respondents

Editing Data

Manually inspecting received forms is necessary- Cases with incomplete answers  Wrong answers  Answers that reflect a lack of interest

Data Cleaning

Errors can be introduced by sloppy editing and data entry • Examine frequency distributions on all variables • Check a sample of questionnaires against the data file (audit) • Enter data into two separate files and then compare for discrepancies (preferred)

Handling Missing Data

With online surveys, you can force the respondent to complete an answer before they can submit the survey. It’s important that your survey has minimum amount of missing data – design the questions carefully

Frequency Tables

Tables representing frequency analysis - Count of number of cases belonging to a particular category

Histogram

a chart used to show frequency of the occurrence of values in a data set. Values are grouped into bins

Pivot Tables

A table of grouped values that aggregates the individual items of a more extensive table within one or more discrete categories ( Excel’s version of frequency table)

Bar Graph

values are represented by height of a rectangle of equal width

Pie Chart

Circular graphic divided into slices and used to represent proportions

Convenience Sample

A type of sampling procedure in which sample is selected based on the unit being present at the right place at the right time

Non Response error occurs when

We fail to acquire response from individuals who are selected to be in the sample

Reject the null hypothesis because the absolute value of -4.5 is 4.5 and that is larger than 1.96

In our analysis using a sample, we get a calculated T value of -4.5. We need to decide if we reject the null hypothesis or fail to reject the nul hypothesis at 95% condience. Which is true

83 ± Sampling Error

To answer the question of how much soda is consumed by an average high school kid in a week, we collected a random sample of 1000 high school kids and asked them about their soda consumption habits. The average soda consumers are calculated to be 83 fl Oz. The true population value is calculated as 

True

T/F Higher standard deviation suggests that values are spread  away from the mean

False

T/F A paired sample T-test is used to compare differences in average of two different groups (men v. women)

Chi Squared Test

Used to test if categorical variables are related to each other

Two Sample T Test

Used to compare two groups to test if their averages are the same or different (men v. women)

Paired Sample T Test

Used to test if difference between two sets of observations for the same individual is zero (ex.  test 1 and test 2 scores comparison)