Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
44 Cards in this Set
- Front
- Back
chi-square goodness of fit test
|
determines is observed sample frequencies differ significantly from expected frequences
|
|
...reject the null hypothesis
|
If the chi-square is larger than the critical value...
|
|
1) variables must be mutually exclusive
2) all categories for variable must be represented 3) < 5 is not enough data |
necessities for running a chi-square test
|
|
Pearson's Chi-Square
|
determines if there is a relationship between 2 nominal variables
|
|
1) nominal data only
2) results are suspect when there are expected values < 5 |
Notes on chi-square
|
|
correlation
|
quantifies the degree to which 2 quantitative variables are related
|
|
regression
|
To what degree does the level of X cause a variation in Y?
|
|
regression is used for...
|
predicting the dependent variable based on specific values of predictor variables
|
|
explained variance
|
how much better the regression line predicts a value than the average line
|
|
unexplained variance
|
the amount of deviation that the regression cannot explain
|
|
total variance
|
Explained Variance + Unexplained Variance
|
|
R-square
|
- measures how much variation is explained by the model
- will increase with each new predictor |
|
adjusted R-square
|
- modification of R-square that adjusts for number of predictors in model
- increases ONLY if new predictor improves model by amount larger than what chance would predict |
|
nominal; categories
|
Regression will not work when a ____ variable has more than 2 ____
|
|
multicollinearity
|
- does not affect predictive power of overall model
- DOES affect predictive power of individual predictors |
|
highly correlated with Y but minimally correlated with each other
|
multicollinearity: want to have model where predictors are _______ but ________
|
|
- nothing (doesn't influence overall model
- drop insignificant variables - collect more data |
what you do if you have multicollinearity
|
|
multiple linear regression
|
for interval/continuous dependent variable, choose...
|
|
multinomial
|
for nominal dependent variable, choose...
|
|
linear; ratio
|
to do a linear regression, you must have ___ or ___ data
|
|
factor analysis
|
allows us to identify and group similar questions w/in our data
|
|
factor analysis
|
- creates new variables (factor scores)
- can turn answers from many ?'s into a few scores |
|
purpose of factor analysis
|
- data reduction (helps with multicollinearity)
- substantive interpretation |
|
factor analysis basics
|
- requires some correlation b/w variables
- assumes that correlation b/w variables is to do their being dependent on the same factor |
|
Establish need for Information
Detail research objectives and information needs Set research design and data sources Design data collection procedure Design sample Collect data Process and code data Analyze the data Presentation of results |
Marketing Research: The Process
|
|
error
|
deviation between what sample reports and the true values of the target group
|
|
population parameters
|
how a target truly thinks or reacts
|
|
exploratory research
|
used to...
- ask important questions - establish hypotheses - break down research interests |
|
covariation
time-order ruling out alternatives |
3 types of evidence needed to infer causality
|
|
Allows for standardization, which makes it easier to understand group characteristics from individual responses
Allows us to measure what's not directly observable |
why quantify responses?
|
|
Interval
|
Likert, Semantic, Differential (all measures)
|
|
Ordinal
|
rank order (%, mode, median)
|
|
Nominal
|
yes or no (%, mode)
|
|
Necessity
Participant knowledge Question wording Frame of reference Order effects Experimenter effects Cultural sensitivity Asking embarrassing questions |
Things to consider in a questionnaire design (Never Pretend Everything Can Amend Or Fake Questions)
|
|
Question validity
Question reliability Question relevance Scaling Categories Organization Sensitive Questions |
Why is pretesting important?
|
|
Exploratory Research
Pretesting a questionnaire Probability sample too costly Sample size too small Can't reach entire population |
Use Non-Random Sampling when...
|
|
cost of mistakes in generalizing results is high
populations are small list of population members easy to obtain |
Random Sampling
|
|
normal distribution pattern
|
This occurs in a distribution of sample means when the sample size gets large.
|
|
population mean; random sample
|
The mean of all possible sample means of given n is equal to the ___ ___ in a ___ ___
|
|
standard error decreased
more centralized mean values n > 3 (rule) |
when your sample size increases...
|
|
"good matching" procedures in sampling
|
true mean of population is close to mean of sample means
|
|
"bad matching" procedures in sampling
|
true mean of population could be far away from mean of sample means
|
|
chance of rejecting true null hypothesis
|
Type 1 Error (alpha)
|
|
chance of accepting false null hypothesis
|
Type 2 Error (beta)
|