Summary and transcripts of last (and most important) lecture
45 views 2 purchases
Course
Marketing Research Methods
Institution
Rijksuniversiteit Groningen (RuG)
This is a document containing transcripts and slide info from the Q&A lecture in the end of the course. This lecture was really important to me because suddenly I started to understand several things I did not know before. I wrote down almost everything he said in the lecture.
In addition to this d...
• Try to understand the formulas.
• He can ask us when he shows us an output of an analysis. Give the estimated total
relationship. E.g. for regression or conjoint analysis. And based on that advertising and price
are provided and he asks for a sales prediction. Or what is the certain utility of a new
product. Which scores highest, what is the relative importance of a certain aspect? Know
how to work with such indirect formulas.
• Yes you might have to do some calculations. Not much/difficult, use basic calculator.
• Degrees of freedom: bunch of data and a certain number of observations and each test you
do will require the usage of some of that information. If you do a t-test for one of the
regression coefficients, one parameter consumes one observation. Number of degrees of
freedom is the excess/addition of information you have relative to what you need. It is the
difference between number of observations and the number of parameters you are
estimating. The more info you have the better, makes your test more reliable. 10 is low.
Significant result is then hard.
• Factor analysis. What is the difference between KMO and Barlett’s test of sphericity? You do
these before the actual factor analysis. Bartlett checks whether it is even a good idea to put
together two variables, whether there are some correlations between variables. H0 is there
are no correlations. You want to reject this one, because otherwise you cannot combine any
variables. KMO is slightly different, it also looks at correlations but also to what extent the
variance of the different variables is due to common variance, which is shared by all the
variables or which part is due to specific variance, covariance with other variables. What you
want is that this is low. You want specific (co)variance between certain variables, because
that it really grouping the variables together, whereas if the variance is shared by all the
variables, that does not group anything. The more common variance, the less likely you can
do the actual factor analysis. KMO compares both. KMO and Barlett are really needed before
your factor analysis.
• Rotation. If you have 2 factors and 8 questions per factor, and you get a result like this:
What items really belong to factor 1 or 2? You do
not know. After an orthogonal rotation you can see
it. You have 3 items scoring high on factor 1. On
second factor their link is
basically 0. The other one
has a strong negative
relationship with 1, and still
not linking with factor 2.
Sometimes original solution
is already pretty clear. And after small rotation a little bit better. A 4 factor solution is not
possible to draw.
• High loading, look at
absolute value, not
negative/ positive
difference.
Commonalities
cannot be higher
than 1, eigenvalues
can. For commonalities the square roots are taken. Take the squared loadings and then sum
them. In exam you can get question like: What are the components (Eigenvalues) of the
, factors, what are the communalities of the variables? Component matrix like this will be
shown then. Determine communalities and Eigenvalues, what is the difference?
• ANOVA without interaction. SS is sum of squares. We divide variation by df, so 14/6=2.33.
why do we do so?
you need to
account for the
fact that the more
levels you have in
your factor, the
more likely that is will also capture more variance. So simply by having more levels in your
factor, you will already having more variance. And in order to make up for that, we simply
divide it by the number of df (the number of levels we have in this case in our factor). By
having more levels in our factor, we already can account for more variation. Why do we
divide coupon and size, so two factors (coupon with store size) by this .67? This is 2-way
ANOVA without interaction.
Basically you are always going to
compare the within group, so the
between level variance so basically
the variance explained by our factor
with the remaining unexplained
variance between individuals. So
the extent to which coupon or store size explains some of the variation among individuals
relative to what is not explained. So what is not explained in this case is the residuals. That’s
the remaining unexplained variance between the individuals. We are going to compare the
explained by coupon and size 49 and 43 relative to what is not explained and that is the
residual variance. So difference between observations which are not explained by any
differences in couponing or size. Any f test is based n what do you explain relative to what
you do not explain. What you do not explain, that’s wats captured in your residuals. And so
that is the reason for dividing the explained by the unexplained (49/0.67=74.00 etc.).
Between group variance is really like to what extent are the groups within such factor
different, you have a certain factor with high medium low, to what extent are these groups
different?
That’s your
between
group sum
of squares
and within
each of
these groups
to what
extent are people there different that’s your within group sum of squares. This last one
should be low.
The benefits of buying summaries with Stuvia:
Guaranteed quality through customer reviews
Stuvia customers have reviewed more than 700,000 summaries. This how you know that you are buying the best documents.
Quick and easy check-out
You can quickly pay through credit card or Stuvia-credit for the summaries. There is no membership needed.
Focus on what matters
Your fellow students write the study notes themselves, which is why the documents are always reliable and up-to-date. This ensures you quickly get to the core!
Frequently asked questions
What do I get when I buy this document?
You get a PDF, available immediately after your purchase. The purchased document is accessible anytime, anywhere and indefinitely through your profile.
Satisfaction guarantee: how does it work?
Our satisfaction guarantee ensures that you always find a study document that suits you well. You fill out a form, and our customer service team takes care of the rest.
Who am I buying these notes from?
Stuvia is a marketplace, so you are not buying this document from us, but from seller verawillems. Stuvia facilitates payment to the seller.
Will I be stuck with a subscription?
No, you only buy these notes for $3.24. You're not tied to anything after your purchase.
