Case 3 Flashcards
Gender
Tel_Cont
are _____ variables
nominal
Satisfaction
Tangi
Assur
Reliab
Loyal
are _____ variables
ordinal
Gender
Tel_Cont
are ____
Frequencies
All variables with a 7-likert scale
are _____
Descriptives
Which type of data do we deal with in case 3?
cross-section data
Whats the SSE?
The difference between the observation (y) and the fitted line (𝑦̂)
o As small as possible, with any other values the SSE becomes larger
–> cannot be explained
Whats the SSR?
Difference between the line (y) and the mean
–> can be explained
b0 and b1 are selected in order to ____
minimize the SSR
Since the sum of the estimated coefficients is close to 1, we suspect that when all 3 constructs simultaneously increase by 1 point…
Satisfaction increases by 1 point as well
To check if the data are consistent with this feature we need to…
Add Tangibles and Assurance to the 1st model and check whether these variables are jointly insignificant
The most plausible reason for the relatively high p-values for Assur1, Assur2 and (especially) Assur5 in the model is…
Collinearity
The fact that the number of dots in the scatterplot is much smaller than the number of respondents is due to the fact that…
the dots with the same values are on top on each other
Independence ⟶ The conditional probabilities of…
one variable being the same across the categories of the other variable
P-value of 0.005 gives us: “if the null were true…
there would be a 0.5% chance of drawing a sample that yields a test statistic as extreme as this one”
The number of respondents who have answered all questions is equal to…
the lowest N in the Correlations table
If all respondents were “fully rational” when it comes to the reverse scale equivalent, then the correlation between loyal2 and loyal3 would be exactly equal to…
-1
If all respondents were “fully rational” when it comes to the reverse scale equivalent, then it would be legitimate to…
base our loyalty-construct on loyal1 and loyal2 only, if you know Loyal2, then you know the answer to Loyal3
What is Cronbachs Alpha
Measurement of the degree to which the included items behave similarly
o The more positively correlated they are the higher the alpha (appeared on every exam)
There is insufficient info to make a motivated choice about the change in Cronbach’s Alpha if we drop…
Loyal1 from our Loyalty construct, and only include Loyal2 and Loyal3R
If we would include Loyal3 rather than Loyal3R in the loyalty construct, the Cronbach’s Alpha would…
probably become negative
If all the respondents are “fully rational” in the sense that they correctly choose the opposite answer in the reversed scale question (Loyal2 and Loyal3), then the R2 of this regression would be…
exactly equal to 1
Since assur3, assur4, and assur5 are missing quite a lot of values, they should not be included in the construct. However, if we do want to include them, then the alternative assurance construct should be…
calculated for somewhat fewer respondents than the smallest number of responses from any of the assur
