2018THESIS Flashcards

1
Q

What else could you have controlled for?

A
  • Financial performance of portfolio companies
  • The number of co-founders
  • Education of co-founders and their previous experience
  • Terms of the investment
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What made you state hypothesis 1 about PCs of public funds having unclear exit strategies?​

A
  • Importance of exits and financial returns is lower due to the dual purpose
  • Lower urgency of returns/lower need to prove exits as public funds do not raise money in the same way
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What made you state hypothesis 2 about PCs of private funds pursuing IPOs more often?

A
  • Reputation benefits from the public disclosure and attention of IPOs
  • Investment focus more often on scalable industries (ICT especially)
  • Also found by Isaksson in Sweden
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What made you state hypothesis 3 about PCs of public funds having a lower intensity of exit-directed activities?

A
  • Lower urgency of returns = Due to the capital source, there is a much lower need to prove exits
  • Poor bonus/incentive structure for MDs of public funds, thus incentivizing exit focus less
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What made you state hypothesis 4 about no significant difference between PCs focus on internationalization?

A
  • Expansion into foreign markets is inevitable to reach a substantial size for successful exits = This goes for all portfolio companies from the small Danish market
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What made you state hypothesis 5 about VCs perceived contribution towards internationalization to be higher among PCs of non-public funds?

A

Knowledge-based / Resource-based view

  • Public funds are geographically constrained in terms of their investments and can thus not get the same knowledge
  • Public funds may want to keep jobs in domestic market due to dual purpose
  • Non-public funds have better MDs due to incentive structure and the knowledge of captives parent-companies
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Why are you not showing the 2 other forms of secondary sale as answer options?

A

The two other forms being VC fund selling share to other VC-fund OR to a financial institution.These are uncommon and IF people had that planned, it could be written in “others”.Also, we took the question from Isaksson’s survey.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Why the chosen methodology?

A
  • Mixed method allowed us to better understand the characteristics of the Danish VC-industry.
  • The survey provided many responses (quantitative) and is believed to much better represent the population rather than a qualitative approach surveying only 1 PC for each VC-fund type
  • Alternatively, one could with very strong cooperation of PCs use number of markets and share of foreign sales (ACTUAL numbers) and already exited PCs exits to answer the research question but this would, first of all, be very hard to find these exited companies and next get them to put in the time necessary for such study.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Why early- and late-stage together?

A

Lockett et al. (2008) found that the later the stage of development, the less VC-funds contribute to the internationalization process due to internally developed resources.We thus chose to separate SEED-stage from other stages as these are the earliest.But stage is not significant for any of the final models - in fact, it is not even included in MODEL 3A and 3C.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What else could you have used instead of the Pearson’s chi-squared test?

A

When having expected counts below 5, Fisher’s exact test is an alternative (for 2x2 at least)But using Chi-square as simply and then also comparable to Isaksson.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What would be the alternative to the use of a principal component factor analysis?

A

Three types of factor analysis: Principal component analysis

  • Goal: DATA REDUCTION Exploratory factor analysis
  • Goal: ESTIMATE THE PATTERNS OF RELATIONS BETWEEN THE COMMON FACTORS AND EACH MEASURED VARIABLE Confirmatory factor analysis
  • Goal: SEEKS TO DETERMINE IF INDICATOR VARIABLES LOAD AS PREDICTED BY THEORY
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What was the process of your backward selection of independent variables in your final regression model?

A
  • Removing the variables with the highest p-value from the non-robust model
  • Next, check if the r-squared ADJUSTED increases or not.
  • If increasing, we will remove another variable until it stops increasing
  • Lastly making the final model robust
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What can explain PCs of public funds higher focus on internationalization? How could this be investigated further?

A

Potentially wrong due to

  • Non-response errors stemming from misunderstandings of question
  • A latent variable not accounted for Potentially the right conclusion. Hard to explain and needs further research:
  • Case studies with PCs may help OR further interviews with VC fund managers focusing on this topic PCs of public funds may find “internationalization to be important” as they are less internationalized than PCs of non-public funds. Being less internationalized is, however, best explained by public funds’ investments in earlier stages
  • but… our models predict that PCs who export more, find international expansion more important
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What is Kaiser-Meyer-Olkin’s (KMO) Measure of Sampling Adequacy?

A

Measures/tests whether a correlation matrix is adequate for factor analysis.“a statistic that indicates the proportion of variance in your variables that might be caused by underlying factors”As long as the KMO value is minimum 0.6, you are good

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What is the K1 approach?

A

Strictly dropping factors with eigenvalues below 1 (Kaiser 1960) => for Kaiser-Meyer-Olkin

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What is Bartlett’s Test of Sphericit?

A

“tests the hypothesis that your correlation matrix is an identity matrix, which would indicate that your variables are unrelated and therefore unsuitable for structure detection.” = and thus unsuitable for factor analysis.As long as you can reject this null hypothesis, you are good.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

What is Principal Component Factor Analysis?

A

“a method to create one or more index variables from a larger set of measured variables”

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

What is an Eigenvalue?

A

The values reported from the Kaiser-Meyer-Olkin test. Must be above 1 when using the K1 approach in determining number of factors.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

What is a Scree plot?

A

Simply a plot of the eigenvalues.Interpreting the scree plot = you should not include factors from the point where it stops dropping substantially / when it flattens out

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

What is the Unrotated solution?

A

The variables will load on both axes making it impossible to see the patterns.

21
Q

What is Rotated factor analysis?

A

In rotated solutions, the reference axes has been changed, thereby making the patterns of loading more obvious.

22
Q

What is an Orthogonal rotation?

A

Rotations produce factors that are uncorrelated = maintain a 90 degree angle.the alternative are “oblique methods” of rotation in which the factors are allowed to correlate = the axes are allowed to assume a different angle than 90 degrees.

23
Q

What is Composite factors?

A

Composite = sammensatteThink this is not a term, just use of a fancy word we didn’t know the meaning of.

24
Q

What is Cronbach’s alpha coefficient?

A

A measure of internal consistency = how closely related a set of items are as a group.Considered to be a measure of reliability (or consistency).Fx used to test whether Likert-scale surveys are reliable.

25
Q

What is Stepwise multiple regression method?

A

In each step, a variable is considered for addition to or subtraction from the set of explanatory variables based on some criterion.The criteria we use is ADJUSTED R-SQUAREDOthers include F-tests or t-tests, Bayesian information criterion, Mallow’s C etc.

26
Q

What is Ordinary least squares (OLS)?

A

A method of estimating the unknown parameters in a linear regression model.MINIMIZING the error of prediction = the squared difference between the real value and the predicted value.

27
Q

What are the 5 Gauss-Markov assumptions?

A

A method of estimating the unknown parameters in a linear regression model.MINIMIZING the error of prediction = the squared difference between the real value and the predicted value.

28
Q

What are the 5 Gauss-Markov assumptions?

A

To properly use, Ordinary-least squares regression, the following 5 assumptions must adhere: Linearity in parameters Random sampling Sample variation (no perfect collinarity) Zero conditional mean (exogeneity) Homoscedasticity

29
Q

What is Cross-sectional regression?

A

Cross-sectional regression is a type of regression in which the explained and explanatory variables are associated with one period or point in time.= Simply cross-sectional regression because we make a cross-sectional study = only one observation for each company at one specific point in time i.e. not a longitudinal study/panel data

30
Q

What is Heteroscedasticity?

A

When the standard deviations of a variable increase with x = when the standard deviations are non-constant.

31
Q

What is White’s test?

A

A test that establishes whether the variance of the errors in a regression model is constant or not = checking for heteroskedasticity (i.e. if the variance / standard deviations are non-constant).

32
Q

What is Breusch-Pagan test?

A

It tests whether the variance of the errors from a regression is dependent on the values of the independent variables = if the variance of errors is constant or notIt tests for heteroskedasticity = just as White’s test..

33
Q

What is Heteroskedasticity-robust procedures?

A

Making the data robust = making sure that your estimations are not unduly affected by outliers = reducing the outliersVCE(robust) uses the robust estimator of variance.

34
Q

What is Variance inflation factors (VIF)?

A

It quantifies the severity of multicollinearity in an OLS regression.Measures how much the estimated variance is increased because of collinearity.

35
Q

What is an F-test statistic?

A

Used for different purposes.We use it for regression models, where the null hypothesis is used to check how well the model fits the dat.

36
Q

What is Standardization?

A

Standardization of our DEPENDENT variables meaning that we standardize our factors to have a MEAN of 0 and a standard deviation of 1.Making it easier to interpret.

37
Q

What is R-squared?

A

Statistical measure that represents the proportion of variance for a dependent variable that is explained by an independent variable.Essentially how much explanatory power the model has for the dependent variable.

38
Q

What is R-squared adjusted?

A

Unlike the normal R-squared, adjusted r-squared only increases when adding useful variables to the model and thus penalizes useless variables.

39
Q

What is Omitted-variable bias (OVB)?

A

A bias that occurs when a statistical model incorrectly leaves out one or more relevant variables.Thus, if a relevant independent variable has been left out, the included independent variables will be overestimated as they compensate for the omitted variable(s).

40
Q

What is Cumulative common variance?

A

For our two factors on internationalisation, together they explain 73.5 % of the common variance of the six sub-questions.

41
Q

What is Uniqueness score?

A

Uniqueness determines whether an individual variable (for us the sub-questions) can be distinguished from all the other members in a sample.Uniqueness is the percentage of variance of a variable that is not explained by the factorsThus, if the uniqueness is low (under 0.50), then the variable is not very unique and can thus better be used in factors.

42
Q

What is Multicollinearity?

A

When one predictor variable can be linearly predicted from other variables.In our case TimeInvestment and CompanyAge have similarity above 0.6. Thus, we omit one of them.

43
Q

What are Residuals?

A

The difference between the observed value and the estimated value.

44
Q

What is Interval data?

A

Interval = we know both the order AND the exact differences between the values.Likert-scale: 1, 2, 3, 4, 5 (this order) and with exactly 1 between each number.

45
Q

What is Synchronous and asynchronous interviews?

A
  • Synchronous = Communicating fast and directly to each other = Warnøe on phone + Bruhn in person
  • Asynchronous = not communicating directly to each other but through e-mails, letters etc. = Brian Mikkelsen on email
46
Q

What is Sampling error and what are its subtypes in the paper?

A

Stems from estimations based on a sample of population. It does not perfectly represent the population
Sampling bias = When some members of the population are less likely to be included
Survivorship bias = We only survey those companies who still exist (not the bankrupt companies)
Self-selection bias = People VOLUNTARILY fill out the survey

47
Q

What are Survey errors and what are the types in the paper?

A

= errors in the formation of the survey
Potential non-sampling errors from invalidated questions = response errors due to misunderstandings of questions
Response biases = biases in terms of the responses due to question order, extremism etc.

48
Q

What are Collection errors and what are the types in the paper?

A

= errors in data collection
Recipient bias = is the recipient is not knowledgeable about the topic questioned
Transient personal factors = people’s mood and fatigue play a role when measuring perceptions
Uncertainty of potential consequences of participation = can cause low response rates if people do not trust the security of information/data