Methods, Measurement and Statistics (2018-2019) – Answers to Exam Questions Series I
Methods, Measurement and Statistics (2018-2019)
Example Exam Questions Series I (ANSWERS)
Q Answer Explanation
1. B Numbers express levels of income, but a score 2 does not mean that the income is
twice as much as an income level of 1, nor is the difference between (for example) 3
and 4 comparable to an income difference between 4 and 5, and 6 and 7, and so on.
So, scores order income level, but scores have no scale level properties.
2. A Measurement level of traveling time is scale level. The distribution of traveling time
is best displayed by means of a histogram.
3. A Correlation is a measure of linear association (i.e., along a straight line). The
relationship in B seems strong but is non-linear and will result in a value for the
correlation coefficient close to 0.
4. C The measurement level follows from the way a variable is operationalized (i.e.,
scored). This is a choice a researcher has to make him/herself, and different
researchers may make different choices even though the same variable is involved.
Think of income: you can measure income in terms of ranks (low, middle, high) or in
terms of monetary units (dollars, euros). The first operationalization results in
ordinal scores, the latter in scale scores. So even though we consider the same
variable, our measurements of the variable may have different levels. So
measurement level is not a property of the variable, but a property of the scores
that reflect the variable.
5. D Researchers are interested in differences in productivity between teams. So object
of study is the team, and thus team is the unit of analysis. Notice that team diversity
and productivity are the variables, and members in the team are the unit of
observation.
6. A Statement is false. One of the requirements of event A being a cause of B is that
event A occurs before even B. Only if event A occurs after event B we can be sure
that A is not a cause of B.
7. A Statement is false. Ecological fallacy means that group results are used for drawing
conclusions about individuals, and not the other way around.
8. A It’s the definition of a sampling frame.
9. C Because participants are asked to come up with specific subjects, it may be best
described by referral sampling.
10. B Because we use respondents who are conveniently available in the theatre on a
random Saturday night, it’s a non-probability sample. The fact that they used a
random selection of persons in the theatre does not change the fact that it’s a
convenience sample (!).
11. D Reliability is a necessary but not sufficient condition for validity:
- A is false: validity can only be high if reliability is also high
- B is false: high reliability does not guarantee validity (reliability is not
sufficient for validity).
- C is false: validity can be low even if the data are reliable
- D is true: if reliability is low, validity cannot be high (reliability is a necessary
condition).
1
Methods, Measurement and Statistics (2018-2019)
Example Exam Questions Series I (ANSWERS)
Q Answer Explanation
1. B Numbers express levels of income, but a score 2 does not mean that the income is
twice as much as an income level of 1, nor is the difference between (for example) 3
and 4 comparable to an income difference between 4 and 5, and 6 and 7, and so on.
So, scores order income level, but scores have no scale level properties.
2. A Measurement level of traveling time is scale level. The distribution of traveling time
is best displayed by means of a histogram.
3. A Correlation is a measure of linear association (i.e., along a straight line). The
relationship in B seems strong but is non-linear and will result in a value for the
correlation coefficient close to 0.
4. C The measurement level follows from the way a variable is operationalized (i.e.,
scored). This is a choice a researcher has to make him/herself, and different
researchers may make different choices even though the same variable is involved.
Think of income: you can measure income in terms of ranks (low, middle, high) or in
terms of monetary units (dollars, euros). The first operationalization results in
ordinal scores, the latter in scale scores. So even though we consider the same
variable, our measurements of the variable may have different levels. So
measurement level is not a property of the variable, but a property of the scores
that reflect the variable.
5. D Researchers are interested in differences in productivity between teams. So object
of study is the team, and thus team is the unit of analysis. Notice that team diversity
and productivity are the variables, and members in the team are the unit of
observation.
6. A Statement is false. One of the requirements of event A being a cause of B is that
event A occurs before even B. Only if event A occurs after event B we can be sure
that A is not a cause of B.
7. A Statement is false. Ecological fallacy means that group results are used for drawing
conclusions about individuals, and not the other way around.
8. A It’s the definition of a sampling frame.
9. C Because participants are asked to come up with specific subjects, it may be best
described by referral sampling.
10. B Because we use respondents who are conveniently available in the theatre on a
random Saturday night, it’s a non-probability sample. The fact that they used a
random selection of persons in the theatre does not change the fact that it’s a
convenience sample (!).
11. D Reliability is a necessary but not sufficient condition for validity:
- A is false: validity can only be high if reliability is also high
- B is false: high reliability does not guarantee validity (reliability is not
sufficient for validity).
- C is false: validity can be low even if the data are reliable
- D is true: if reliability is low, validity cannot be high (reliability is a necessary
condition).
1
