Question 7: Significance tests are used on samples of data. So, if we find a sig

Question

Question 7: Significance tests are used on samples of data. So, if we find a significant difference in the sample, what are we assuming about the underlying population and the sample.

Question 8: We learned that when two variables correlate, that does not mean their relationship is causal. Let’s do a logical extension of this claim. I maintain that, while all correlational relationships are not causal, all causal relationships (between A the cause and B the effect) are correlated. Explain

Question 9: Would you need statistical inference techniques if, in fact, you actually had all of the population data? Explain.

Question 10: Sampling error is apparent in nearly all published studies where sample data is used instead of population data. Explain two research methodologies to remedy or at least reduce (1) sampling error and (2) non-sampling error.

Question 11. Most data in the standard normal curve lies within the IQR. Explain this statement with words and an image of the normal curve.

Explanation / Answer

ANSWERS

Q7

If the test shows significant difference, we assume the difference in population is also significant.

Q8

Existence of a causal relationship does not necessarily imply correlational relationship because correlation

implies linear relationship between the response (effect) variable and causative variable(s). Thus, a causal relationship

might as well be quadratic or exponential or logarithmic. In such cases, the correlation coefficient would be insignificant.

Q9

Strictly speaking, No. but, practically, even when population data is fully collected, a sample audit is very useful.

Population studies involve 100% enumeration which can be tedious and a certain amount of fatigue sets in the

researcher which can adversely affect the results and the conclusions drawn there-from. So, it is always advisable to have a sample audit to

to validate the population studies.   

Q10

Two tools available at the disposal of an experimentor to minimize error are randomization and replication.

The former eliminates or at least minimizes bias and the latter can help detecting out-liers and also facilitate

to bring out interaction effects.

Related Questions

Navigate

• Browse (All)
• Browse Q
• Subjects

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.