Do you believe in the paranormal? Do you own a cat?

Question

A questionnaire survey was conducted whereby a large number of people were asked a number of questions. Two of the questions were:

• Do you believe in the paranormal?

• Do you own a cat?

The data obtained from this survey led the authors to conclude that people who believe in the paranormal are significantly more likely than non-believers to own a cat.

Which TWO of the following five statements, if true, would explain the above correlation?

A More women than men keep a cat as a pet and more women than men believe in the paranormal.

B People who believe in the paranormal are more likely than non-believers to agree to requests to fill out questionnaires.

C The questionnaire contained so many questions that some significant associations were bound to be found just by chance.

D People who believe in the paranormal tend to answer survey questions unpredictably.

E Cat lovers tend to enjoy answering questions.

Answer 1

The correct answer is (A & C)

~~~Justification~~~

A} This would explain the correlation. The more basic / ‘real’ associations are:

(1) gender is associated with cat ownership (2) gender is also associated with belief in the

paranormal. These two real associations ‘create’ the illusion that cat ownership is associated

with belief in the paranormal.

C} This would explain the conclusion. When stating that two variables are significantly associated, a

researcher would also state what the probability level is. In other words, s/he would state how

probable it is statistically that the association observed could have simply been due to random /

chance factors (in sampling, etc.). The researcher would normally state (for example) that p =

0.05 – this meaning that there is an observed association, BUT that there is a 5% chance that

this (empirically) observed association simply happened by chance. Hence, if a researcher

looked for 20 possible associations in his/her data, then (approx.) one of these associations

should be found to be significant (at p = 0.05) simply by chance / due to random sampling.

~~~Distractors~~~

B} This would NOT explain the conclusion. If this statement were true, then the only effect of this on

the research data would be that more believers than non-believers (in the paranormal) would be

represented in the survey’s sample. Logically, this sampling bias would not (in itself) produce a

link between cat ownership and belief in the paranormal. If the link / association were actually

there, then this would be revealed with or without this sampling bias.

D} If it were the case that believers in the paranormal answer questions in an unpredictable

manner, then this would NOT explain the conclusion. All that this would mean would be that they

would be just as likely to answer ‘no’ (unpredictably) to the questions as they would be to answer

‘yes’ (unpredictably). An association / correlation would, however, be generated if it were true

that people who believe in the paranormal tend to answer ‘yes’ to questions posed in survey

questionnaires.

E} This would NOT explain the conclusion. If this statement were true, then cat lovers would enjoy

answering ‘no’ as much as they would enjoy answering ‘yes’: they would simply enjoy answering

the questions.

===================