Brighton Hove and Sussex Sixth Form College

Data Handling – inferential tests – Validity – 17 marks

Q1.
Read the item and then answer the questions that follow.

A psychologist wanted to see if verbal fluency is affected by whether people think they are
presenting information to a small group of people or to a large group of people.

The psychologist needed a stratified sample of 20 people. She obtained the sample from
a company employing 60 men and 40 women.

The participants were told that they would be placed in a booth where they would read out
an article about the life of a famous author to an audience. Participants were also told that
the audience would not be present, but would only be able to hear them and would not be
able to interact with them.

There were two conditions in the study, Condition A and Condition B.

Condition A: 10 participants were told the audience consisted of 5 listeners.

Condition B: the other 10 participants were told the audience consisted of 100 listeners.

Each participant completed the study individually. The psychologist recorded each
presentation and then counted the number of verbal errors made by each participant.

The results of the study are given in the table.

Mean number of verbal errors and standard deviations for both conditions

Condition A Condition B
(believed audience (believed audience
of 5 listeners) of 100 listeners)

Mean 11.1 17.2

Standard
1.30 3.54
deviation

(a) What conclusions might the psychologist draw from the data in the table? Refer to

Page 1
 Brighton Hove and Sussex Sixth Form College
the means and standard deviations in your answer.
(6)

(b) Read the item and then answer the question that follows.

The psychologist had initially intended to use the range as a measure of

dispersion in this study but found that one person in Condition A had made
an exceptionally low number of verbal errors.

Explain how using the standard deviation rather than the range in this situation,
would improve the study.
(3)

(c) Name an appropriate statistical test that could be used to analyse the number of
verbal errors in the table. Explain why the test you have chosen would be a suitable
test in this case.
(4)

(d) The psychologist found the results were significant at p<0.05. What is meant by ‘the
results were significant at p<0.05’?
(2)

(e) Briefly explain one method the psychologist could use to check the validity of the
data she collected in this study.
(2)
(Total 17 marks)

Page 2
 Brighton Hove and Sussex Sixth Form College
Mark Scheme

M1.
(a) [AO2 = 2 and AO3 = 4]

Level Marks Description

Conclusions in respect of both means and standard deviations are

presented with clarity. Understanding of the relevance of each statistic
3 5–6
is demonstrated. Justifications for each make good use of the values
given.

Conclusions and justification in respect of both means and standard

deviations are relevant, but there is some lack of clarity in both.
2 3–4
Or, one is done well and justified appropriately (most usually this will
be the mean).

One conclusion is drawn or two are partially correct. Any justification is

1 1–2
limited. The answer lacks clarity.

0 No relevant content.

Means

• Conclusion: when people believe they are presenting to a large

audience they are less fluent in their spoken communication than when
they believe the audience is small (or vice versa).
• Justification / Application: this is supported by the difference in the mean
fluency scores which show more verbal mistakes (on average 6 more
mistakes) when the audience is believed to be large (or vice versa).

Standard deviations

• Conclusion: performances of participants in Condition A where audience

is believed to be small are less varied / dispersed / spread out than in
Condition B where audience is believed to be large (or vice versa).
• Justification / Application: lower SD in Condition A suggests that
individual performances in Condition A were more similar to each other
and / or all quite close to the mean of 11.1.

(b) [AO3 = 3]

1 mark – this would be an improvement because the SD is a measure of

Page 3
 Brighton Hove and Sussex Sixth Form College
dispersion that was less easily distorted by a single extreme score.

Plus

1 mark – one that takes account of the distance of all the verbal error scores
from the mean.

Plus

1 mark – not just the distance between the highest verbal error score and the
lowest verbal error score.

(c) [AO2 = 4]

1 mark for naming the t-test for independent / unrelated groups or a Mann-
Whitney test.

Plus

Up to 3 marks for explanation for unrelated t-test. Credit relevant points as

follows:

• can assume interval data because verbal errors can be assumed to be

of equal size (ie one verbal error is equivalent to any other verbal error)
• the experimental design is independent groups
• the psychologist is looking for a difference between the two conditions.

OR

Up to 3 marks for explanation for Mann-Whitney test. Credit relevant points

as follows:

• data should be treated as ordinal. Cannot assume interval data because

verbal errors cannot be assumed to be of equal size (ie one verbal error
is not equivalent to any other verbal error)
• the experimental design is independent groups
• the psychologist is looking for a difference between the two conditions
• SDs are quite different.

(d) [AO1 = 2]

2 marks for a clear and appropriate definition as follows:

This means that there is a less than 5% likelihood that this difference would
occur if there is no real difference between the conditions OR the researchers

Page 4
 Brighton Hove and Sussex Sixth Form College
would have a 95% confidence level.

1 mark for a less clear answer which shows some understanding, eg this
means the researcher can conclude that the difference was not due to chance.

Accept any other valid answer.

(e) [AO2 = 2]

2 marks for a clear and detailed explanation applied to this study.

1 mark for a partial or muddled explanation or one that is only loosely applied
to the study.

Credit answers based on any type of validity. Most answers will refer to either
face or concurrent as follows:

• asking other people if verbal errors are a good measure of verbal fluency
(face validity)
• giving participants an alternative / established verbal fluency test and
checking to see that the two sets of data are positively correlated
(concurrent validity).

Page 5