Research Methods in Tourism, Hospitality and Events Management

4.5.1 Pragmatic determination of sample size

At several points in this book, we have returned to the question ‘what is the purpose

of the survey?’ and in deciding on sample size this should again be borne in mind.

Where a high level of accuracy is required from the results, and forecasting or other

generalisations are required from the sample about the rest of the population, then

the size of the sample becomes crucial. However, in more descriptive studies, if 350

people out of 500 interviewed requested extra refreshment facilities at a tourist

attraction, then a management decision might reasonably be based on this alone.

Thus, the need for complicated, predictive statistics requiring large samples may be

unnecessary. Perhaps the first question to ask is the level of accuracy required from

the results in relation to the survey objectives.

---

At a second level, the researcher should consider what statistical tests will be

performed on the data. Here again, there is a relationship with the objectives of the

survey. In 

Chapter 7

, we will look at analysing results from surveys and this will

show that some statistical tests require a minimum number of responses on which the

test can be satisfactorily executed. If an objective of the survey is to investigate how

age affects the reaction to a product, then the results might be tabulated as seen in

Table 4.3

.

In 

Table 4.3

, there are 20 cells (5 columns of opinion scores multiplied by 4 rows

relating to age category). For there to be a possible relationship between age and

opinion to be tested, the sample size needs to be sufficiently large for there to be

responses in each cell. The pilot stage in a survey is useful in showing whether the

classification of categories is appropriate.

A third level of questioning which helps to determine sample size is the amount of

resources the researcher has available. There will be a finite budget in terms of the

time available and the cost of the project. The larger the sample, the greater the

resources required. Larger sizes may therefore result in the need for more staff and

an increased number of survey days for interviews. Hence, sample size is often

influenced by the constraints on the resources available.

At a fourth and final level, the researcher must consider the anticipated response rate

shown by the pilot stage. This is particularly applicable to postal or online surveys. If

the requirements of the survey in terms of the accuracy and statistical analysis call

for 1,000 completed questionnaires and a response rate of 50% are anticipated, then

clearly 2,000 questionnaires would be distributed.

To summarise, to determine sample size in a pragmatic way, the researcher should

consider:

the accuracy required of the results in relation to the objectives of the survey

the requirements of statistical tests in the analysis stage

the available resources for the project

the anticipated response rate.

Illustration 4.7 A sampling excerpt from an undergraduate dissertation

‘Determining the level of consumer-based brand equity (CBBE) in the

airline industry’ by Lauren Polhill, supervised by Andreas Walmsley,

Plymouth University, 2016.

Purpose

---

The purpose of this study is to provide a framework to determine a level

of consumer-based brand equity (CBBE) in the airline industry, through

assessing brand knowledge amongst consumers and quantifying the

results, in particular drawing on existing studies by Keller (1993), Aaker

(1996) and Yoo and Donthu (2001), and their approaches to measuring

CBBE. Despite CBBE being widely researched, there is no consensus for

quantifying it, which this study aims to do for Emirates, British Airways

and American Airlines.