DATA COLLECTION

1
SAMPLING
Is the method used to make statistically valid inferences when it is not
possible to measure the total population. We have to find a representative
sample.

Why sample?

 Its quicker
 Its cheaper
 Useful where population is changing rapidly

The size of a sample

The greater the size of the sample population, the more closely it will
approximate the population, so one should always chose the largest possible
sample. Take into consideration the aspect of non-response from some of
the sample population

The sampling frame

 The total number of items from which the sample is taken (i.e. the
population) eg police lists of reported crimes in a particular suburb the
sampling frame are the cases that have been reported.
Sampling techniques

1. Random sampling
 Every element of the population has an equal chance of selection. Done
by numbering the entire population or sampling frame and then drawing
numbers from a random number table until the correct sample size has
been reached.

2
 Examples
 If the population consists of 100 items, numbered 1 – 100, then only the
first 2 columns in the random number table are used, with 00 representing
100. If a sample of ten items is to be chosen, the first 10 numbers in
these 2 columns are 89, 86, 98, 94, 33, 85, 32, 17, 04 and 79.

3
  If the population is made up of 10 000 items numbered 1 – 10 000, then
the first four columns would be used, with 0000 representing 10 000.
These numbers would be 8990, 8638, 9810, 9418, 3369 …….. etc.
NB. If numbers occur which would exceed the number of the total pop,
they are discarded and the next random number is used. Numbers can be
read singly, in pairs or in multiples of 3, 4, 5 etc so the first column can be
read all 8, 89, 899, 8990, 89902 etc. Can be ready either vertically or
horizontally (read throughout in the chosen manner)

Advantages

 No biases are introduced into the choice.

 Gives a balanced cross-section of population conditions.
 Can be used with large populations
 It is quicker and simple
Disadvantages

 Can lead to a biased sample if there is a distinct spatial clustering of the

particular categories of the features which are being studied.
 Items can fall into discrete groups such as villages, towns and cities. Random
sampling may create bias towards one particular group. Cannot work where
the total pop is not known eg shoppers.
2. Systematic sampling
 Enables items to be picked up at some regular systematic interval eg every
tenth item on a list, every sixth house on a street, every twentieth shopper
coming out of a shop etc.
 The starting point of the systematic sample should be obtained from the
random number table, so as to ensure objectivity.

Advantages

4
  There is no need to make a pre-survey.
 Requires no previous information, so is excellent for reconnaissance survey.
 Provides an even and complete cover of the population.
 Is simple and easy to apply, so is usually quicker than random sampling.
 It makes the mapping of sample data easy.
Disadvantages

 May introduce bias into the sample, if the sample points coincide with some
other regular feature in the study area.

3. Stratified sampling
 Used when there are very different characteristics between items within the
sampling frame and one wishes to be sure that none are missed.
 The population is divided into strata or sub-sets. Samples are chosen from
each group eg in a survey of population characteristics of an area, the total
population might be divided into urban pop and rural pop. Each stratum
would then be sampled by means of a random sample (stratified random
sampling) or by a systematic sample (stratified systematic sampling)
Total pop

Urban pop Rural pop

High medium low high medium low

income income income income income income

Advantages

 Each stratum or group of items is adequately covered

5
  Innumerable geographical problems which involve large numbers of items can
be analyzed in this way.
 Useful when the items to be studied are not scattered randomly over an area
but are clustered or patterned in some way.
 Useful for small samples
Disadvantage

 To delimit strata accurately one often needs a great deal of basic information
about the sampling frame.
Methods of sampling aerial distributions

Point sampling – where individual points are chosen and the feature is sampled at
those points

Line sampling – where lines are drawn across the map and the occurrence of
features along the line are noted.

Area or quadrat sampling – where squares or quadrats are chosen on the map, or
on the ground, and the occurrence of the features are investigated within those
squares.

6
INTERVIEWS

Is a conversation between the interviewer and the interviewee with a purpose. It is

one initiated by the interviewer for the specific purpose of obtaining relevant
information for the research.

Types

1. Fully – structured- has a predetermined set of questions with responses

recorded on a schedule.
2. Semi – structured – the interviewer works out a set of questions in advance,
but free to modify them during the conversation i.e. change wording or leave
out certain aspects.
Important considerations

 Exercise a lot of tolerance

 Do more listening than talking
 Questions have to be clear and non-ambiguous
 Avoid lead questions eg cutting down trees has led to land degradation in this
area. What is your view? Yes No
 Be confident
 Take a full record of the interview i.e. take down all
 Avoid long questions
Advantages

 These are a flexible and adaptable way of data collection

 Face-to-face interviews offer the possibility of modifying one’s line of enquiry
i.e. make follow-up, probe interesting items.
 One can observe the non-verbal responses

Disadvantages

7
  It is time consuming
 They require careful preparation eg arrangements to visit, securing the
necessary permission and confirming the arrangements.
 Some people are just unwilling to co-operate.
 Recording tapes are expensive’
 Compiling notes into a report is a mammoth task.
THE QUESTIONNAIRE

Contains a list of questions that you as a researcher intend to ask each respondent.
Its purpose is to collect accurate and appropriate information from respondents.

Attributes of a good questionnaire

It must

 Provide valid and reliable answers.

 Be specific and avoid ambiguity.
 Use short questions.
 Avoid questions that call for too much effort from the respondent.
 Be understood by all respondents.
 Be clear and precise
 Have clear questions and answer categories that do not conflict.
 Use familiar language
 Begin with simple questions
 Instruct how to answer eg tick
 Reminds the questioner to thank respondents for their co-operation.
 Avoid questions that are too personal.
The layout of the questionnaire

(a) General information

Comes first/at the beginning;
 The title
 Brief explanation of the survey, who is administering it and why
 Questionnaire number, so it can be identified
 The area being sampled
 The date
 The fieldworker administering each questionnaire
(b) Socio-economic data
Gives information about who is answering what. The following may be
needed to be known
 Age
 Sex
 Marital status
 Family size or household composition
8
  Highest education level attained
 Occupation for at least the household head
 Nationality or citizenship
 Home language
 Race/ethnic group
 Income levels (household income)
(c) Questionnaire data
 These are the questions that will give you the information which you are
trying to find from the survey.
 Ask only what you need to know
Pre-testing the questionnaire

 Is the first test done to an instrument before the actual test is done on the
intended ground
 It is a test for workability i.e. how the items in the instrument are capable of
collecting the intended data.
 This has to be done in an environment similar to the one where the main
survey will be conducted.
Importance

 It tests whether the questions actually ask what the researcher intended.
 Shows where some ideas or questions have been completely missed out.
 Highlights where people do not understand or where they find difficulties.
 It gives the researcher the last chance to make amendments before the main
survey is done.
Advantages

 The same questions are repeated to different respondents thereby enabling

the researcher to ask all respondents exactly the same questions.
 Provides a facility for data processing since in some cases the variables are
pre-coded.
 Information can be transformed into numbers by counting the number of
respondents who gave a particular response.
 No deviation from answers supplied as possible answer (closed-ended
questionnaire).
 Are answered individually and confidentially.
 Researcher’s absence could enhance the removal of bias.
Disadvantages

 People may simply deny responding to it, because they may feel it’s a waste
of time.
 Respondents may respond dishonestly.
 Its time consuming in preparation of

9
  Expensive – stationary
 Limit number of questions that can be asked and the kind of answers that are
obtained.
 The illiterate may not understand what is required of them.
 Do not probe respondents
 Closed-ended questions limit and restrict respondents to specific answers.
 No enough room for respondents to add more information on every question.

Steps to take to make meaningful surveys

 Obtain permission prior to survey.

 Obtain background information where possible eg a map of the town.
 Undertake a pre-survey to check whether interview guide of questionnaire is
well-structured and it really works, also to check on accessibility.
 Decide whether to work solo or as a team. If teamwork is needed, come out
with meaningful groups to carry out the survey.
 Carryout the survey as you have planned.
 All data obtained should be recorded on an information sheet, questionnaire,
map plan, field notebook etc
 Analyzing the results and making sense and comments out of it.
 Presentation of results i.e. maps, diagrams etc
Problems

Each survey has its own problems. Most common include;

 Language if it is interview based.

 Accessibility if there is an element of observation
 Purposeful lying by respondents on questionnaire based interviews.
 Non-co-operation from some sections of the population.
 Some areas are prohibited from entering.

END OF TOPIC

10