DATA COLLECTION

INTRODUCTION
 This topic will focus on aspects of data-collection
procedures and techniques
 The process of data collection is of critical importance to
the success of a study. Without high quality data
collection techniques, the accuracy of research
conclusions is easily challenged
 It is therefore of utmost importance that the researcher
should be well aware of the various data-collection
techniques, including their advantages and
disadvantages, so that he/she can select the most suitable
and appropriate technique for the study purpose, the
setting and the proposed study population
DATA-COLLECTION PROCESS
 When planning the process of data collection, the
researcher is guided by important questions: What?
How? Who? Where? When?
WHAT DATA WILL BE COLLECTED?
 Researchers must carefully consider exactly what type of
information is needed to answer the research questions e.g.
does the question call for knowledge, attitudes or behaviors?
 If the researcher is concerned with the way crisis situations
affect Cadets, the what of data collection becomes cadets’
behaviors or responses in crises
 Researchers must also consider if they are going to quantify
the data or analyze it qualitatively. If they intend quantifying
the collected data, a decision must be made regarding the
level of measurement or measurement scale to be used. Four
levels of measurements or measurement scales have been
identified: nominal, ordinal, interval and ratio
Nominal scales
- Are used when persons or events or other phenomena are
named or categorized into mutually exclusive categories
e.g. black or white, married or single, divorced or
widowed, dead or alive. Even feelings can be classified
using nominal scales: a person can be happy, sad, angry
etc.
Ordinal scales
- Are used for variables that can be categorized and ranked
ordered or assessed incrementally e.g. the feelings of a
person are not just classified into happy or unhappy, but
very happy, indifferent, unhappy, very unhappy, thus
enabling the comparison between degrees of happiness of
different persons
- Other examples of ordinal scales are slight, moderate or
intense pain
- It would be appropriate to conclude that intense pain is
higher than moderate pain. One would however not
determine the exact quantity of pain difference between
moderate and intense
Interval scales
- Variables within the interval scale of measurement are assigned real numbers
that are categorized and ordered with equal measurement between each category
- The categories in interval data are the actual numbers on a scale (such as on a
thermometer )
- If body temp. was being measured, a reading of 36,2 degrees Celsius, 37.0
degrees Celsius could be another category, and 37.8 degrees Celsius could
constitute a third category
- The researcher would be correct in saying that there is a difference of 0,8
degrees Celsius between the first and second categories and between the second
and third categories (that is there are equal intervals)
- Similarly, if the researcher undertook a study in which a psychological test was
used, the scores on the test would represent interval data.
- Suppose that 200 people did the test; 90 obtained scores between 42 and 49, 30
between 50 and 59, 60 between 60 and 69 and 20 between 70 and 79. these
scores are categorized into interval classes, they are ranked and the
measurements between each class are equal
Ratio scale
- Ratio level of measurement includes data that can be
categorized and ranked, the distance between ranks can be
specified and, in addition, a true or natural zero point can be
identified
- The amount of money in your bank account could be
considered ratio data because it is possible (and quite likely at
times) to be zero
- If a researcher wants to determine the number of pain
medication requests made by two group of patients, it is
possible that some subjects in one group would never ask for
pain medication. This type of data would be considered ratio
data
- Other obvious ratio scales include time, length and weight
HOW WILL THE DATA BE COLLECTED?
 Choosing a data collection instrument is a major decision
that should be made only after careful consideration of
the possible alternatives
WHO WILL COLLECT THE DATA?
 If the researcher is going to collect the data, this question
is easy to answer
 Often however, there are teams of researchers collecting
data, and sometimes other people outside the research
may be used in the data-collection phase
 Sometimes data collectors are paid for their services
 Whenever more than one person is involved, it is
necessary to ensure that the data is being gathered in the
same manner
 Training will be needed for the data collectors, and
checks should be made on the reliability of the collected
data
WHERE WILL DATA BE COLLECTED?
 The settings for data collection should be carefully
determined; e.g. will it be in a carefully controlled
laboratory, a classroom, a ward, a clinic, a home, a
community center, within a specific region, and so on?

When will the data be collected?

- Decisions must be taken with regard to the month, day
and sometimes even the hour, of data collection as well
as how long data collection will take. Frequently, the
only way to answer this question is through a trial run
DATA-COLLECTION TECHNIQUES
 Techniques most frequently used by researchers include:
observations,use of Questionnaires and Interview
Guides, self-report and physiological methods
 An overview of these and other less commonly used
data-collection techniques will be discussed
OBSERVATION
 This is the technique for collecting descriptive data on
behavior, events and situations
 It is very useful in studies because one can observe
behavior/phenomena as it occurs
 To be considered scientific, observation must be made
under precisely defined conditions in a systematic and
objective manner and with careful record-keeping
 All observations must be checked and controlled. These
criteria make the difference between simply observing
the world around one and collecting research data
through observation
 Observations may be structured or unstructured.
 Structured observations involve specifying in advance precisely
what events or behaviors are to be observed and how they will be
recorded, and preparing forms for record-keeping such as check-
lists, categorization systems and rating scales.
 Structured observation is the method of observation most
commonly used in quantitative studies, where the researcher or a
trained observer simply observes and record certain aspects of
subjects’ behavior e.g. nurses willingness to interact with and
listen to patients, or children’s reactions to removal of a leg plaster
 In then latter, the behaviors to be observed may be distress, co-
operation and searching for info. A rating scale could be prepared
that provides a score on distress, cooperation and searching for
info. The researcher then records on the scale what she observes
 Unstructured observation involves the collection of descriptive info. That is analyzed
qualitatively rather than quantitatively
 In this observation, the researcher attempts to describe events or behaviors as they
occur, with no preconceived ideas of what will be seen
 This method of data collection is commonly used in qualitative nursing research
studies
 An unstructured observation may for example be used to describe the behavior of
nurses immediately following the death of a patient. It would involve the full
description of everything the nurse says and does at this time
 The most common forms of record-keeping in this observation are logs and field
notes
 A log is a daily record of events and conversations that took place
 Field notes may include the daily log, but tend to be much broader than a simple
listing of occurrences
 Observations can also be categorized according to the degree of investigator
involvement
 Spradley (1980) outlines five types of participation ranging from non-participation,
where there is no involvement with the research subjects, to complete participation,
where there is total involvement with subjects and environment
TABLE 11.1 DEGREE OF INVOLVEMENT AND
PARTICIPATION IN OBSERVATION RESEARCH
INVOLVEMENT PARTICIPATION EXAMPLE
High Complete The researcher joins a group under
investigation as one of its mentors,
sharing in all activities e.g. working
on a ward as a nurse, doing active
patient care while studying nurse-
patient interaction. Thus the
researcher participates actively
without the knowledge or consent
of the participants

Similar to above, but the researcher.

Though considerably involved with
the activities in the group,
Active participates as researcher and not as
one of the group. The researcher
thus discloses his or her identity
Moderate and purpose to the other patients

The researcher interacts with

members of the group while
studying the activities, e.g. nurse-
patient interaction, but is less
involved with the actual activities
Low Passive The researcher studies
the activities e.g. nurse-
patient interaction,
merely as an observer,
with little interaction or
involvement with the
group
None Non-participant The researcher studies, the
activities, e.g. nurse-patient
interaction, in situations
removed from the subjects’
immediate environment
e.g. from a viewing gallery
or one-way mirror – no
interaction or involvement,
only direct observation
TIMING OF OBSERVATIONS
 It is necessary to plan how observations will be made
 The two main methods are time sampling and event
sampling

Time sampling
- Involves observation of events during certain specified
times e.g. to observe nurse-patient interaction for fifteen
minutes periods during an eight-hour shift would provide
a good sample of interactions. The periods can be either
randomly selected or pre-determined according to the
daily routine of the ward
Event sampling
- Involves observation of an entire event e.g. if a researcher
is interested in determining the nurse-patient interaction
during admission to the hospital, event sampling would
be appropriate because the researcher would observe the
whole admission procedure. In the case of event
sampling, the investigator must either have some
knowledge concerning the occurrence of events or be in
a position to wait for their occurrence
ADVANTAGES AND DISADVANTAGES OF
OBSERVATION
Advantages
 There are many problems that are better suited to an
observational approach than to questionnaires or
interviews because (what people say is often not what
they really do)
 No other data-collection method can match the depth and
variety of information that can be collected when
utilizing observational techniques
 Observational techniques are also quite flexible, and can
be used in both experimental and non-experimental
designs, and in laboratory and field studies
Disadvantages
- There are problems concerning the reactivity of the
observed when they are aware that they are being
observed
- There are also ethical problems if consent is not obtained

- The data obtained is viable to bias and distortion

- Emotions, prejudices and values can also influence the

way behaviors and events are observed
- It is also very time-consuming

- Costly particularly when observers have to be trained

GUIDELINES FOR CRITIQUING
OBSERVATIONAL METHODS
1. Is observation an appropriate approach to obtain the
necessary info?
2. What or who has to be observed?
3. Was a structured or unstructured approach used and to
what extent was the researcher involved?
4. Where did the observations actually take place?
5. How were the data recorded?
6. What steps were taken to minimize observer bias?
SELF-REPORT TECHNIQUES
 When the researcher’s objective is to find out what people
believe, think or know, the easiest and most effective
method is to ask questions directly or the person concerned
 The purpose of questions is to find out what is going on in
the minds of subjects, their perceptions, attitudes, beliefs,
feelings, motives, plans, past events, knowledge levels and
recall, as well as to gather factual info about the subjects
 As subjects have to answer the questions about the study
variable, these questioning techniques are known as self-
report techniques
 Self-report instruments include questionnaires, scales and
interviews
TABLE 11.2 CRITERIA FOR SELECTING THE
INTERVIEW OR QUESTIONNAIRE
INTERVIEW QUESTIONNAIRE
ADVANTAGES/STRENGTHS
The subject need not be able to read or Questionnaires are a quick way of
write obtaining data from a large group of
people
Responses can be obtained from a wide Questionnaires are less expensive in terms
range of subjects (almost all segments of of time and money
the population)
Responses and retention role is high Questionnaires are on of the easiest
research instruments to test for reliability
and validity
Non-verbal behavior and mannerism can Subjects feel a greater sense of anonymity
be observed and are more likely to provide honest
answers
Questions may be clarified if they are mis- The format is standard for all subjects and
understood is not dependent on mood of interviewer
In-depth responses can be obtained
 DISADVANTAGES/WEAKNESSES
Training programs are needed for Mailing of questionnaires may be
interviewers expensive
Interviews are time consuming and Respondents may provide socially
expensive acceptable answers
Arrangements for interviews may be Respondents may fail to answer some of
difficult to make the items
Subjects may provide socially acceptable There is no opportunity to clarify any
responses items that may be misunderstood by
subjects
Subjects may be anxious because answers Subjects may be illiterate
are being recorded
Subjects may be influenced by interviewer The subjects who respond may not be
characteristics representative of the population
Interviewers may misinterpret non-verbal
behavior
QUESTIONNAIRES
 A questionnaire refers to a self-report instrument where
the respondent writes his or her answers in response to
printed questions on a document
 A well designed questionnaire is easy for the respondent
to fill out and is very easy for the researcher to
administer and score
 Yet, a good questionnaire is difficult to develop

 Each aspect, from the items themselves, to the colour of

the paper used, can influence the responses of the
respondents
 The researcher must therefore pay very careful attention
to the development or construction of questionnaires
A well designed questionnaire should:
- Meet the objectives of the enquiry

- Obtain the most complete and accurate info possible; and

- Do so within reasonable limits of time and resources

IMPORTANT POINTS TO BEAR IN MIND
WHEN DEVELOPING A QUESTIONNAIRE
1) Firstly, list the specific research issues to be investigated by the
questionnaire. Clearly specified goals and objectives or research
questions and/or unambiguous hypotheses are the important
precursors of asking the right question
2) Ensure that you have a thorough understanding of the relevant
literature
3) Analyze what kind of info is needed to study the research
questions. Here, use a specification matrix of the various
content areas in which questions covering specific situations are
needed is helpful
4) Finally, formulate the specific questions. Decisions must be
taken on several aspects when formulating questions. Aspects
such as the structure of questions, the phrasing of questions, the
length of the questionnaire and the organization of the questions
will be briefly dealt with later
STRUCTURE OF QUESTIONS
 The researcher has to choose between using
unstructured, open-ended questions or structured, close-
ended questions
 An unstructured open-ended question allows the
respondent to answer the question in any way he or she
sees fit e.g.
- What do you see as a major problem facing nursing
education today?
- are there any circumstances under which the use of
changaa should be legalized?
 Close-ended questions are of great variety such as yes/no
questions, multiple-choice questions, check-list type questions,
true/false questions and ,matching questions e.g.
- Are you male or female?
- Please indicate your annual income level for the previous
year with a tick against the appropriate number
1) R 000,00 to R9 999,99
2) R10 000,00 to R24 999,99
3) R25 000,00 to R34 999,99
4) R35 000,00 and above
 Open-ended questions are not based on preconceived
answers hence appropriate for explanatory studies, case
studies or studies based on qualitative analysis of data as
they generally provide richer, more diverse data than
close-ended questions which limit the answers to the
options provided by the researcher
ADVANTAGES OF CLOSE-ENDED
QUESTIONS
- Facilitates the coding and analyzing of data
- Secondly, respondents are able to complete more close-
ended questions in a given amount of time
- Thirdly, respondents are more willing to complete close
ended items as they are direct and short

Disadvantages
- More Difficult to construct than open-ended times

- The researcher may neglect or overlook some potentially

important responses
- They may be superficial and some respondents may
become frustrated with the limited responses
PHRASING OR WORDING OF THE
QUESTIONS
1) Questions should be simple and short. Complex questions
should be broken down to simpler ones
2) Questions should be unambiguous. Words used should not be
vague or too general
3) Questions should be understandable. Use vocabulary adapted
to the level of education of the participants
4) Questions should not be double-barelled, that is contain two
questions in one e.g. ‘Do you plan to pursue a master’s
degree in military studies and seek administrative position
upon graduation?’
5) Leading questions should be avoided e.g. ‘Don’t you agree
that…?’ or ending with ‘….Is it not so?’
6) Questions should be stated in an affirmative rather than a
negative manner
THE LENGTH OF THE QUESTIONNAIRE
 It must be long enough to obtain the necessary info, but
not so long that it fatigues the respondent
 Treece and Treece (1986) recommended that a
questionnaire should take no more than 20 to 25 minutes
to complete
QUESTION ARRANGEMENT OR
ORGANIZATION
 Arrangement of questions in a questionnaire is critical
 The questions must be arranged in a logical and relevant way
to the respondent. Various strategies exist to help the researcher
1) Group similar questions or all questions about a certain topic
together
2) Ask interesting and/or easier questions first
3) Ask for sensitive info last. It is thought that the respondent is
more likely to answer sensitive questions when they are
placed at the end of the questionnaire
4) Arrange questions from general to specific
5) Repeat the content of a question , formulated in different
parts of the questionnaire. This is a method checking the
truthfulness of answers and the honesty of the respondent
REVIEW OF NEWLY DEVELOPED ITEMS
 When the questionnaire has been drafted, it should be
critically reviewed by others who are knowledgeable
about instrument construction and about the content and
by a lay person who can give meaningful insight based
on his knowledge of the topic and the sample
 The instrument should also be pretested with a small
sample of respondents and then revised if necessary. A
pretest is a trial run to determine, in so far as is possible,
whether the instrument is clearly worded and free form
major biases and whether it is appropriate for the type of
info envisioned
OVERALL APPEARANCE OF
QUESTIONNAIRE
 They should be neat in appearance, grammatically
correct and should contain no typing or spelling errors
 They should not have a cluttered or crowded appearance

 Question should be well spaced and there should be

adequate margins
DISTRIBUTION
 Questionnaires can be mailed, delivered by hand, given in
groups, or one on one, or even given by computer. Each
method has its own advantages
 A covering letter should accompany the questionnaire. The
covering letter is a extremely important and may be the single
most important factor in motivating the respondents to
complete questionnaires
 When a covering letter is being constructed, the researcher
should try to imagine himself or herself as the recipient.
 Complete instructions should be submitted with the
questionnaire. These must be clear and concise
 It is very helpful to provide respondents with an example of
the appropriate way to respond to a particular type of question
INTERVIEWS
 This is a method of data collection where the interviewer
obtains responses from a subject in face-to-face
encounter or through a telephone call or electronic
means
 They are commonly used in explanatory and descriptive
research and in case studies. They are the most direct
method of obtaining facts from the respondent.
 They can also be useful in ascertaining values,
preferences, interests, tasks, attitudes, beliefs and
experiences
TYPES OF INTERVIEWS
 Classified as either structured or unstructured. Most interviews
however, range between the two ends of the continuum and are
referred to as semi-structured

Structured interviews
- Are formalized so that all respondents hear the same question sin
the same order and in the same manner.
- They are most appropriate when straight forward factual info is
desired.
- The instrument used for data collection is the interview schedule

- The interview schedule is a questionnaire with a with close

ended or fixed alternative questions as well as more or less
precise indications of how to answer each question
Unstructured interviews
- More free-flowing with its structure limited only by the focus of research

- It leaves the wording and organization of questions and sometimes even the
topic, to the discretion of the interviewer
- These interviews are conducted more like a normal conversation but with a
purpose.
- They are particularly appropriate for exploratory or qualitative research studies
where the researcher does not possess enough knowledge about the topic to
structure questions in advance of data collection
- They produce more in-depth info on subjects, beliefs and attitudes that can be
obtained through any other data gathering procedure

Semi-structured interview
- Interviewers are generally required to ask a certain number of specific
questions, but additional probes are allowed or even encouraged.
- Both closed-ended and open-ended questions are included in a semi structured
interview
INTERVIEW QUESTIONS
 Question development, sequencing and wording
considerations are similar to those under questionnaire
development and should be reviewed prior to the
interviews
 In addittion, all interview schedules should be pretested
and assessed for reliability and validity
RECORDING INTERVIEWS
 For structured interviews, the data obtained are usually
recorded directly onto the interview schedule or
otherwise on a separate coding sheet
 The process of recording responses should be absolutely
clear to the interviewer
 Data obtained from semi-structured and unstructured
interviews can be recorded on audiotapes or videotapes.
Field notes and logs are frequently the record-keeping
devices for interviews
INTERVIEWER TRAINING
 Training should be provided for all interviewers who will
collect data during a study
 Training should be carried out in groups, so that all
interviewers receive the same instructions
 The more unstructured the interview, the more training
and experience is required of the interviewers
TIME AND SETTING FOR INTERVIEWS
 All interviews should occur at a time that is mutually
convenient for the researcher and the respondent
 Allowing adequate time is crucial to the completion of
the interview schedules
 Interviews may occur in a variety of settings e.g. home,
ward, clinic, school and so on
 Regardless of the setting, th interviewer should attempt
to seek as much privacy as possible for the interview
INFLUENCE OF INTERVIEWERS ON
RESPONDENTS
 In face-to-face interviews, the interviewer can have a
great deal of influence on the outcome
 Studies have shown that gender, ethnic origin, manner of
speaking and clothing influence the answers provided by
respondents
 In telephone interviews, the interviewer’s verbal
mannerisms, such as tone of voice and dialect, can be a
positive or negative factor in obtaining co-operation
from respondents
FOCUS GROUP INTERVIEWS
 Another form of interview that deserves mention is the focus group
 Focus group interviews with groups of about 5 to 15 people whose
opinions and experiences are requested simultaneously
 Apart from the various practical advantages of interviewing several
people at the same time, it is often useful to allow participants to share
their thoughts with each other
 In this way they spark off new ideas and consider a range of views
before answering the researcher’s questions
 One disadvantage, is that some people are uncomfortable talking in
groups
 The researcher who wants to use focus groups must be skilled at
facilitating group discussions
 Focus groups are particularly useful in participatory and action research
where members of the communities are equal participants in planning
and implementation of research and where the topic of research is a
practical community concern
SCALES
 Are self-reported data-collection instruments that ask
respondents to report their attitudes or feelings on a continuum
 A scale is composed of a set of numbers, letters, or symbols, that
have rules and that can be used to locate individuals on a
continuum.
 There are different types of scales, the most common being
semantic, differential, rating scales, summated rating scales,
Likert scales, Guttmann scales and visual analogue scales
 A likert scale is an example of a summated rating scale which is
frequently used to test attitudes or feelings
 It is summative in that item scores are added to obtain the final
result. It consists of a number of declarative statements about the
topic and five or seven responses for each statement, ranging
from strongly agree to strongly disagree
EXAMPLE
Statement 1 Nurses should practice therapeutic touch in patient care
situations
Strongly disagree Disagree Uncertain Agree Strongly
disagree

1 2 3 4 5

An approximately equal number of positively and negatively worded items should

be included on Likert instrument.
To score a Likert scale, the score responses of all items are added to obtain a total
score. The values obtained are treated as interval data
If five responses are used, as in the example above, scores on each item generally
range from 1 to 5. if 20 items are included on an instrument, the total score could
vary from 20 to 100.
A score of 1 is usually given to ‘strongly disagree’ an a score of 5 to ‘strongly agree’
usually negative worded items are reverse scored: ‘strongly disagree’ is given a score of
5 and ‘strongly agree’ a score of 1
PHYSIOLOGICAL MEASURES
 Because of strong connections between physiological
measures and clinical nursing practice, many nurse
researchers use physiological measures in research
 Among the most familiar measures used are blood
pressure values, urine values and electro cardiograms
 One of the greatest advantages of physiological
measures is their precision and accuracy
OTHER TECHNIQUES
Vignettes
- They are short, descriptive sketches about a situation or event to which
subjects must respond

Records and available data

- A researcher need not collect new data to undertake a specific
investigation. Nurse researchers may for example refer to hospital records,
admission charts, incident reports and sick leave records
- The use of records is an economical source of info, it permits an
examination of trends over time and the investigator does not have to be
concerned with obtaining cooperation from participants
- However, the use of records may be exposed to many sources of error. The
records may contain institutional biases, facts ,may be distorted, some facts
may be omitted, record keeping may be erratic, the collection of some data
may have been stopped for political or financial reasons, and some data are
not readily available because of their confidential nature
Critical incidents
 The critical incident technique is a set of procedures for
collecting direct observations of human behavior in such
a way as to facilitate potential usefulness in solving
practical problems
ANALYSIS OF QUANTITATIVE DATA
 The most powerful tool available to the researcher in
analyzing quantitative data is statistics.
 Without the aid of statistics, the quantitative data would
be little more than a chaotic mass of numbers
 Statistical methods enable the researcher to reduce,
summarize, organize, manipulate, evaluate, interpret and
communicate quantitative data
 Statistical methods can be roughly divided into two
broad categories or groups, namely descriptive and
inferential
ANALYSIS OF QUANTITATIVE DATA
 Descriptive statistics are used to describe and summarize
data
 They convert and condense a collection of data into an
organized, visual representation of data (a picture) in a
variety of ways, so that the data have some meaning for
the readers of research reports
 A descriptive approach employs measures such as
frequency distributions, measures of central tendency
and dispersion or variability and measures of
relationships
ANALYSIS OF QUANTITATIVE DATA
 Inferential statistics are concerned with populations and use sample data
to make an inference about the population
 They permit the researcher to infer that particular characteristics in a
sample exist in the larger population
 Inferential statistics help the researcher to determine whether the
difference that is found between two groups, such as an experimental and
a control group, is a real difference or whether it is only a chance
difference that occurred because a non-representative sample was chosen
from the population
 When an inferential approach is used, P values (the probability that the
outcome was due to chance) are used to communicate the significance or
lack of significance of the data
 Inferential statistics also help the researcher to test hypothesis
 Inferential statistics employed for the latter purpose include chi-square
test, t-test, analysis of variance analysis of co-variance, factor analysis and
multi-variate analysis
DESCRIPTIVE STATISTICS
 Descriptive statistics can be subdivided into groups according to
the summary functions they provide
 An overview of five groups of descriptive statistics will be given
 The intention is that you become aware of the kinds of methods
used to reach particular goals
 Simple formulas and calculations are referred to
 More detail on these measures can be obtained in any handbook on
statistics
 The five groups discussed in the chapter are:
i. frequency distributions and frequency count;
ii. Simple descriptive statistics;
iii. Measures of central tendency;
iv. Measures of dispersion or variability; and
v. Measures of relationship
INFERENTIAL STATISTICS
 There are two kinds of inferential statistical tests, parametric and non-
parametric.
a) Parametric Statistics
 Parametric statistics are applied to data where certain underlying
assumptions have been made.
These are:
1. The variables of concern must be normally distributed within the
targeted population
2. The selected sample must be representative of the target population
3. The variables are measured by an interval or ratio scale
4. The variances of groups compared should be approximately the
same
5. The tests require estimates of parameters (for example, mean and
standard deviation)
Examples of the most common parametric statistical tests used in nursing
research are the t-test and analysis of variance (ANOVA)
INFERENTIAL STATISTICS

b) Non-parametric statistics
 Non-parametric statistics are sometimes referred to as
distribution-free statistical tests, since they are applied to
data where no assumptions are made regarding the
normal distribution of the targeted population
 Non-parametric statistics are usually applied when the
variables have been measured on a nominal or ordinal
scale
 The chi-square is one of the most widely used non-
parametric statistical tests in nursing research
 The chi-square test is appropriate for comparing sets of
data that are in the form of frequencies
INFERENTIAL STATISTICS
 Other examples of non-parametric tests are the Mann-
Whitney test, which is an alternative to the independent form
of the t-test;
 The Wilcoxon test for correlated samples, which is an
alternative to the dependent form of the t-test; and the
Kruskal-Wallis test, which is an alternative to ANOVA for
comparing significant differences between several groups
 When using inferential statistical techniques to test a
hypothesis, the researcher must be well acquainted with
several concepts, such as the probability of P value or
significance, degrees of freedom, critical values, one-tailed
and two-tailed tests of significance, type I and type II errors
 Statistical textbooks must be consulted for clarification of
these terms
INFERENTIAL STATISTICS
 In order to choose the most appropriate inferential
procedure for a study, the researcher must consider several
factors.
Some of the more important are;
1. Are we interested in testing for differences or for
relationships?
2. What is the level of measurement of the variables
(nominal, ordinal, interval/ratio)?
3. Does the level of measurement permit the use of
parametric statistics?
4. What is the size of the sample?
5. How many groups or sets of scores are being compared?
6. Are the observations or scores dependent or independent?
SUMMARY
 In this chapter attention has been paid to data-analysis strategies
 Statistical, graphical and narrative strategies were discussed
 The statistical methods were subdivided into two groups:
descriptive techniques and inferential techniques
 The descriptive techniques were considered and frequency
distributions, measures of central tendency and dispersion or
variability and correction techniques were discussed
 The inferential techniques and graphics were dealt with only briefly
in this chapter
 The function of the various techniques was mentioned, but
calculation methods were not described
 Attention was also given to the analysis of qualitative data
 General criteria for critiquing data-analysis strategies were
proposed to facilitate the reader’s consideration of these aspects
EXERCISES
1) The following scores were obtained from parents of terminally ill children who responded to
a test evaluating their ability to cope with their problem (hypothetical data). Calculate the
mean, median and mode.
2) A nurse researcher is studying the relationship between hopelessness and job satisfaction.
She administers a job satisfaction inventory and a hopelessness inventory to 100 nurses and
compares the results. The mean score on the hopelessness inventory is 65.8 and the mean
score on the job satisfaction inventory is 72. What statistical test should she use?
3) Give an example of how standard deviation could be used in reporting an experience that
you have had.
4) If we applied rank order correlation to some data and found a correlation of 0.88, what does
it show?
5) A researcher has collected data on pulse rate and scores on a final exam for ten students and
would like to know if there is a relationship between the two measures. What statistical test
should be used?
6) What inferential statistic would you choose for the following sets of variables? Variable 1 is
whether an amputee has a leg removed above o r below the knee. Variable 2 is whether or
not the amputee has shown signs of aggressive behaviour during rehabilitation.
7) Ask ten colleagues to describe their conception of preventive health carte and what it means
in their daily lives. Develop a coding scheme to organize the data. What are the major
themes that emerge?