Chapter 16: Research findings and dissemination
Regarding our findings in quantitative studies, we have descriptive
statistics representing the sample, and inferential statistics
demonstrating the differences between different the parts of data.
In qualitative data, we have findings in the forms of descriptions of
key themes and accounts of differences between the two themes. In
both research styles, we need to critically evaluate the findings to
detect any errors.
There are various types of errors:
Measurement errors: This refers to data that is wrong or
inaccurate. This occurs when the research design or data
collection is not accurate.
Classification errors: This is when data is wrongly identified
and put in an inappropriate class. This could be, for example,
when refugees are interviewed and classified as citizens of the
country.
Both classification errors and measurement errors can be
eliminated by being more careful and asking more questions.
Constant errors: These are systematic, repeated errors
throughout the research that can introduce significant bias. An
example of this would be if a researcher unintentionally avoids
certain questions in the interview every time they interview a
participant.
Random errors: This occurs on some occasions, but not others.
It is not systematic and is unpredictable.
NB: Know the difference between errors and blunders.
Errors: Errors introduce bias and inaccuracies in measurement, but
their source can be detected, their seriousness can be evaluated
and their effects can be improved.
Mistakes or blunders: On the part of the researcher, these are
generally unpredictable and undetectable. These are mostly related
to the inexperience or incompetence of the researcher
There are various sources of errors, such as:
1. Vagueness of definitions and inaccuracy of hypotheses or
research questions.
If the hypothesis or research question is too vague, the whole study
can simply confirm the researchers' own biases. Objectivity is
achieved when the hypothesis, research question and definitions
are very specific and clear.
2. Inadequacy of design and planning of research.
This occurs when researchers don't identify all the important
variables, have too many uncontrolled variables or have too many
, sources of errors that they didn't detect.
3. Sampling errors and other errors.
This occurs when the sample either does not represent the
population (quantitative research) or the phenomenon (qualitative
research). The sampling frame could either be faulty (I.e. The
researchers did not use an appropriate sampling technique) or the
respondents themselves did not respond accurately.
4. Imperfection in the research instrument.
These can be faults with the research instrument, such as an
inappropriate length of the questionnaire, the venue, the order of
the presentation of the questions, the type of questions, and the
content, etc. An example of this would be a questionnaire that is too
long, or uses words that the participants don't understand. This
problem causes the researcher now to test the hypothesis or
answer the research question.
5. Bias.
This is the introduction of extraneous variables that distort or
disguise the relationships between variables. The difference
between bias and errors can be explained as follows: If I read a
book that is written in Xhosa, I can prevent errors by looking up the
words that I do not know in a dictionary. However, if I read someone
else's translation, I run the risk of being affected by their bias – the
way that they interpreted the text.
There are several sources of bias:
a) Interviewer bias: The interviewer can affect the respondents'
answers by the way that he reacts and behaves around them, e.g.
his attitude, race, body language, etc. can all affect participants
responses.
b) Respondent bias: These are the result of uncooperative and
unresponsive participants, as well as those who lie or give
inaccurate responses. This can also include the responses of
participants who misunderstood the question. These problems can
be addressed by pilot testing every new data collection instrument.
c) Analyst bias: These include all the errors made while processing,
coding and analysing the data.
d) Researcher bias: The researcher's own philosophical, political and
religious beliefs can bias the researcher during the research
process.
There are various ways that we can avoid bias, such as:
Making sure that all research staff are aware of, and trained in
the issue of bias.
Try to keep the research staff unaware of the research
hypothesis.
Design the data collection tools carefully, making them easy to
Regarding our findings in quantitative studies, we have descriptive
statistics representing the sample, and inferential statistics
demonstrating the differences between different the parts of data.
In qualitative data, we have findings in the forms of descriptions of
key themes and accounts of differences between the two themes. In
both research styles, we need to critically evaluate the findings to
detect any errors.
There are various types of errors:
Measurement errors: This refers to data that is wrong or
inaccurate. This occurs when the research design or data
collection is not accurate.
Classification errors: This is when data is wrongly identified
and put in an inappropriate class. This could be, for example,
when refugees are interviewed and classified as citizens of the
country.
Both classification errors and measurement errors can be
eliminated by being more careful and asking more questions.
Constant errors: These are systematic, repeated errors
throughout the research that can introduce significant bias. An
example of this would be if a researcher unintentionally avoids
certain questions in the interview every time they interview a
participant.
Random errors: This occurs on some occasions, but not others.
It is not systematic and is unpredictable.
NB: Know the difference between errors and blunders.
Errors: Errors introduce bias and inaccuracies in measurement, but
their source can be detected, their seriousness can be evaluated
and their effects can be improved.
Mistakes or blunders: On the part of the researcher, these are
generally unpredictable and undetectable. These are mostly related
to the inexperience or incompetence of the researcher
There are various sources of errors, such as:
1. Vagueness of definitions and inaccuracy of hypotheses or
research questions.
If the hypothesis or research question is too vague, the whole study
can simply confirm the researchers' own biases. Objectivity is
achieved when the hypothesis, research question and definitions
are very specific and clear.
2. Inadequacy of design and planning of research.
This occurs when researchers don't identify all the important
variables, have too many uncontrolled variables or have too many
, sources of errors that they didn't detect.
3. Sampling errors and other errors.
This occurs when the sample either does not represent the
population (quantitative research) or the phenomenon (qualitative
research). The sampling frame could either be faulty (I.e. The
researchers did not use an appropriate sampling technique) or the
respondents themselves did not respond accurately.
4. Imperfection in the research instrument.
These can be faults with the research instrument, such as an
inappropriate length of the questionnaire, the venue, the order of
the presentation of the questions, the type of questions, and the
content, etc. An example of this would be a questionnaire that is too
long, or uses words that the participants don't understand. This
problem causes the researcher now to test the hypothesis or
answer the research question.
5. Bias.
This is the introduction of extraneous variables that distort or
disguise the relationships between variables. The difference
between bias and errors can be explained as follows: If I read a
book that is written in Xhosa, I can prevent errors by looking up the
words that I do not know in a dictionary. However, if I read someone
else's translation, I run the risk of being affected by their bias – the
way that they interpreted the text.
There are several sources of bias:
a) Interviewer bias: The interviewer can affect the respondents'
answers by the way that he reacts and behaves around them, e.g.
his attitude, race, body language, etc. can all affect participants
responses.
b) Respondent bias: These are the result of uncooperative and
unresponsive participants, as well as those who lie or give
inaccurate responses. This can also include the responses of
participants who misunderstood the question. These problems can
be addressed by pilot testing every new data collection instrument.
c) Analyst bias: These include all the errors made while processing,
coding and analysing the data.
d) Researcher bias: The researcher's own philosophical, political and
religious beliefs can bias the researcher during the research
process.
There are various ways that we can avoid bias, such as:
Making sure that all research staff are aware of, and trained in
the issue of bias.
Try to keep the research staff unaware of the research
hypothesis.
Design the data collection tools carefully, making them easy to
