Explain when a z- test would be appropriate over a t-test. Due 3/5/2014
Researchers routinely choose an alpha level of 0.05 for testing their hypotheses. What are
some experiments for which you might want a lower alpha level (e.g 0.01)? What are some
situations in which you might accept a higher level (e.g o.1)? Due 3/7/14
Research in Healthcare
- Researchers routinely choose an alpha level of 0.05 for testing their hypotheses. What
are some experiments for which you might want a lower alpha level (e.g 0.01)? What are
some situations in which you might accept a higher level (e.g o.1)?
The alpha level is a statistical measure of probability of error. Pagano (2012) asserts that
there are various statistical error types that researchers’ are usually subjected to when conducting
their statistical researches and that the various alpha levels are associated with the different types
and incidences of errors. For instance, the alpha level can assert the probability of stating a
positive hypothesis when the study should actually be based on a null hypothesis. In conducting
statistical analyses, two types of alpha levels are normally used and these are low alpha level and
high alpha levels (Pagano, 2012). Each alpha level is applied according to the nature and type of
statistical data that is used in conducting the study.
A low alpha level of 0.01 is used in situations where there is a low probability of Type 1
error. A Type 1 error is occasioned by the rejection of a null hypothesis which happens to be
true. A null hypothesis outlines the default position that the research has assumed and the
researcher’s role is usually to either reject or affirm the null hypothesis (Pagano, 2012). A null
hypothesis which is proved true through the research process should be affirmed while one
which returns a false result should be rejected. Jones (2002) explains that rejecting a null
hypothesis which turns out to be true might significantly interfere with the entire integrity of the
RESEARCH IN HEALTHCARE 2
study being conducted. It is therefore important that researchers conducting a research in which
the integrity of hypothesis is central to the research employ the use of low alpha of 0.01 so as to
limit the probability of a Type 1 error whose occurrence will in turn interfere with the integrity
and credibility of the entire research process (Jones, 2002). Some of the experiments in which
the alpha level should be low include:
A research conducted to establish the relationship between cholesterol levels and cardiac arrest
incidences among middle aged men.
A research conducted to establish the link between the quality of pre-natal healthcare offered to
mothers and infant mortality rates in the Sub-Saharan Countries.
As already stated, low alpha level is associated with low type 1 error. From this, it follows
that high alpha level is associated with high type 1 error (Jackson, 2008). Unlike low Alpha
Level where the integrity and credibility of the entire research process is hinged on correct
treatment of the null hypothesis, a high Alpha level can accommodate errors associated with the
treatment of a null hypothesis. This implies that rejecting a null hypothesis which later turns out
to be true might not significantly impact the integrity of the study (Jackson, 2013). Researches
that usually employ the use of high Alpha level usually are much less sensitive than researches
that usually employ the use of low Alpha level. In such researches, the researchers can actually
afford to make a fundamental mistake and get away with it. Some examples of high Alpha Level
research studies include:
A research conducted to establish the relationship between the type of music that elementary
school children listen to and their mode of dressing.
A research conducted to establish the relationship between the income status of consumers and
the nature of recreation locations they visit.
RESEARCH IN HEALTHCARE 3
References
Jackson, S (2008). Research Methods and Statistics: A Critical Thinking Approach. New York:
Cengage Learning.
Jackson, S (2013). Statistics Plain and Simple. New York: Cengage Learning.
Jones, D (2002). Pharmaceutical Statistics. Michigan: PP Press.
Pagano, R (2012). Understanding Statistics in the Behavioral Sciences. New York: Cengage
Learning.