Statistical hypotheses are of two types

Task 2
Hypothesis Testing
In some cases, it happens that a researcher in interested to study a particular phenomenon
of a population that is totally unknown to him and he makes a guess about that feature based on a
random sample taken from that population (Bonett and Wright 2015). This type of problem is
called testing of hypothesis. It refers to the statistical procedure used to accept or reject the
assumption made by the researcher. Statistical hypotheses are of two types-
 Null hypothesis- It is nothing but the assumption made by the experimenter about the
population parameter and denoted by H0.
 Alternative hypothesis- It is a contrary to the null hypothesis and is denoted by H1.
If the test result is significant then the null hypothesis is rejected and alternative
hypothesis is accepted. The degree of significance which helps to decide whether to reject or
accept the null hypothesis is called the level of significance. There are mainly two types of test-
One tailed test: H0:μ=μ0 vs H1: μ>μ0 or H1: μ<μ0.
Two-tailed test: H0:μ=μ0 vs H1: μ≠μ0
For conducting a null hypothesis, at first the null and alternative hypothesis are
constructed. Then a relevant sample is selected and based on the sample, analysis is done. The
final part is to decide whether to reject or accept the null hypothesis based on the results. It may
happen that a null hypothesis is rejected when it is actually true. This is called type-I error.
Another is type-II error that occurs when a false hypothesis is accepted.

Reliability
In statistical theory, reliability shows the consistency of a measure. High reliability of a
measuring test indicates that the measure will generate similar results in consistent conditions
(Bajpai and Bajpai 2014). There are mainly four types of reliability-
 Inter-rater reliability- It measures the degree of agreement between two or more raters.
For example, if a patient is suffering from fever and different doctors give him same
medicine, then it is a case of inter-rater reliability.
 Inter-retest reliability- It measures the consistency of scores from one period to another.
 Inter-method reliability- It shows the degree of repeatability in test scores when different
methods are used.
 Internal consistency reliability- It is used to determine the consistency of scores across
the items in a test.
Reliability plays a very important role in psychological tests since a test would be valuable
only when it produces consistent results in repetitive experiments (Gnedenko, Belyayev and
Solovyev 2014). It should be noted that reliability does not imply validity. The validity of a test
checks whether the test measures the claim or not, whereas reliability shows the consistency of a
test.
Confidence Level
Confidence level measures the percentage of samples that would contain the true value of
the population parameter. Though the purpose of confidence level and confidence interval are
same, there is a minor difference in the concepts (Hinton 2014). If repetitive samples are taken,

then a certain percentage of confidence interval will contain the true value of the parameter. This
percentage is called confidence level (Schweder and Hjort 2016). For example, a 90%
confidence level implies that 90% of the confidence intervals would contain the true value of the
concerned population parameter. A 0% confidence level tells that if the survey is repeated, then
every time the results will be different. On the other hand, a 100% confidence level tells that no
matter how many repetitive samples are taken, the test result will be same for all the samples. In
practice, a 100% confidence level is not possible. In general, a higher confidence level shows a
greater accuracy of the model.
Internal Validity
The term validity refers to the situation where a researcher is assured that his test results are
correct. There are two types of validity that measure if the test results are truthful or not- internal
and external validity (Baldwin 2018). Internal validity is defined as the degree to which a test
constructs a dependable cause and effect relationship between the given treatment and outcome.
In other words, it shows the casual relationship between a dependent and an independent
variable. The measure of internal validity depends on the number of confounding variables that
have been used in the experiment. If the number of confounding variables is minimised, then
internal validity will be high. Otherwise internal validity will get reduced as more confounding
variables are introduced to the experiment. Moreover, it will imply that the dependent variable is
significantly influenced by the presence of other independent variables apart from those in the
model. There are many factors that can improve the internal validity. Some of them are-

 Randomisation- If the participants are allocated randomly to the treatment and control
groups, then the bias due to sampling will decrease. Consequently, validity will be
higher.
 Blinding- It refers to the method where patients and sometimes the researcher are also
unaware of the treatment under study. This reduces the bias due to knowledge.
 Experimental manipulation- Sometimes the independent variable in the study is
manipulated to get a better internal validity.
Overall, it can be deduced that internal validity is a method to check the durability of the
model.

References
Bajpai, S. and Bajpai, R., 2014. Goodness of measurement: Reliability and
validity. International Journal of Medical Science and Public Health, 3(2), pp.112-115.
Baldwin, L., 2018. Internal and External Validity and Threats to Validity. In Research Concepts
for the Practitioner of Educational Leadership (pp. 31-36). Brill Sense.
Bonett, D.G. and Wright, T.A., 2015. Cronbach's alpha reliability: Interval estimation,
hypothesis testing, and sample size planning. Journal of Organizational Behavior, 36(1), pp.3-
15.
Gnedenko, B.V., Belyayev, Y.K. and Solovyev, A.D., 2014. Mathematical methods of reliability
theory. Academic Press.
Hinton, P.R., 2014. Statistics explained. Routledge.
Schweder, T. and Hjort, N.L., 2016. Confidence, likelihood, probability (Vol. 41). Cambridge
University Press.