Sie sind auf Seite 1von 20

Introduction to Statistical Tests

Week 18
Ways in which data is analysed
• When analysing data, it is possible to use both
descriptive and inferential statistics in your
analysis.

 Descriptive means to describe something


 Inferential means to reach a conclusion based on
evidence
Descriptive and Inferential Statistics

Typically, in most research conducted on groups of


people, you will use both DESCRIPTIVE and
INFERENTIAL statistics to analyse your results and
draw conclusions.

So what are descriptive and inferential statistics?


And what are their differences ?
Descriptive Statistics
Definition:
Descriptive statistics is the term given to the analysis of data that helps
describe, show or summarize data in a meaningful way such that, for
example, patterns might emerge from the data.

Descriptive Statistical analysis may involve:


•Calculation of the measures of central tendency or spread.
•The use of graphs eg. scatter graph, bar chart etc.

Analysis using Descriptive Statistics would enable a researcher to make


conclusions about the set of data given; patterns or trends in the data.
Limitations of Descriptive Statistics
• Descriptive statistics DO NOT, however, allow us to make
conclusions beyond the data we have analysed or reach
conclusions regarding any hypotheses we might have
made. They are simply a way to describe our data.

• You cannot use the data you have collected to generalize


to other people or objects. For example, if you tested a
drug to beat cancer and it worked in your patients, you
cannot claim that it would work in other cancer patients
only relying on descriptive statistics
Inferential Statistics
Definition:
Inferential statistics is the term given to the analysis of data that uses
techniques that allow us to make generalizations about the
populations from which the samples were drawn. Inferential statistics
is the testing of statistical hypotheses, using mathematical
techniques of the mean values in samples and populations.

Inferential Statistical Analysis may involve:


• Comparing the mean values of one or more groups to other groups
or to a hypothetical value

Analysis using Inferential Statistics would enable a researcher to


make conclusions about a wider population based on a sample,
giving support to a research theory with a degree of confidence
Limitations of Inferential Statistics

• Inferential Statistics is not an exact science. You are


providing data about a population that you have not fully
measured and, therefore, cannot ever be completely sure
that the values/statistics you calculate are correct.

• Remember, inferential statistics are based on the concept


of using the values measured in a sample to estimate/infer
the values that would be measured in a population; there
will always be a degree of uncertainty in doing this.
Inferential Statistics and Hypothesis testing

• Very simply, Inferential Statistics tests whether a


particular theory in research is correct or not
• A theory in research is called an Hypothesis
• Inferential Statistics can not conclude that a
particular Hypothesis is true or not, but it can
‘support’ or ‘reject’ a Hypothesis with a certain
degree of confidence
Hypothesis
• An Hypothesis is a theory about the relationship between two
or more variables

• An Hypothesis stating that there is a correlation or that there


is a difference between groups is called the ‘Alternative
Hypothesis’ and is abbreviated as H(1)

• An Hypothesis stating that there is no correlation or that there


is no difference between groups is called the ‘Null
Hypothesis’ and is abbreviated as H(0)
Hypothesis testing
Once data is analysed using Inferential Statistics, the following
conclusions could be made:

• There is not enough support for the Alternative Hypothesis so


it is rejected and the Null Hypothesis is accepted

• There is not enough support for the Null Hypothesis so it is


rejected and the Alternative Hypothesis is accepted

• Remember, in research, it is just as valuable to accept the Null


Hypothesis as it is to accept the Alternative Hypothesis !
Example of Hypotheses
A medical researcher wants to test a new drug and whether it
increases blood pressure. She divides patients into 2 similar
groups and gives Group A the real drug and Group B a placebo
drug.

H(1) Group A blood pressure will be significantly higher than that


of Group B
H(0) There will be no difference in the blood pressure between
Group A and Group B

Once the data is gathered, they are analysed using one or many different types of
Statistical tests. Which one to use depends on a number of factors ...
Which Statistical Test to use ?
• There are a wide range of Statistical tests and which one
a researcher uses depends on what he or she is
attempting to find out and the types of groups used in
the research.

• For example the researcher could be finding out


whether there is a relationship / correlation between
variables or whether there are differences between
groups. Comparisons could also be made, not between
groups, but between one group and a hypothetical value
Examples of Statistical tests to use
• If finding out the degree of relationship or correlation
between groups –
Spearman Rank Correlation
Pearson Correlation

• If comparing data from one group to a hypothetical


value –
One sample t test
Examples of Statistical tests to use
• If comparing differences between 2 unpaired groups –
Chi Squared test of Independence

• If comparing differences between 3 or more unpaired groups –


One way ANOVA

• If comparing differences between 2 or more paired groups –


Wilcoxon Signed Rank Test

• If comparing 3 or more paired groups –


One way repeated measures ANOVA
How to use Statistical tests
• Once data is gathered, analysis can be done ‘by hand’
using complex mathematical formulas. However,
most researchers would simply input data into a
Statistics package on a computer and have the
calculations done for them !

(University of Liverpool commonly use a package


called SPSS)
Significance levels
• If a researcher rejects a Null Hypothesis and accepts an
Alternative Hypothesis, he or she needs to be very confident that
their data supports this.

• It would be very unlikely to be able to support an Alternative


Hypothesis with 100% confidence, but researchers would like to
be at least 95% or even 99% confident

• A significance level is called a p value.

• If the p value is 0.05 this means that the probability of the same
results occurring by chance is 5% (therefore a researcher can be
95% confident in supporting the Alternative Hypothesis)
How to use Significance levels
• A Statistical test will result in a calculated value and this value will be
compared to a figure on the appropriate statistical table

• A statistical table will have ‘Degrees of Freedom’ along the rows and ‘p
values’ along the columns

• Degree of Freedom is calculated using the formula n-1 and n is the number
of people in the sample

• In the appropriate ‘Degrees of Freedom’ row, the calculated value is


compared to the table value to see if it is greater or not. For example, if the
calculated value is greater than the table value under the column ‘p = 0.05’
then the results support the Alternative Hypothesis with 95% confidence
Example
(Going back to our earlier example, comparing
the blood pressure in Group A and Group B)

1. The data (blood pressure figures) was input and a t test was
used to analyse the data.
2. The t test calculated a T value of 3.796
3. The degrees of freedom were 10 (as 11 people in the sample)
4. The calculated T value and tabulated T values were compared
(see table on next slide)
Using a Statistical Table
At 10 degrees of freedom,
the calculated T value of
3.796 is greater than the
table value in the 0.005
column !

This means that the results


between the two groups are
significantly different and we can
say this with 99.5% confidence !
Conclusion
• The Alternative Hypothesis H(1) (that stated that Group A blood pressure
will be significantly higher than that of Group B) can be accepted with
99.5% confidence.

• The Null Hypothesis H(0) (that stated that there will be no difference in
blood pressure between the groups) can be rejected

• Being 99.5% confident that the differences in blood pressure


between groups A and B were due to the drug is very high.
It means that there is only 0.5% probability that the
differences between the groups was due to chance !

Das könnte Ihnen auch gefallen