Statistical Tests
Statistical Analysis
- A process of mathematically treating the gathered
data in order to make proper conclusions and
interpretations.
- The type of statistical tool used depends on the type
of data acquired, along with other criteria.
Parametric Tests Non-parametric Tests
Statistical tests used for data Statistical tests used for data
with the following properties: with either of the following
properties:
1. normally distributed data
2. continuous or interval in 1. not normally distributed
nature. 2. nominal or ordinal
3. both of the above
Parametric Tests Non-parametric Tests
1.) T-test 1.) Mann-Whitney U Test
2.) Paired Sample T-test 2.) Wilcoxon Sign Rank Test
3.) Two-sample T-test 3.) Friedman Test
4.) Z-test 4.) Kruskal-Wallis Test
5.) Two-sample Z-test 5.) Chi-Squared Test
6.) 1-way ANOVA 6.) Spearman Rank
7.) 2-way ANOVA
8.) Pearson R
How do we determine whether a
parametric or non-parametric
test is to be used?
Does your data involve
numerical or categorical data?
Numerical Categorical
Use Non-
Does your data involve
parametric
continuous or ordinal?
tests
Continuous/Interval Ordinal
Use Use Non-
Parametric parametric
tests tests
Is your dataset per variable
normally distributed?
1. If p-value is greater than
alpha, the data is normally
Use Wilk-Shapiro Test
distributed.
2. If p-value is less than
YES NO
alpha, the data is not
normally distributed.
Use Use Non-
Parametric parametric
tests tests
Determining the specific
parametric or non-parametric
test to be used.
Testing significant difference (Parametric)
Testing Significant Difference
1 sample group 2 sample More than 2 sample
and population groups groups
Sample size Independent 1 independent 2 independent
Sample size Paired groups
is less than groups variable variables
is at least 30
30
Sample size Paired-
Sample size 1 Way ANOVA 2 Way ANOVA
is less than sample T-
1-sample 1-sample is at least 30
30 test
T-test Z-test
2-sample 2-sample Tukey Test to test which
T-test Z-test groups are different
Testing Significant Difference (Non-parametric)
Testing Significant Difference
2 sample More than 2 sample
1 sample group
groups groups
Wilcoxon Independent Independent Dependent
Sign Rank Paired groups groups
groups groups
Test
Mann- Paired Friedman
Kruskal
Whitney U Wilcoxon Sign Test
Wallis Test
Test Rank Test
Post-Hoc Dunn’s Test to test
which groups are different
1-sample situations
1) You want to test the effectivity of the prototype filter that
you have made. You test the amount of impurities of 10
water samples which have been filtered by your prototype
filter. You compare the average amount of impurities to the
scientifically accepted amount.
2) You want to compare the durability of your prototype
bioplastic to the commercial plastic by measuring the shear
stress of 15 sample bioplastics.
2-sample situations (Independent groups)
1) You want to compare the difference in the bacteria inhibition of
your prototype bandage with chitin and without chitin
2) You want to compare the solubility of coconut-based soap
sheets and papaya-based soap sheets.
2-sample situations (Paired)
1) You want to study the difference in the ratings of two different
flavors of ice creams, where each person gives a rating for
both flavors.
2) You want to test the effectivity of a energy boosting vitamin by
measuring the activity level of certain people before and after
taking the vitamin
1-way ANOVA situations
1) You want to see if there is a difference in the durability of cement
bricks of three different base materials.
a. Coffee bean
b. Egg shell
c. Coconut husk
2) You want to test the if there is a difference in the transparency of
three different bioplastics.
a. Mango-based bioplastic
b. Seaweed-based bioplastic
c. Honey-based bioplastic
2-way ANOVA situations
1) You want to see if there is a difference in the durability of
cement bricks of two different base materials and three
different curing time.
a. Coffee bean a. 7-day curing time
b. Egg shell b. 14-day curing time
c. 28-day curing time
Testing Relationships
Relationships
Numerical Categorical
Parametric Non-parametric Chi-Squared
Test
Pearson R Spearman Rank
Pearson R Sample situations
1) You want to test if there is a negative correlation between the
number of hours spent exercising per week and systolic blood
pressure.
2) You want to determine if there is a linear correlation between
daily temperature (°C) and the growth rate (cm/day) of a
specific plant species.
3) You want to determine the correlation with the concentration
of fruit extract to the pH level of your cosmetic product.
Spearman Rank Sample situations
1) You want to determine whether higher satisfaction survey
scores correspond to higher star ratings in online reviews.
2) You want to relate the stress levels and the academic
performance of students.
3) You want to correlate the concentration of mango extract
to the taste level of your mango ice cream.
Chi-square test situations
1) You want to determine if the province of residence is
associated with academic performance.
2) You want to determine whether the type of herb used in
producing different prototypes of cooking is associated to
the spice level of the cooked food.
