You are on page 1of 4

Guide to Interpreting Statistical Data

Template for Interpreting Descriptive Data


1. Explain the method/ formula used.
2. Explain how the answer was derived using the formula.
3. Write down the implications of the data/ Solve the problem of the study.

For example:

Company X wanted to adjust current salaries on their employees. They wanted to know first how much are their
employees getting per cutoff.

Mean: Php 32,000. 00

The mean is the average value of a set of values. It is computed by adding all given values and dividing it on the
number of frequencies. Accounting for the salaries of the employees at Company X, it was computed that the
average salary for an employee as Php 32,000.00. This means that the salaries of most employees cost around more
or less Php 32,000.00.

Template for Interpreting Inferential Data


1. State the present hypothesis.
2. Explain the method/ formula used.
3. Explain how the answer was derived using the formula.
4. Make a decision on the hypothesis.
5. Describe the implications of the data.

For example:

A teacher wanted to find out if there is a positive effect of differentiated instruction in discussing basic shapes in
the kinder level. She set up two groups: one control group and one experimental group.

df: 23 confidence level: 5% critical value: 1.714


t-value (control group): 0.896
t-value (experimental group): 1.985

The hypothesis of the study was to investigate if differentiated instruction has a positive effect in discussing basic
shapes in the kinder level. The kinder class was divided into two groups: a control group and an experimental group.
A t-test of correlated means* was used to compute for the difference between the control group and experimental
group to uncover the effect of differentiated instruction. The results showed that the t-value for the control group
was 0.896, while the experimental group was 1.985. Since the study had a df of 23 and a confidence level of 5%, the
critical value was at 1.714. The control group rated lower than the critical value, while the experimental group rated
higher than the critical value. It means that the pupils who had differentiated instruction performed significantly
better than the other group. Therefore, the null hypothesis is rejected. The results also show that differentiated
instruction is an effective strategy to teach shapes in the kinder classroom.

*A t-test of correlated means was used because it is assumed that the study has two samples taken from the same
population, and that there was a pre/post treatment analysis.
What qualifies as descriptive data? What qualifies as inferential data?

Mean, median, mode


Percentages (contingency tables) t-test results
Percentile, quartile, decile information Chi-square
Skewness of distribution ANOVA/ANCOVA
Kurtosis of distribution Correlation (Pearson & Spearman)
Range Regression
Standard deviation

Descriptive data: Functions and assumptions


Measures of Central Tendency

Mean: is the average value of the given data set


Median: is the middle value in the given data set
Mode: is the value with the most number of frequencies in a given data set.

Example:

A restaurant owner evaluated the quality of customer service her employees are providing to customers with the
use of a survey questionnaire. The survey consists of four responses: 4 for Very Satisfied; 3 for Satisfied; 2 for
Dissatisfied; and 1 for Very dissatisfied. The results of the survey are as follows:

Mean: 3.02 Median: 3 Mode: 4

With a mean of 3.02, it means that the customers are satisfied with the customer service they are getting. The
median response was 3, implying that in the distribution of scores, 3 is in the middle. The modal score was 4,
indicating that more customers responded with a 4 or Very satisfied rating.

Among the three measures, mean is the more powerful tool to describe a given data. The median and mode, on the
other hand, are used to merely describe the datas arrangement and frequency.

Measures of Position

Quartile: divides the data into four equal groups


Decile: divides the data into ten equal groups
Percentile: divides the data into 100 equal groups

Example:

A teacher gives a 20-point test to 10 students. (What is the percentile rank of a score of 12?)

18 15 12 6 8 2 3 5 10 20

Following the formula, the answer would be 65. Hence, any student scoring 12 in this quiz would be higher than the
65% of the population that took the same quiz.
However, this works more clearly for percentile because it has more rank scores than decile and quartile; the more
rank scores there are, the better estimate it makes for the rank score. But it does not mean that it is not accurate.
D1 merely corresponds to P10, D2 to P20, etc., as Q1 is to P25, Q2 to P50, etc. The only difference is that a quartile
has four groups (Q1-Q4); decile has 10 groups (D1-D10) and a percentile has 100 groups (P1-P100).

Skewness and Kurtosis

Measures of Variability

Range: difference between the highest and lowest value


Standard deviation: amount of variation or dispersion in a given data set

Exercises: Interpret the following results.

Problem 1
Company Y evaluated their companys CSR programs and activities. They utilized a questionnaire with a 5-item
Likert scale, with the following responses: 5 for Strongly Agree; 4 for Agree; 3 for Uncertain; 2 for Disagree;
and 1 for Strongly Disagree. They administered a survey to the companys employees, which involved the
following factors: employees perception on the CSR activities; employees commitment to CSR; and relevance of
CSR activities. The data was analyzed and the following results came up for each factors.

Employees Perception on CSR Employees Commitment to CSR Relevance of CSR Activities


Mean: 3.42 Mean: 4.78 Mean: 3.22
Mode: 4 Mode: 4 Mode: 3
Standard Deviation: 0.87 Standard Deviation: 1.34 Standard Deviation: 2.21

Inferential data: Functions and assumptions


Chi-square test of independence
Nominal level data
Measured from the same population

*To interpret a chi-square (x2) answer, the value must be compared with a distribution table, varying on the df and
confidence level given in each problem

Students t (One-sample t-test)


Sample size is > 30
Comparing a sample mean from a norm population mean

T-test for two independent samples


Normal distribution
Samples are independent: not related to each other

T-test for two correlated samples


Normal distribution
Experimental setup: pre/post treatment

One-way analysis of variance (ANOVA) (F-test)


Two or more independent samples from the same population

*To interpret

Spearman rho rank correlation


Ordinal level and interval/ratio levels
Paired ranks: data ranked in two variables

Correlation coefficient
Regression
Mann-Whitney U Test
Kruskal-Wallis One-Way ANOVA
Wilcoxon Signed Rank Test (z-test)

You might also like