STATISTICS IN RESEARCH

A. Types of Statistics Used in Research

Research deals with two types of statistical data: descriptive and inferential.
Descriptive statistics involves numerical descriptions of a particular group.
Inferential statistics is concerned with the process of sampling, the selection for study of a small group
that is assumed to be representative of the large group from which it is drawn.

The researcher may utilize various sampling techniques such as: simple random sampling; probability
sample; area sampling; stratified sampling; and cluster sampling.

B. Statistics and Its Assistance to the Researcher

Statistics assists the researcher in the understanding the data expressed in numerical values, more
completely in four ways:

1) it may indicate the central point around which the data revolve;
2) it may indicate how broadly the data are spread;
3) it may show the relationship of one kind of data to another kind of data; and
4) it may provide certain techniques to test the degree to which the data conform, or depart from
the expected operations of the law of chance or approximate an anticipated standard.

FOUR STEPS IN MAKING SENSE OF DATA

A. Gathering Data

Intelligent data gathering demands that a conscious, concerted efforts must be focused on the
following steps:
1) specifying the objective of the data—gathering exercise;
2) identifying the variable/s of interests;
3) choosing an appropriate design for the survey or scientific study; and
4) collecting the data.

These steps on the other hand, require that the researcher perform the following tasks: understand
the problem under study; examine the objective of the data-gathering exercise; and choose methods
of collecting the data.

In statistics, data can be gathered by the way of a survey, a study, or a combination of the. two. With
reference to gathering data on existing conditions, attitudes or behaviors, surveys are usually passive.
On the other hand, scientific studies are more active, with the researcher varying certain conditions to
reach on the conclusion.

B. Summarizing Data

To make meaningful interpretations, the researcher should summarize the data. Summarizing data
can be conducted by using graphical displays and by focusing on the average or typical value in a
data set and spread or variability of the measurements.
Graphical descriptions of data are very useful and they are easily comprehended by the lay person
and the scientist engaged in economic or sociological research. Pie charts, bar charts, histograms,
frequency polygons, and stem-and-lead plots are graphical methods.

The computation of one or more numerical descriptive measures is the second way of summarizing of
data. These measures are suitable for developing both descriptions and inferences.

C. Analyzing Data

Analysis of data entails utilizing large amounts of data, weighing each bit of relevant information and
arriving at a good inference. The statistician, rather than depending upon his or her own intuition,
utilizes statistical results to help him in making inferences.

In making inferences about parameters, the researcher either estimates the value of the population
parameter of interests or he tests a hypothesis about the value of the parameter. Estimation and
hypothesis testing, two methods of statistical inferences, follow different procedures, and answer two
different questions about parameter. The question “What is the value of the population parameter?”
estimates a population parameter. On the other hand, the question “Is the parameter value equal to
this specific value?” tests a hypothesis.

A statistical test is composed of the following elements: a null hypothesis, a hypothesis about a
population parameter; an alternative research hypothesis, a hypothesis which is accepted if the null
hypothesis is rejected; a test statistic, a quantity computed from the sample data; and a rejection
region, a set of values for the test statistics that are contradictory to the null hypothesis and imply its
rejection.

D. Reporting Results

The final step in making sense of data is the communication of results. The researcher is reminded to
be careful not to distort the results of his own analyses.

Distortions of results may include graphical distortions, biased sampling, and omitting a discussion of
the sampling technique and sample size. Understanding these obstacles may encourage researchers
to strive for more effective communication of results of analyses.

SELECTION AND APPLICATION OF STATISTICAL PROCEDURES: GUIDELINES

A. Talligram (tabulation table), ordered arrangement of scores, score distribution, class (grouped)
frequency distribution or scattergram may be used depending upon what is desired to be
known or what is to be known or what is to be computed, in organizing the data.

B. To know certain proportions of the population based on certain variables such as age, height,
income, etc., frequency counts with their frequency percents may be used.

C. To know the typical, normal or average, the measures of central tendency such as the median,
the mean or the mode may be computed and used.

D. For variables that are abstract or continuous and cannot be counted individually such as
adequacy, efficiency, excellence, extent, seriousness, and the like, the weighted mean may be
computed and used if the average is desired to be known.

E. To know the variability of the population, the measures of variability such as the range,
quartile, deviation, average deviation or standard deviation may be counted and used.
F. To know the relative placements of scores or positions, ranking, quartile or percentile rank
may be computed and used.

G. To know the significance of the trend of reaction or opinion of persons such as a group toward
certain issue ,situation, value or thing but in which there is a neutral position, the chi-square of
equal probability, single group may be computed and interpreted.

H. To determine the significance of the difference between the reactions, or opinion of two distinct
groups in which there is a neutral position, the chi-square of equal probability, two groups, is
computed and used,

I. To find out how one variable varies with another, the coefficient of correlation is computed.

J. To study the significance of the difference between the perceptions of two groups about a
certain situation, the computation of the difference between means is to be used.

K. To find out the relative effectiveness randomized groups are respectively exposed and only a
post test is given to the to the different groups, analyses of variance may be used.

L. To find out the effects of some variable to which they are related, partial and multiple
correlation may be used.

M. To find out the association between two independent variables, the chi-square of
independence or chi-square of multiplication may be use