Business research

tools

Dr Smita Gupta
Assistant Professor
Indus Business Academy Bangalore
Variables & research design

 Variables are the main focus of research in science, something we can measure,
something that can vary in terms of precision.

 An independent variable refers to the variable which is being manipulated by the

researcher. The researcher wants to study the effect of this manipulation on the
dependent variable.

 A dependent variable refers to the variable which shows us the effect of the
manipulation.
Variables & research design
 Experimental designs:

 The laboratory experiment is more artificial compared to more naturalistic research

settings.

 Experiments involve the manipulation of one variable systematically to see what effect it
has on the other variables in the study.

 Random allocation of participants to conditions of the independent variable is a major

feature of experiments.
Variables & research design
 Quasi experimental designs:
 Investigates groups of individuals and does not use random allocation of participants to
conditions. Because participants are not randomly allocated to various conditions, we
cannot be certain that our pseudo manipulation of the independent variable is
responsible for any differences between conditions.

 Correlational designs:
 Correlational designs allow a researcher to examine the degree and direction of the
relationship between two characteristics or variables. This design examines relationships
between variables and cannot infer causation.  

 Confounding variables:
 Sometimes the difference a researcher has observed in a dependent variable as a result
of manipulating the independent variable may not be due to the manipulation but due to
confounding variables.
Variables & research design
 Using a within participants design means that :
 The same people can be measured twice on the dependent variable.
 You get participants to complete all the various experimental conditions but in
different orders, known as counterbalancing, in order to reduce the order
effects.
 It provides for a more sensitive test of the differences between conditions
because it controls for the differences between individuals.

 A problem with the within participants design is that there can be effects
of a participants serving in more than one condition of the study.
Variables & research design

 Counterbalancing: A solution typically used for dealing with the effects that
can occur in a within-participants design as a result of participants doing the
conditions in a particular order.

 This is the term given when the order in which a participant undertakes two or
more tasks or conditions, affects how they respond e.g. through fatigue,
practice, boredom, etc.
Variables & research design
 Between-participants designs should be considered when:
 The independent variable doesn’t lend itself well to a repeated measure, eg., gender.
 Order effects are likely
 Participants might be affected by demand effects

 If in SPSSFW there is a grouping variable in the data screen this tells you that it is a
Between Participants Design.
Variables & research design
 Type of Research Design and corresponding statistical test to be applied:

 It is important to distinguish between the major sorts of research design because they
use different methods of analysis.

 Experimental designs and quasi-experimental designs: t tests

 Quasi experimental designs: Wilcoxon
 Correlational designs: Spearman’s Rho

 Streiner (2002) argues that in analyses you are only two thirds as likely to detect
relationships among variables if you dichotomise continuous variables. Thus there is a
loss of sensitivity in the study.
Descriptive statistics
 Population refers to a distinct group of people. Sample refers to the people in your study.

 Levels of data: Nominal, Ordinal, Interval, Ratio

 We use descriptive statistics when we want to describe the data.

 Inferential statistics are used to help us generalize from the sample to the whole
population.
 Inferential statistics deals with making conclusions and generalizations about population
from our sample data.

 Mean, median, mode : understand how each one of these is calculated

 The larger the sample size, the closer the sample mean will be to the population mean.
Descriptive statistics

 In probability theory, the central limit theorem establishes that, in some situations, when
independent random variables are added, their properly normalized sum tends toward a
normal distribution even if the original variables themselves are not normally distributed.

 Central Limit Theorem refers to the tendency of the mean of sample means to equal the
population mean.
 Study the application of central limit theorem discussed in class thoroughly.

 Normal distribution theory tells us that for large samples, what percentage of sample means lie
within how many standard deviations above and below the population mean?

 The shape of the sampling distributions is expected to be normal.

Descriptive statistics

 In statistics, sampling error is the error caused by observing a sample instead of the
whole population. The sampling error is the difference between a sample statistic used
to estimate a population parameter and the actual but unknown value of the parameter.

 Thus the degree to which sample statistics differs from the equivalent population
parameters is known as sampling error.

 Sampling bias:
 In statistics, sampling bias is a bias in which a sample is collected in such a way that
some members of the intended population have a lower or higher sampling probability
than others.
Descriptive statistics
 Selection error:
 An error that occurs in sampling when the researcher is pursuing sampling
procedures that are either improper or incomplete.

 Selection bias:
 Selection bias is the bias introduced by the selection of individuals, groups or
data for analysis in such a way that proper randomization is not achieved, thereby
ensuring that the sample obtained is not representative of the population intended
to be analyzed. It is sometimes referred to as the selection effect.
Descriptive statistics, sample distribution
 Read on histogram, scattergram, box and whisker plot.

 We can use box and whisker plot and scattergram to detect outlying scores. However, the
former is more useful in this regard than the latter.

 When we want to find whether a distribution deviates from normality, it may be more difficult
when using a box and whisker plot.

 Read up on range, variance, standard deviation, and how these are statistically computed.

 To be defined as 'normal’, a distribution should be symmetrical about the mean, it should

meet the x axis at infinity, and it should be bell shaped.

 We must be cautious about using the mean as a measure of central tendency, when the data
is skewed, either positively or negatively.
Probability, sampling, distribution
 A sample mean is a point estimate and we do not know how close it is to the population mean.

 In order to use the standard normal distribution you need to transform the scores in
the sample to the standard normal scores.
 The concept and formula for standard error of the mean, z scores must be known thoroughly.

 Given the sample size, mean, standard deviation, how do you calculate the 95% confidence
interval?
 Refer to class notes done earlier, to calculate the confidence interval using formulas.

 How to calculate probability.

 Relationship between sample size and sampling error… read up.

 How do you interpret z scores, when raw scores, mean, standard deviation values are given??
THANK YOU