INFERENTIAL STATISTICS SEM-4

INTRODUCTION TO INFERENTIAL STATS

Statistics is a branch of mathematics that deals with samples in order to derive
conclusions about parameters or population values. It is science of classifying,
organizing and analyzing data. There are 2 types of statistics: Descriptive and
inferential. Descriptive stats involve methods for summarizing and describing
data. It provides a clear and concise summary of the main features of the
dataset, allowing for interpretation.
Inferential statistics: In general, inference means making a guess about
something. So, statistical inference means, making inference about the
population. Inferential statistics is a branch of stats that allows us to make
predictions, inferences, or generalizations about a population based on data
collected from a sample. It helps in drawing conclusions beyond the immediate
data, estimating population parameters, and testing hypothesis. Unlike
descriptive stats, which summarizes data, inferential stats use probability
theory to assess the reliability and significance of the conclusions drawn from
the sample data. In inferential stats, measures are used to draw conclusions or
make inferences about a population based on a sample. The measures include
t-Test, ANOVA, confidence intervals, etc.
Inferential statistics use a random sample of data taken from a population to
describe and make inferences about the population. Inferential statistics are
valuable when examination of each member of an entire population is not
convenient or possible. For example, to know about the opinion of the entire
country about causes of rape is impractical. Here one can survey a randomly
selected sample from the nation. One can use the information from the sample
to make generalizations about the entire population.

PARAMETRIC AND N0N-PARAMETRIC STATS

Introduction:
Statistics can be broadly classified into parametric and non-parametric
methods based on the assumptions made about the underlying data
distribution.

Parametric Statistics:
Parametric statistics involve statistical techniques that assume the data follows
a specific distribution, usually a normal distribution. These methods rely on
parameters such as the mean, variance, and standard deviation.
 Key Characteristics:
o Assumes underlying statistical distribution (e.g., normal
distribution).
o Involves parameters like mean and variance.
o Requires data to be on an interval or ratio scale.
o More powerful when assumptions are met.
 Examples of Parametric Tests:
o t-test (for comparing means)
o ANOVA (Analysis of Variance)
o Linear Regression
 Advantages:
o Higher statistical power when assumptions hold true.
o Efficient for large sample sizes.
 Disadvantages:
o Not suitable if assumptions (like normality) are violated.
o Sensitive to outliers.

Non-Parametric Statistics:
Non-parametric statistics do not assume any specific distribution for the data.
They are also known as distribution-free tests and are useful when dealing
with ordinal data, skewed data, or small sample sizes.
 Key Characteristics:
o No strict assumptions about data distribution.
o Suitable for ordinal or nominal data.
o Robust against outliers.
 Examples of Non-Parametric Tests:
 o Mann-Whitney U Test (alternative to the t-test)
o Wilcoxon Signed-Rank Test
o Kruskal-Wallis Test
o Chi-Square Test (for categorical data)
 Advantages:
o Flexible with fewer assumptions.
o Can handle outliers and non-normal data effectively.
 Disadvantages:
o Less powerful compared to parametric tests when parametric
assumptions are met.
o Results may be harder to interpret in some cases.

Differences Between Parametric and Non-Parametric Statistics:

Aspect Parametric Statistics Non-Parametric Statistics
Assumes specific distribution
Assumptions No distributional assumptions
(e.g., normal)
Interval or ratio data Ordinal, nominal, or non-normal
Data Type
interval data
Robustness Sensitive to outliers Robust to outliers
Statistical Higher (if assumptions are
Lower (but flexible)
Power met)
t-test, ANOVA, Linear Mann-Whitney U, Chi-Square,
Examples
Regression Kruskal-Wallis

RANDOM SAMPLING DISTRIBUTION OF MEANS

(Suruchi mam ne gc pr handwritten notes upload kre h iske which are very
good and short)

NULL AND ALTERNATE HYPOTHESIS

A hypothesis is a predictive, testable statement that establishes a relationship
between an independent and a dependent variable. It acts as a proposition
that guides scientific research to determine its validity.
Characteristics of a Good Hypothesis:
Clear and Precise – Should be easily understood.
Testable – Must be capable of scientific verification.
Relational – States the relationship between variables.
Specific – Limited in scope, focusing on narrow aspects.
Simple Language – Easy to comprehend without losing significance.
Consistent with Facts – Aligns with established knowledge.
Timely Testing – Can be tested within a reasonable timeframe.
Explains Observations – Provides logical reasoning for the problem.

Functions of Hypotheses:
Testing Theories – Converts theories into testable propositions.
Suggesting Theories – Hypotheses can inspire new theories.
Describing Phenomena – Provides insights about social behaviors.
Informing Policy & Practice – Impacts education, social policies, etc.

Types of Hypotheses:
1. Based on Generality:
o Universal Hypothesis: Applies universally (e.g., "Light improves
reading efficiency").
o Existential Hypothesis: Applies to at least one case (e.g., "At least
one schizophrenic has no delusions").
2. Based on Purpose:
o Causal Hypothesis: Identifies cause-effect relationships (e.g.,
"Boring ads cause channel switching").
o Descriptive Hypothesis: Describes behaviors without explaining
causes.
3. Other Common Types:
o Simple Hypothesis: Involves one or two variables (e.g., "Reward
improves learning").
 o Complex Hypothesis: Involves multiple variables (e.g., "Socio-
economic status affects adult adjustment").
o Research Hypothesis: Derived from theoretical assumptions, also
called a working hypothesis.
Alternate hypothesis and null hypothesis are two other important types of
hypotheses.
ALTERNATIVE HYPOTHESIS: An alternative hypothesis is a hypothesis about a
population parameter that contradicts the null hypothesis. HA is a symbol for
the alternative hypothesis. If the null hypothesis is rejected by the results of
the experiment, an alternative hypothesis is accepted and the research
hypothesis is strengthened. Alternative hypothesis is sometimes also called 2
experimental hypothesis or statistical hypothesis. The alternative hypothesis
may be directional or nondirectional depending on the purpose of the study.
There are two important types of alternative hypothesis: - One-tailed
(directional) and - Two-tailed (non-directional) hypothesis
One-Tailed and Two-tailed Hypothesis
Two-Tailed Hypothesis: When the alternative hypothesis is non-directional, a
two-tailed test results, and it is possible to detect a difference between the
true value and the hypothesized value of the parameter regardless of the
direction of the difference. We would want this capability when examining
many research questions. For example, in most cases in which the performance
of a group is compared with a known standard, it would be of interest to us to
discover that the group is superior or to discover that the group is substandard.
One-Tailed Hypothesis: Sometimes a directional test (a one-tailed test) is more
appropriate. In a directional (one-tailed) test, our interest is in discovering
whether or not there is a difference in a particular direction. Some examples of
situations in which a one-tailed test might be suitable are as follows:
1. A social psychologist wishes to know whether children display more
aggressive behaviors after watching violence on television.
2. The physical fitness of schoolchildren is tested. If mean performance is
substandard, it will be necessary to institute a special physical training
program.
 3. A new teaching method is proposed and it will be used only if it can be
demonstrated that learning accomplished under the new method is superior to
that under the standard method.
In a one-tailed test, the alternative hypothesis limits what the researcher has a
chance to discover.
The major disadvantage of a one-tailed test is that it does not allow for any
chance of discovering that reality is just the opposite of what the alternative
hypothesis says.
In general, a directional alternative hypothesis is appropriate only when there
is no practical difference in meaning between retaining the null hypothesis and
concluding that a difference exists in a direction opposite to that stated in the
directional alternative hypothesis. The decision to use a one-tailed alternative
should always flow from the logic of the research question.
Therefore, the time to decide on the nature of the alternative hypothesis is at
the beginning of the study, before the data are collected. It is not proper to
observe the sample outcome and then set the region of rejection in the tail of
the sampling distribution toward which the sample outcome tends.
NULL HYPOTHESIS: A null hypothesis is a hypothesis about a population
parameter that a researcher tests. H ० is the symbol for the null hypothesis. As
Per Minium, Although Ho is referred to as the null hypothesis, no special
meaning should be attached to the term null. It does not mean “zero” or “no
difference.” The null hypothesis is simply whatever hypothesis we choose to
test. As Per Kothari, null hypothesis denotes that there is no significant
difference between the two groups or variables under study. ● Both Ho and HA
are always statements about the population parameter. They are never
statements about the sample statistic.
WHEN DO WE RETAIN AND WHEN DO WE REJECT THE NULL HYPOTHESIS?
Deciding when and when not to retain null hypotheses is not quite simple. This
decision is somewhat arbitrary, but common research practice is to reject Ho if
the sample mean is so deviant that its probability of occurrence by chance in
random sampling is .05 or less. Such a criterion is called the level of significance
and is symbolized by the Greek letter - α (alpha). The level of significance is the
probability value that is used as a criterion to decide that an obtained sample
statistic has a low probability of occurring by chance if the null hypothesis is
true (resulting in rejection of the null hypothesis).
CONFIDENCE LEVEL AND SIGNIFICANCE LEVEL
Confidence Level
A confidence level refers to the percentage of all possible samples that can be
expected to include the true population parameter. For example, suppose all
possible samples were selected from the same population, and a confidence
interval were computed for each sample. A 95% confidence level implies that
95% of the confidence intervals would include the true population parameter.

Level of significance
The significance level, also denoted as alpha or α, is the probability of rejecting
the null hypothesis when it is true. The rejection or acceptance of the null
hypothesis is based upon the level of significance, which is used as a criterion.
The levels of significance are also known as alpha levels. In psychological,
sociology and educational resources there are two levels of significance which
are commonly used for testing the null hypotheses.
One is the 0.05 or 5% level and another is the 0.01 or 1% level of significance. If
the null hypothesis is rejected at the 0.05 level, it means that 5 times in every
hundred applications of the experiment the null hypothesis is true and 95
times this hypothesis would be false. In other words, this suggests that a 95%
probability exists that the obtained results are due to the experimental
treatment rather than due to some chance factors.
Rejection of the null hypothesis when, in fact, it is true, constitutes a Type l
error or alpha error. Thus, it can be said that at the 0.05 level of significance,
the experimenter commits a 5% Type I error when he rejects a null hypothesis.
Some investigators may want a more stringent test and the 0.01 level is one
such level where the investigator commits a Type l error of 1% only. The 0.01
level suggests that a 99% probability exists that the obtained results are due to
the experimental treatment, and hence once in 100 replications of the
experiment, the null hypothesis would be true. In testing the significance of the
obtained statistics, sometimes the investigator accepts the null hypothesis
when, in fact, it is false. This error is technically known as the Type Il error or
beta error.

INTERPRECTING THE RESULTS OF HYPOTHESIS TESTING:

TYPE-I AND TYPE-II ERROR
POWER OF A TEST
Factors Affecting the Power of a Test
1. Effect Size (Difference Between Hypothesized Mean and True Mean):
o Power increases with a larger difference between the true
population parameter and the value specified in the null
hypothesis.
o Formula: Effect Size = True Value - Hypothesized Value
 2. Sample Size (n):
o A larger sample size reduces the standard error, increasing the
power of the test.
o Key Point: Increasing sample size is the simplest way to boost
power.
3. Variability of the Measure (Standard Deviation, Sd):
o Reducing variability (standard deviation) increases power.
o This can be achieved by controlling extraneous variables through
careful planning.
4. Level of Significance (α):
o A higher significance level (e.g., 0.05 instead of 0.01) increases
power because the rejection region is larger.
o Trade-off: Reducing α lowers the chance of a Type I error but
increases the risk of a Type II error, reducing power.
5. Choice of One-Tailed or Two-Tailed Test:
o One-tailed tests have more power when the direction of the effect
is correctly specified because the rejection region is concentrated
in one tail.
o Two-tailed tests are less powerful in comparison since the
rejection region is split between both tails.

DEGREE OF FREEDOM
📊 Degree of Freedom (DF) in Statistics
The concept of Degrees of Freedom (DF) is fundamental in statistics, especially
when dealing with hypothesis tests, confidence intervals, and variance
calculations. It might seem abstract at first, but once you grasp the idea, it
becomes much clearer. Let's break it down step-by-step:

🎯 1. What Is Degree of Freedom?

  Definition:
The degree of freedom (DF) refers to the number of independent values
or observations in a statistical calculation that are free to vary while still
satisfying a given constraint.
 Simplified Meaning:
Think of DF as the number of choices you have when making
calculations. It represents how much "flexibility" your data has.

🧩 2. Basic Example to Understand DF

Example 1: Simple Numbers
Imagine you have 3 numbers that must add up to 100:
 You can freely choose the first two numbers (say, 40 and 30).
 But the third number is fixed because it has to make the sum 100 (i.e.,
100 - 70 = 30).
In this case:
Degrees of Freedom (DF)=3−1=2\text{Degrees of Freedom (DF)} = 3 - 1 = 2
 Why? Because after fixing the first two numbers, the third is no longer
free to vary—it’s constrained.

📏 3. Degrees of Freedom in Formulas

✅ (a) For Sample Variance (s²):
When calculating the variance or standard deviation of a sample, we use the
formula:
s2=∑(xi−xˉ)2n−1s^2 = \frac{\sum (x_i - \bar{x})^2}{n - 1}
 Here, n−1n - 1 is the degree of freedom.
 Why subtract 1? Because once you’ve calculated the sample mean (xˉ\
bar{x}), only n−1n - 1 data points can vary freely. The last data point is
determined by the mean constraint.

✅ (b) For t-Test (Comparing Means):

When comparing means using the t-test, the degrees of freedom depend on
the type of test:
 One-sample t-test: DF=n−1DF = n - 1
 Two-sample t-test (independent groups): DF=n1+n2−2DF = n_1 + n_2 - 2
 The larger the DF, the closer the t-distribution is to the normal
distribution.

✅ (c) For Chi-Square Test:

In a Chi-Square Test, the DF is based on the number of categories:
DF=(r−1)×(c−1)DF = (r - 1) \times (c - 1)
Where:
 rr = number of rows
 cc = number of columns
This formula helps when testing for independence or goodness of fit.

🚀 4. Why Are Degrees of Freedom Important?

1. Affect Test Statistics:
DF determine the shape of statistical distributions (like t-distribution,
chi-square distribution).
2. Critical Values Depend on DF:
When you look up critical values in statistical tables (for t-tests, chi-
square tests), they depend on DF.
3. More DF = More Accurate Estimates:
o Higher DF: Data behaves more like the normal distribution.
o Lower DF: Results are more variable, leading to wider confidence
intervals.

🧠 5. Intuitive Summary
  DF = The number of independent data points you have to work with
after applying constraints.
 “Freedom” is reduced when you introduce rules (like fixing the mean or
total).
 In most cases, DF = sample size minus the number of constraints.

🔍 Quick Examples Recap:

Scenario Formula for DF Why?
Sample Variance n−1n - 1 Mean is a constraint.
One-sample t-test n−1n - 1 Mean reduces 1 freedom.
1 mean per group = 2
Two-sample t-test n1+n2−2n_1 + n_2 - 2
constraints.
Chi-Square Test (2x2 (r−1)×(c−1)(r - 1) \times (c Rows & columns have
table) - 1) constraints.