Hypothesis testing and

making inferences
 Scientific
research

Observational True
studies experiments

Natural Independent Dependent

observations Variable variable

Does change in one variable produce changes in

Parameter another variable?
Estimations

Correlational
Studies
Terms Introduced in Prior semester

• Population  all possible values

• Sample  a portion of the population
• Statistical inference  generalizing from a sample to a population with
calculated degree of certainty
• Two forms of statistical inference
• Hypothesis testing
• Estimation
• Parameter  a characteristic of population, e.g., population mean µ
• Statistic  calculated from data in the sample, e.g., sample mean (x )
Descriptive statistics and inferential statistics
Descriptive statistics
• Used to describe a sample
• Purpose: external validity, generalisability of study
• For example, report age and sex distribution of participants
Inferential statistics
• Relates to populations
• Usually too expensive and time consuming to measure quantity of
interest in whole population
• Use a sample to make inferences about the population
• Two main purposes:
1. HYPOTHESIS TESTING
2. PARAMETER ESTIMATION
 HYPOTHESIS

In statistics, a hypothesis is a claim or

statement about a property/parameter of a
population.
A hypothesis test (or test of significance) is
a standard procedure for testing a claim about
a property of a population.
It is the tentative answer to the research
problem.
 Components of a
Formal Hypothesis Test
 Null/ Statistical
Hypothesis: H0
v The null hypothesis (denoted by H0) is a statement that the value
of a population parameter (such as proportion, mean, or
standard deviation) is equal to some claimed value.
v It states that on general population there is no change, no
difference and no relationship.
v We test the null hypothesis directly. It predicts that Independent
variable has no difference or no effect on dependent variable.
v Either reject H0 or fail to reject H0.
 Alternative/Substantive
Hypothesis: H1/A
v The alternative hypothesis (denoted by H1 or Ha or HA) is the statement that
the parameter has a value that somehow differs from the null
hypothesis.
v The symbolic form of the alternative hypothesis must use one of these
symbols: , <, >.
v Ha can be of two types
o Directional (single-tailed)
o Non-directional (two-tailed)
Directional vs. non-directional
hypothesis
Non-Directional hypothesis:
• Specifies that the independent variable has
an effect on the dependent variable.
• Does not specify the direction of the effect.
•
Directional Hypothesis:
• Specifies that the IV has an Effect on the DP
+ predicts the types of relationship between
the two variables( increase/
decrease; yes/no).
• Uses words like: More than; greater than; positively correlated.
•
Testing Statistical Hypotheses -
steps
Step 1: A specific hypothesis called the null hypothesis is formulated about a
parameter of the population (e.g: about the population mean) along with an
alternative hypothesis.
Step 2: a random sample is drawn from the population of observations, the
value of the sample statistics (e.g.: the value of the sample mean) is
obtained.
Step 3: characteristics of the random sampling distribution of the statistics
under consideration are examined to learn what sample outcomes would
occur (and with what relative frequencies) if the null hypothesis is true.
Step 4: the null hypothesis is retained if the particular sample outcome is in
line with the outcomes expected if the hypothesis is true; otherwise, its
rejected and the alternative hypothesis is accepted.
Formulating hypotheses – class activity
• Research question
• Are IISU students more intelligent than the general population?
• What we expect to find (the alternative hypothesis – H1)
• IISU students are more intelligent than the general population
• What we try to disprove (the null hypothesis – H0)
• There is no difference in the intelligence of IISU students and the
intelligence of the general population
• The null hypothesis – more specific
• The mean IQ score for IISU MSc students is the same as
the mean IQ score in the general population
• The null hypothesis – quantified
• The mean IQ of MSc students = 100
Illustrative Example: “Body Weight”
• The problem: In the 1970s, 20–29 year old men in the
India had a mean μ body weight of 170 pounds.
Standard deviation σ was 40 pounds. We test
whether mean body weight in the population now
differs.
• Null hypothesis H0: μ = 170 (“no difference”)
• The alternative hypothesis can be either
Ha: μ > 170 (one-sided test) or
Ha: μ ≠ Pass
170 (two-sided test)
outcomes of this test:
1. We can reject the null hypothesis
2. We can fail to reject the null hypothesis
Levels of significance
• We always evaluate the results of an experiment by assessing the null
hypothesis.
• We directly assess the null hypothesis because we can evaluate the
probability of chance events, but there is no mathematics for the
probability of alternative hypothesis.
• We start any research by assuming that the Null hypothesis is true
and then test the reasonableness of this assumption by calculating

• This critical probability level is called alpha (α) level or level of

the probability of getting the results if chance alone is operating.

significance or p-value
Strength of evidence
1

α levels are 0.05

• The commonly used No evidence against the

(5%) and α
null hypothesis

0.01(1%) 0.1
Weak evidence against the
• BUT p values are null hypothesis
more nuanced
0.01
• The frequent use of
0.05 and 0.01 levels Increasing evidence against the
null hypothesis
of significance is a
matter of convention 0.001
having little Strong evidence against the
scientific basis. null hypothesis
0.0001
Null hypothesis

-9 -6 -3 0 3 6 9 12 15 18
Treatment effect
Alpha criterion

-9 -6 -3 0 3 6 9 12 15 18
Treatment effect
Reject the null hypothesis or not

Do not reject H0 Reject H0

-9 -6 -3 0 3 6 9 12 15 18
Treatment effect
Alternative hypothesis

Do not reject H0 Reject H0

-9 -6 -3 0 3 6 9 12 15 18
Treatment effect
Errors of inferences
Probability of type 1 error: α

Probability of type 2 error: β

 e.g.: treatment has an effect when
infact treatment has no effect.

e.g.: treatment has no effect when

infact treatment has an effect.
Student t-test
• The t-test is any statistical hypothesis test in which the test statistic
follows a Student's t-distribution under the null hypothesis. A t-test
is most commonly applied when the test statistic would follow a
normal distribution.
• The t test tells you how significant the differences between groups
are; In other words it lets you know if those differences (measured
in means/averages) could have happened by chance.
• It was developed by W. S. Gossett (1908) who wrote under the pen
name “student” (thus student t-test)
Assumptions of t-test
1. the observations must be independent. This implies 2 things –(i)
selection of samples from the population should be random and
the assignment of scores should be impartial
Properties of t-distribution
• The statistic t is a small sample statistics.
• t distribution forms a bilaterally symmetrical curve, with no skewness.
At the unimodal point t=0.
• It look like bell shaped curve. But its distribution is more variable
(though 0 skewness), and kurtosis greater than 3. t distributions are
more leptokurtic.
• When the n is small, the t distribution lies under the normal curve,
but the tails of the curve are higher than corresponding Normal
curve.
•
Properties of t-distribution
• t distributions vary in shape according to the number of df and n
• A critical t score is the highest t score with a given df, upto which the
observed results have a specified probability (p) of occurring by
mere chance.
Degree of freedom
• Degrees of freedom of an estimate is the number of independent
pieces of information that went into calculating the estimate. It’s
not quite the same as the number of items in the sample. In order
to get the df for the estimate, you have to subtract 1 from the
number of items.
• Why do we subtract 1 from the number of items? Another way to
look at degrees of freedom is that they are the number of values
that are free to vary in a data set.
DF- your hats
• First, forget about statistics. Imagine you’re a fun-loving person who loves
to wear hats. You couldn't care less what a degree of freedom is. You
believe that variety is the spice of life.
• Unfortunately, you have constraints. You have only 7 hats. Yet you want to
wear a different hat every day of the week.
• On the first day, you can wear any of the 7 hats. On the second day, you
can choose from the 6 remaining hats, on day 3 you can choose from 5
hats, and so on.
• When day 6 rolls around, you still have a choice between 2 hats that you
haven’t worn yet that week. But after you choose your hat for day 6, you
have no choice for the hat that you wear on Day 7. You must wear the
one remaining hat. You had 7-1 = 6 days of “hat” freedom—in which the
hat you wore could vary!