You are on page 1of 3

A hypothesis is a statement about the value of a population parameter (such as the mean or proportion) developed for the purpose

of testing, or a statement about a population parameter subject to verification. Hypothesis testing is a procedure, based on sample
evidence and probability theory, used to determine whether the hypothesis is unreasonable statement and should be rejected or is
reasonable and should not be rejected.

Hypothesis Testing Steps. 1) State null & alternate hypothesis. 2) Select a level of significance (the probability of rejecting H0
when it is true. 3) Identify the test statistic (t/z) 4) Formulate a decision rule (where will you reject H0?) (.5- α, find that # on the
t/z chart. Coordinating # is critical point. 5) Take a sample, arrive at decision (complete formula t/z). 6) Fail to reject or reject H0

H0 is a statement that needs to be tested about the numerical value of a population parameter (no difference/no change/the current
or reported condition). H1 is a statement that describes what you will conclude if you reject H0 (alternate hypothesis/research
hypothesis/boastful claim).

H0 status quo—nothing changes in situation (smoking is NOT harmful to your health (nothing changes whether you smoke or not).
H0 (this medication is NOT effective in treating this disease).

α Level of significance (level of risk) represents the probability of rejecting the null hypothesis when its true/value between 0 &
1. This represents rejection region .05 consumer research projects. .01 quality assurance. .10 political polling.

The direction of the test involving claims that use the words “has improved”, “is better than”, and the like will depend upon the
Inequality variable being measured. For instance, if the variable involves time for a certain
Keywords Symbol Part of:
medication to take effect, the words “better” “improve” or more effective” are translated
Larger (or as “<” (less than, i.e. faster relief). On the other hand, if the variable refers to a test
> H1
more) than
score, then the words “better” “improve” or more effective” are translated as “>”
(or less)
< H1 (greater than, i.e. higher test scores)
No more than ≤ H0 Notation Meaning
At least ≥ H0 n sample size
Has increased > H1 x sample mean
Is there
≠ H1 S2 sample variance
Has not S sample standard deviation
= H0
changed ρ sample proportion
Has N population size
“improved” “is
See right H1 µ (mu) population mean
better than” “is
σ2 population variance
more effective”
σ population standard deviation
Anything related to population = parameter.
π population proportion %

Mutually exclusive—exact opposites

collectively exhaustive—cover all

P-value the probability of finding a value of the test statistic this extreme when the null hypothesis is smallest level of
significance where you will reject H0 when H0 is true (prob Type I error). If P-value is less than α of: a) .10 = some evidence H0 is
not true. b) .05 = strong evidence H0 is not true . c) .01 = very strong evidence H0 is not true d) .001 = extremely strong evidence
H0 is not true. If P-value < α; reject H0 If P-value > α; do not reject H0

d.f = degrees of freedom (n-1) 1 population sampled

p = x/n (sample proportion, x = # of successes, n = # of

observations P-Value Rules
True: If reject H0 when α is .05, you will reject it when α is .01
In real life most times standard deviation not known.
False: If reject H0when α is .01, you will reject it when α is .05
You should be able to reject at any smaller alpha α
How do we differentiate between dependent and independent samples?
1.Dependent sample is characterized by a measurement followed by an intervention of some
kind and then another measurement. This could be called a “before” and “after” study.
2.Dependent sample is characterized by matched pairs; that is, matching or pairing
observations for each subject in the sample.
Why do we prefer dependent samples to independent samples?
By using dependent samples, we are able to reduce the variation in the sampling distribution.
Comparing Two Population Means (Case of Independent Samples) Comparing Tw o Population M ea ns (Ca se of Indep en dent Samples)
with Unk n own and Unequal Populatio n Variances
with Unknown but Equal Population Variances (the Pooled t-test)
Finding the value of the test statistic requires two steps. the degrees of freedom are computed by an approximation
1. Both populations must follow the normal distribution. formula, and then rounded down . The reason for rounding down
2. Both populations must have equal variances. is to reduce the number of degrees of freedom in the test, which
3. The samples are from independent populations. will require a larger value of the test statistic to reject the null

Pool the sample variances

(n1 −1) s12 + (n 2 −1)s22

s 2p =
n1 + n 2 − 2

Use the pooled sample variance in the formula to calculate the value
of the test statistic.
X1 − X 2
t =
 1 1 
s 2p  + 
 1 n 2 

Chapter 11 formulas:
Case 3: either n1 or n2 or both < 30, unknown but equal Case 4:
σ σ 2

either n1 or n2 < 30; & Case 1: n1&n2 > 30;

σ σ2
s σ u2 σ s2
kn own; (case 2) unknown calc:
σ σ 2

(((8*9.492)+(6*6.882))/14) enter known, but unequal

(unequal variances t-test)
(b efore & after tests)—dependent Sample calc: (((52/15) +
(152/12))2 / (((52/15)2/14) + ((152/12)2/11)))

Make sure if it only gives you s—be sure to make it s2 to get

variance for equation

Mean of difference between paired or related observations

sd=(d-d )2n-1 can use this or other to the left, then square it

Dependent: before & after samples/matching or pairing observations. Dependent better than independent samples, reduces variation in sampling distribution.

Chapter 10: formulas

αβ ρ π σµ